Download Document 8652271

CHAPTER 10
Summarizing discussion and future perspectives
10
CHAPTER 10
T
his thesis has focused on two unique lymphocyte subsets that play
an important role in anti-tumor immune responses and interactions
between these subsets. iNKT cells constitute a lymphocyte lineage sharing
characteristics of both T cells and NK cells. They display a highly restricted TCR
repertoire and recognize antigens in the context of a monomorphic CD1d antigen
presenting molecule [1,2]. iNKT cells have been shown to play crucial roles in
various immune responses, including antitumor responses, based on flexibility
with regards to their predominant cytokine profile. As immune-regulatory cells,
iNKT cells constitute a bridge between the innate and adaptive immune system
by the rapid release of high levels of cytokines creating a micro-environment
in which developing adaptive immune responses can be skewed [2,3]. As such,
manipulating iNKT cells potentially constitutes a valuable therapeutic approach.
Vγ9Vδ2-T cells are the predominant γδ-T cells subset in human peripheral
blood and have well-established anti-tumor effector functions [4]. They can be
activated and expanded by phosphoantigens (pAg). Aminobisphosphonates
(NBP) support the intracellular accumulation of endogenous pAg by inhibiting
the mevalonate metabolism, inducing mobility/conformational changes of the
ubiquitously expressed CD277/ BTN3A1 that triggers Vγ9Vδ2-T cell activation and
expansion [5]. Upon stimulation, Vγ9Vδ2-T cells acquire the capacity to kill solid
tumors of diverse origins, produce a diverse set of cytokines and chemokines
to regulate other cells, and are able to present antigens for αβ T cell priming.
Their abundance and high in vitro anti-tumor potential warrant further studies
to enhance their antitumor effects (in vivo and) in patients.
The fact that iNKT and Vγ9Vδ2-T cells are clinically relevant in multiple
tumor types [6–9] has triggered clinical trials focusing on these immune subsets.
Interestingly, therapeutic manipulation of both iNKT cells and Vγ9Vδ2-T cells has
been shown to induce immunological, biochemical and even clinical responses
in patients treated with specific activating ligands without causing substantial
toxicity upon activation. However, clinical results lacked consistency. Several
factors can/may play a role in the variable induction of antitumor immune
responses in clinical trials that have been performed to date, including the low
number of iNKT cells in patients with cancer, occurrence of anergy induction
upon intravenous antigen stimulation and the role of micro-environmental
conditions. In this thesis we have studied aspects of iNKT cells and Vγ9Vδ2-T
cells as well as potential interactions between these subsets in an effort to gain
more insight into cross-talk in the induction and effector phase between these
unique immune subsets. In addition we explored new therapeutic approaches
to exploit the antitumor effector functions of these cells in patients with cancer
150
Summarizing discussion and future perspectives
(summarized in figure 1).
Most striking clinical and preclinical effects with the iNKT cell activating
ligand α-GalCer have thus far been obtained when α-GalCer was loaded onto
dendritic cells (DC), either ex vivo or in situ, e.g. on skin-resident subsets [10–
12]. As described in Chapter 3 we have developed and optimized protocols to
generate immature and mature monocyte-derived DC (moDC) from peripheral
blood and to optimally expose them to α-GalCer to allow activation and biased
cytokine production in responding iNKT cells. Furthermore, as direct ligation
of CD1d using monoclonal antibodies has also been shown to induce Th1biased immune responses, this methodology was also described as a promising
approach to the immunotherapy of cancer. The impact of NKT cells in terms of
tumor control has been demonstrated by their prognostic value in patients with
cancer.
In Chapter 4, we have provided an updated analysis with a median
follow-up time of 8.7 years of a study in which circulating iNKT cell numbers were
assessed in a group of 47 patients with HNSCC before the start of curative intent
radiotherapy. This study demonstrated that patients with HNSCC and a severe
deficiency of iNKT cells have a strikingly poor clinical outcome, with respect to
overall survival, disease specific survival and loco-regional control compared
to patients without this deficiency. Furthermore, it was found that the survival
benefit of patients with intermediate/high iNKT cell numbers was not related to
differences in HPV infection. Since the correlation between a low frequency of
intra-tumoral iNKT cells and poor prognosis was also noted in neuroblastoma
[8] and colorectal cancer[9], the predictive value of iNKT cells could be a more
general phenomenon and, despite their small numbers, underscores their
important role in the regulation of anti-tumor immune responses.
