University of Groningen
Modeling of MLL-AF9-rearranged pediatric leukemia Carretta, Marco
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2018
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Carretta, M. (2018). Modeling of MLL-AF9-rearranged pediatric leukemia: Identification of mechanisms and potential targets. University of Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
CHAPTER
6
Summary, general discussion and future perspectives
The work presented in this thesis focuses on the establishment of novel in vitro and in vivo models to study mechanisms of leukemic transformation and to identify new potential druggable targets. MLL-AF9, a particular subtype of pediatric leukemia, was chosen as the main model for these studies because of its unique biological features and also because of its high frequency in pediatric leukemias.
The past thirty years have seen increasingly rapid advances in the development of in vitro and in vivo models to study human hematopoiesis, malignant hematopoiesis and hematopoietic stem cells (HSC). The knowledge derived from humanized mouse models has enabled researchers to start dissecting the physiology and pathophysiology of human hematopoiesis. Whereas techniques such as multicolor flow cytometry and genetic analysis constantly improve and allow in depth molecular characterization, the xenograft models remain a critically important assay to functionally define HSCs and their malignant counterparts. However, current humanized xenograft models still possess limitations as normal and malignant human hematopoiesis cannot be faithfully recapitulated and a large subset of primary patient material still cannot engraft 1. Besides the characteristics of the immunodeficient mice strains, several other factors are known to affect the engraftment of human acute myeloid leukemia (AML) cells such as conditioning of the animal, the sex, and the route of administration of the cells 2. Moreover, it is clear that human hematopoiesis is regulated by a specialized microenvironment also termed the bone marrow (BM) niche, where a complex network of different cell types and signaling molecules are regulating stem cell survival, maintenance and differentiation. Therefore, the functionality of human hematopoietic cells in a mouse environment may change, probably due to the lack of cross-reactivity of specific factors between the mouse host and the human graft.
Previously in our lab, we have been able to establish various retroviral/lentiviral models to study human myeloid and lymphoid leukemias. In vitro, CD34+ cord blood (CB) cells transduced with BCR-ABL or MLL-AF9 oncogenes (either alone or in combination with other hits) could induce transformation along myeloid or lymphoid lineages. When the same cells where injected intravenously in NOD-SCID-based xenograft models, including the NOD-SCID IL2Rγ−/− (NSG) mice, transformation was often heavily lymphoid biased and myeloid transformation has been more difficult to achieve. In chapter 2 of this thesis, we took advantage of an in vivo model in which NSG mice are subcutaneously implanted with ceramic scaffolds seeded with human mesenchymal stromal cells in order to generate a human bone marrow (huBM-sc)-like niche. We hypothesized that the presence of a human microenvironment could provide extrinsic and species-specific factors, which might dictate the lineage fate of the leukemic clones. Furthermore, this specialized microenvironment may provide a platform for human-specific direct interactions of leukemic cells with stromal cells such as MSCs and the extracellular matrix.
Injection of non-sorted BCR-ABL/BMI1-transduced CB CD34+ cells in huBM-sc mice resulted in the development of tumors at the site of injection in 11 out of 13 mice, and in contrast to the NOD-SCID/NSG mouse in which exclusively CD19+ B-lymphocytic acute leukemia (B-ALL) was observed 3, both myeloid and lymphoid disease could be recapitulated. In previous studies, it has been shown that in NOD-SCID/NSG mice a serially transplantable CD19+ B-ALL could be induced only when additional hits were provided, such as co-expression of BMI1 4. In the huBM-sc xenograft model
that we presented here, both BCR-ABL-only expressing cells as well as with cells that expressed both BCR-ABL and BMI1 were capable of engrafting and inducing
leukemia. These data suggest that ectopic expression of BMI1 is not an absolute requirement to allow transformation of human CB CD34+ cells transduced with only BCR-ABL. Interestingly, transcriptome analyses of leukemic cells derived from murine BM niches versus leukemic cells derived from the huBM scaffold niches revealed that endogenous BMI1 was strongly upregulated in the human BM niche. Hence, it could be hypothesized that the human BM niche provides factors that induce BMI1 expression, possibly through activation of SALL4 or the Sonic Hedgehog pathway, which both have been shown to be able to induce BMI1 expression 5–8. Moreover, transcriptome analyses also revealed that in the humanized microenvironment stemness of HSCs/LSCs was better preserved, as the top differentially expressed genes were enriched for cell cycle inhibitors (like CDKN1C/p57) and hypoxic signature-related genes. Efficient engraftment in the huBM-sc xenograft model was also observed when primary CD34+ CML patient cells were injected and, in contrast to the murine niche, leukemic cells derived from the huBM scaffold more closely resembled the original patient immunophenotype, in particular with regard to the percentage of CD34+ cells.
The myelomonocytic and B-ALL phenotypes that are observed in MLL-AF