Numerous studies have been carried out to characterize the cross-talk between
iNKT cells and other conventional effector cell subsets encompassing NK, T
and myeloid subsets in the immune response against cancer (see Figure 1 for
a schematic overview). In aggregate, these studies point to the versatile direct
and indirect ways in which iNKT cells can contribute to the antitumor effector
response and the elimination of tumor-mediated immune suppression, both
systemically and in the tumor microenvironment.
In contrast, relatively little is known about cross-talk between iNKT cells
and Vγ9Vδ2-T cells and how these subsets can mutually modulate each other’s
antitumor effector functions.
151
10
CHAPTER 10
Figure 1 Overview of direct and indirect antitumor effects of iNKT cells and mechanisms
of cross-talk between iNKT and other myeloid and lymphoid immune effector subsets,
resulting in reduced immune suppression and the boosting of antitumor immunity.
Adopted from [13,14].
In Chapter 5 we therefore set out to explore whether the antitumor activity of
Vγ9Vδ2-T cells could be potentiated by iNKT cells.
We found that activation of iNKT cells using α-GalCer resulted in an
enhancement of the activation of Vγ9Vδ2-T cells induced by pAg. Co-activation
of iNKT cells also enhanced the IFN-γ production as well as the cytolytic potential
of Vγ9Vδ2-T cells against tumor cells. This potentiating effect of iNKT cells was
mediated by their production of TNF-α. These observations have thus expanded
152
Summarizing discussion and future perspectives
the spectrum of immune cells affected by iNKT cells from conventional T cells,
NK cells, B cells, DC and Tregs, to include Vγ9Vδ2-T cells and have clearly shown
how this interaction may contribute to antitumor immunity.
As mentioned before, clinical trials with α-GalCer–pulsed moDC have
shown anecdotal antitumor activity in advanced cancer. However, there are
several practical limitations that hamper further clinical exploration of treatment
with α-GalCer–pulsed moDC. As Vγ9Vδ2-T cells have been shown to be able to
acquire professional antigen-presenting capacities upon their activation by pAg
[15,16], we explored their capacity to present glycolipid antigen to iNKT cells in
Chapter 6. Vγ9Vδ2-T APC were indeed found to be able to present α-GalCer to
iNKT cells resulting in their activation. However, glycolipid antigen presentation
was shown not to result from the de novo synthesis of CD1d by Vγ9Vδ2-T, but
to depend on trogocytosis of CD1d-containing membrane fragments from pAgexpressing cells with which Vγ9Vδ2-T cells interact. Although we found that
a-GalCer loaded moDC clearly outperformed Vγ9Vδ2-T cells functioning as APCs
on a per cell basis, Vγ9Vδ2-T cells had the advantage of a quicker maturation
into APCs as well as that of quantitative superiority, being more numerous as
compared to moDC precursors. Furthermore, culture supernatants of iNKT
activated via Vγ9Vδ2-T cells as APCs were more biased toward a Th1 type
cytokine profile –with a higher IFNγ/IL-4 ratio– than those iNKT cells stimulated
by moDC’s. These features make Vγ9Vδ2-T cells an interesting platform for
consideration in future clinical trials as reviewed in Chapter 7.
We next determined the effects of NBP on glycolipid Ag-induced iNKT
cell activation in vitro and found that NBP resulted in a striking reduction of
moDC-mediated α-GalCer-induced iNKT activation (Chapter 8). This inhibitory
effect was found to be caused by an NBP-induced reduction in the production
of apolipoprotein E (apoE), which is a known facilitator of trans-membrane
transport of exogenously derived glycolipids by DC. These data on the inhibitory
effect of NBP on glycolipid-mediated iNKT cell activation should be taken into
account in the design of future iNKT cell-based anti-cancer therapies, as NBP are
regularly prescribed in cancer patients with bone metastases.
Though it is known that intravenous administration of NBP can induce
powerful systemic activation of Vγ9Vδ2-T cells in vivo and that these effects are
strongly associated with the occurrence of an acute phase response (APR), the
effect of NBP on the induction of APC markers on circulating Vγ9Vδ2-T cells and
the potential of statin pretreatment of patients to inhibit Vγ9Vδ2-T cell activation
and the related occurrence of an APR have not been extensively studied. In
Chapter 9 we confirmed that in vivo treatment with NBP induced Vγ9Vδ2-T cells
153
10
CHAPTER 10
to develop towards an activated and Th1 cytokine producing effector phenotype
accompanied by clinical signs of an APR. Circulating Vγ9Vδ2-T cells did not
acquire phenotypic characteristics of APC. Although treatment with simvastatin,
started at least 1 week before the first NBP administration, was found to limit
perforin, granzyme B and HLA-DR expression by Vγ9Vδ2-T cells activated by NBP,
simvastatin was not able to fully prevent either the development of an activated
Th1 cytokine producing effector phenotype by Vγ9Vδ2-T cells or the occurrence
of APR upon NBP-infusion. This limited effect of statins could be related to an
incomplete inhibition of the mevalonate pathway at the doses used in our study
resulting in sufficient pAg accumulation in response to NBP administration for
Vγ9Vδ2-T cells to still become activated. Though higher therapeutic dosing of
statins could be more effective, another approach to prevent the APR response in
patients treated with NBP could consist of the direct blocking of pAg recognition
by Vγ9Vδ2-T cells, e.g. by using neutralizing Vγ9Vδ2-TCR specific antibodies. For
an overview of the findings from this thesis, see Figure 2.
Figure 2 Overview of the findings from this thesis outlining the interactions between
iNKT and Vγ9Vδ2-T cells with possible consequences for the treatment of cancer.
Future clinical perspectives
A general drawback of many of the clinical trials thus far performed with iNKT
cells and Vγ9Vδ2-T cells is that, though they are effective in inducing the specific
(systemic) activation of these cell subsets, they provide no specific trigger for
these cells to accumulate at the tumor site where they would be expected to
exert their antitumor effect. Also, readouts of iNKT and Vγ9Vδ2-T cell responses
154
Summarizing discussion and future perspectives
are primarily from peripheral blood rather than (tumor) tissues. Analysis of iNKT
cells and Vγ9Vδ2-T cells in the tumor-environment (e.g. cytokine responses) is
crucial, especially since down-regulation of the TCR upon activation complicates
unambiguous identification of these cells. Indeed, promising clinical responses
have been reported in patients with HNSCC after local activation of iNKT cells,
suggesting that the activation of iNKT cells within the tumor micro-environment
could be effective [17]. In light of this approach, therapies specifically targeting
iNKT cells and Vγ9Vδ2-T cells to the tumor microenvironment are of interest.
Therapeutic antibody fragments in development, e.g. single-chain Fv fragments
(scFvs), camelid VHH domains, human domain antibodies (dAbs), and an ever
expanding array of bispecific/dual-targeting variants (e.g. Bi-specific T cell
engagers (BiTEs)), are aiming at inducing Th1-biased responses in the generally
suppressive tumor-microenvironment ([18,19] and reviewed in [20]). These
targeting approaches could possibly be further supported by increasing the
frequency of Vγ9Vδ2-T or iNKT cells, for example by means of adoptive transfer
or in vivo stimulation and expansion of Vγ9Vδ2-T or iNKT cells, approaches that
have been shown to be safe therapeutic modalities resulting in objective clinical
responses in a variety of cancer types (reviewed in [21]).
One can envision that the stimulatory effect of iNKT cells on the effector
function of Vγ9Vδ2-T cells, as described in chapter 5, can be used to strengthen
future Vγ9Vδ2-T cell based immunotherapeutic approaches. Combined
presence of activated iNKT and Vγ9Vδ2-T cells at the effector site could
conceivably potentiate anti-tumor immune responses. This strategy is broadly
applicable, since both iNKT and Vγ9Vδ2-T cells can kill a wide variety of tumor
targets and have well conserved Ag recognition receptors. Another advantage
of this strategy, beside the enhancement of the anti-tumor effector functions of
Vγ9Vδ2-T cells, is the ability of Vγ9Vδ2-T cells to act as professional APC at the
tumor site. The APC function in human Vγ9Vδ2-T cells has been described by
Brandes et al. [15] and might be highly relevant to immunotherapy. As described
above, we have shown in chapter 6 [22] that Vγ9Vδ2-T APC can present antigens
to iNKT, as a result of trogocytosis of CD1d-containing membrane fragments
from pAg-expressing cells. CD1d-expressing Vγ9Vδ2-T cells were able to activate
iNKT in a CD1d-restricted and a-GalCer-dependent fashion. In order to achieve
a more effective APC platform for iNKT cell–based immunotherapy, one could
imagine using retroviral CD1d transduction of Vγ9Vδ2-T cells, hypothesizing that
the limited CD1d-restricted APC function of Vγ9Vδ2-T cells could be related to
the relatively limited surface density of CD1d, when compared with transfectants
or moDC.
155
10
CHAPTER 10
Other aspects of cross-talk between Vγ9Vδ2-T and iNKT cells in the tumor
micro-environment should be carefully studied when aiming for a (Th1-biased)
anti-tumor immune response. Vγ9Vδ2-T cells show remarkable plasticity that is
regulated by different stimuli. For instance, it has been found that activation of
Vγ9Vδ2-T cells in the presence of differential stimuli could either promote Th17or Treg-like-Vγ9Vδ2-T cells, resulting in potentially tumor-promoting responses,
favoring tumor cell proliferation and metastatic spread (reviewed in [23]).
These observations clearly suggest that identification and elimination of these
stimuli in the micro-environment could be important in establishing an antitumor immune response without unwanted collateral effects. This concept of
environmental programming is also valid for iNKT cells, as for instance Tahir et al.
found that the functional defect in iNKT cells in patients with advanced prostate
cancer, appeared to be reversible by administration of the pro-inflammatory
cytokine IL-12 [24].
Furthermore, repeated i.v. administration of α-GalCer induced iNKT cell
anergy, while repeated administration of CD1d targeted antibody constructs
resulted in sustained iNKT cell activation [25,26], suggesting that tumortargeting of iNKT cells or of α-GalCer–pulsed APC might constitute a powerful
but as yet unexplored clinical approach. Other approaches that warrant further
exploration with respect to their capacity to enhance the efficacy of iNKT cell
based immunotherapy include alternative routes of administration of α-GalCer
[12], and the use of alternative glycolipid ligands such as the non-glycosidic
compound threitolceramide, which was shown to efficiently activate iNKT cells,
resulting in maturation of dendritic cells and the priming of Ag-specific T and B
cells, while preventing ligand-induced iNKT cell lysis of pulsed dendritic cells as
well as activation-induced anergy of iNKT [27]. New and highly promising in the field of immunotherapy are immune
checkpoint inhibitors (e.g. anti-CTLA4 and anti-PD1/PD-L1 mAbs), leading to
durable immune responses and long-term disease control in different types
of cancer (e.g. in melanoma [28,29]). However, impressive responses occur in
a limited group of patients, emphasizing the need for further improvement of
this approach. We hypothesize that checkpoint inhibitors could have a beneficial
effect on antitumor responses of iNKT and Vγ9Vδ2-T cells, and vice versa that
specific activation of iNKT and Vγ9Vδ2-T cells could possibly enhance the
immune responses of checkpoint inhibitors. As mentioned previously, therapies
specifically targeting iNKT and Vγ9Vδ2-T cells to the tumor microenvironment
are of interest and could be of benefit in combination with immune checkpoint
blockade. The development of therapeutic bispecific/dual-targeting antibody
156
Summarizing discussion and future perspectives
fragments (BiTEs) seems very promising, as for example blinatumomab is already
registered for treatment of relapsed/refractory B-precursor acute lymphoblastic
leukemia (ALL) leading to durable remission by engaging T cells through CD3
of the TCR complex to induce lysis of CD19 expressing cells [30]. Many of the
current bispecific approaches target CD3 and do not discriminate between proinflammatory T cells and regulatory T cells and can thus also target Tregs to the
tumor microenvironment. Specific targeting of conserved iNKT or Vγ9Vδ2-T cells
could avoid this detrimental targeting of Tregs and perhaps further potentiate
the generation of a durable pro-inflammatory anti-tumor response.
Concluding Remarks
In summary, in this dissertation we have studied the functional interactions
between iNKT and Vγ9Vδ2-T cells, and how these can interfere with or enhance
anti-tumor immune effector responses mediated by these conserved T cell
subsets. Although we have shown that concurrent activation of iNKT cells
and Vγ9Vδ2-T cells is very potent, many challenges remain to overcome the
susceptibility of these cells to the immunosuppressive environment that cancer
cells induce. Crosstalk between these cells raises possibilities for novel iNKT
and Vγ9Vδ2-T cell based therapeutic approaches. In order for these approaches
to be successful, it is essential to also address the immune suppressive
microenvironment by novel or additional strategies. Indeed, targeting these
invariant T cell subsets to the tumor may conceivably contribute to this, by tipping
the balance in the tumor-microenvironment toward a Th1 biased response,
thereby enhancing anti-tumor immunity. Between these in-betweeners novel
solutions may thus arise for the effective immunotherapy of cancer.
References
1. H.J.J. Van Der Vliet, J.W. Molling, B.M.E. Von Blomberg, N. Nishi, W. Kölgen, A.J.M. Van
Den Eertwegh, et al., The immunoregulatory role of CD1d-restricted natural killer T
cells in disease, Clin. Immunol. 112 (2004) 8–23.
2. D.I. Godfrey, M. Kronenberg, Review series Going both ways : immune regulation via
CD1d-dependent NKT cells, J Clin Invest. 114 (2004) 1379–1388.
3. L. Wu, C.L. Gabriel, V. V. Parekh, L. Van Kaer, Invariant natural killer T cells: Innate-like
T cells with potent immunomodulatory activities, Tissue Antigens. 73 (2009) 535–545.
4. P. Vantourout, A. Hayday, Six-of-the-best: unique contributions of γδ T cells to
immunology., Nat. Rev. Immunol. 13 (2013) 88–100.
5. C. Harly, Y. Guillaume, S. Nedellec, C.-M. Peigné, H. Mönkkönen, J. Mönkkönen, et al.,
Key implication of CD277/butyrophilin-3 (BTN3A) in cellular stress sensing by a major
157
10
CHAPTER 10
human γδ T-cell subset., Blood. 120 (2012) 2269–2279.
6. M.S. Braza, B. Klein, Anti-tumour immunotherapy with Vgamma9Vdelta2 T
lymphocytes: from the bench to the bedside, Br J Haematol. 160 (2012) 123–132.
7. F.L. Schneiders, R.C.G. de Bruin, A.J.M. van den Eertwegh, R.J. Scheper, C.R. Leemans,
R.H. Brakenhoff, et al., Circulating invariant natural killer T-cell numbers predict
outcome in head and neck squamous cell carcinoma: updated analysis with 10-year
follow-up., J Clin Oncol. 30 (2012) 567–570.
8. L.S. Metelitsa, H.-W. Wu, H. Wang, Y. Yang, Z. Warsi, S. Asgharzadeh, et al., Natural
Killer T Cells Infiltrate Neuroblastomas Expressing the Chemokine CCL2, J Exp Med.
199 (2004) 1213–1221.
9. T. Tachibana, H. Onodera, T. Tsuruyama, A. Mori, S. Nagayama, H. Hiai, et al.,
Increased intratumor Valpha24-positive natural killer T cells: a prognostic factor for
primary colorectal carcinomas., Clin Cancer Res. 11 (2005) 7322–7327.
10. M. Nieda, M. Okai, A. Tazbirkova, H. Lin, A. Yamaura, K. Ide, et al., Therapeutic activation
of Valpha24+Vbeta11+ NKT cells in human subjects results in highly coordinated
secondary activation of acquired and innate immunity., Blood. 103 (2004) 383–389.
11. D.H. Chang, K. Osman, J. Connolly, A. Kukreja, J. Krasovsky, M. Pack, et al., Sustained
expansion of NKT cells and antigen-specific T cells after injection of alpha-galactosylceramide loaded mature dendritic cells in cancer patients, J Exp Med. 201 (2005)
1503–1517.
12. H.J. Bontkes, M. Moreno, B. Hangalapura, J.J. Lindenberg, J. de Groot, S. Lougheed, et
al., Attenuation of invariant Natural Killer T-cell anergy induction through intradermal
delivery of α-galactosylceramide, Clin. Immunol. 136 (2010) 364–374.
13. E. Vivier, S. Ugolini, D. Blaise, C. Chabannon, L. Brossay, Targeting natural killer cells
and natural killer T cells in cancer, Nat. Rev. Immunol. 12 (2012) 239–252.
14. V. Cerundolo, J.D. Silk, S.H. Masri, M. Salio, Harnessing invariant NKT cells in
vaccination strategies., Nat. Rev. Immunol. 9 (2009) 28–38.
15. B. Moser, M. Brandes, Gammadelta T cells: an alternative type of professional APC.,
Trends Immunol. 27 (2006) 112–118.
16. M. Brandes, K. Willimann, B. Moser, Professional antigen-presentation function by
human gammadelta T Cells., Science (80-. ). 309 (2005) 264–268.
17. N. Kunii, S. Horiguchi, S. Motohashi, H. Yamamoto, N. Ueno, S. Yamamoto, et
al., Combination therapy of in vitro-expanded natural killer T cells and alphagalactosylceramide-pulsed antigen-presenting cells in patients with recurrent head
and neck carcinoma., Cancer Sci. 100 (2009) 1092–1098.
18. K. Stirnemann, J.F. Romero, L. Baldi, B. Robert, V. Cesson, G.S. Besra, et al., Sustained
activation and tumor targeting of NKT cells using a CD1d – anti-HER2 – scFv fusion
protein induce antitumor effects in mice, J Clin Invest. 118 (2008) 994–1005.
19. S. Corgnac, R. Perret, L. Derré, L. Zhang, K. Stirnemann, M. Zauderer, et al., CD1dantibody fusion proteins target iNKT cells to the tumor and trigger long-term
therapeutic responses, Cancer Immunol. Immunother. 62 (2013) 747–760.
20. Lameris R, de Bruin RC, Schneiders FL, van Bergen En Henegouwen PM, Verheul HM,
de Gruijl TD, Vliet van der H.J, Bispecific antibody platforms for cancer immunotherapy,
Crit Rev Oncol Hematol. 92 (2014) 153–165.
21. C.L. Deniger DC, Moyes JS, Clinical applications of gamma delta T cells with multivalent
158
Summarizing discussion and future perspectives
immunity, Front Immunol. 5 (2014) 1–10.
22. F.L. Schneiders, J. Prodöhl, J.M. Ruben, T. O’Toole, R.J. Scheper, M. Bonneville, et al.,
CD1d-Restricted Antigen Presentation by Vγ9Vδ2-T Cells Requires Trogocytosis.,
Cancer Immunol. Res. 2 (2014) 732–740.
23. B.N. Lafont V, Sanchez F, Laprevotte E, Michaud HA, Gros L, Eliaou JF, Plasticity of γδ T
Cells: Impact on the Anti-Tumor Response, Front Immunol. 5 (2014) 1–13.
24. S.M. Tahir, O. Cheng, A. Shaulov, Y. Koezuka, G.J. Bubley, S.B. Wilson, et al., Loss of
IFN-gamma production by invariant NK T cells in advanced cancer., J. Immunol. 167
(2001) 4046–4050.
25. T. Iyoda, M. Ushida, Y. Kimura, K. Minamino, A. Hayuka, S. Yokohata, et al., Invariant
NKT cell anergy is induced by a strong TCR-mediated signal plus co-stimulation., Int.
Immunol. 22 (2010) 905–913.
26. B.A. Sullivan, M. Kronenberg, Activation or anergy: NKT cells are stunned by alphagalactosylceramide., J Clin Invest. 115 (2005) 2328–2329.
27. J.D. Silk, M. Salio, B.G. Reddy, D. Shepherd, U. Gileadi, J. Brown, et al., Cutting edge:
nonglycosidic CD1d lipid ligands activate human and murine invariant NKT cells., J.
Immunol. 180 (2008) 6452–6456.
28. J.D. Wolchok, H. Kluger, M.K. Callahan, M. a Postow, N. a Rizvi, A.M. Lesokhin, et al.,
Nivolumab plus ipilimumab in advanced melanoma., N. Engl. J. Med. 369 (2013) 122–
33.
29. M.A. Postow, J. Chesney, A.C. Pavlick, C. Robert, K. Grossmann, D. McDermott et al.,
Nivolumab and Ipilimumab versus Ipilimumab in Untreated Melanoma., N Engl J
Med. 372 (2015) 2006–2017.
30. M.S. Topp, N. Gökbuget, A.S. Stein, G. Zugmaier, S. O’Brien, R.C. Bargou, et al.,
Safety and activity of blinatumomab for adult patients with relapsed or refractory
B-precursor acute lymphoblastic leukaemia: a multicentre, single-arm, phase 2
study., Lancet Oncol. 16 (2015) 57–66.
159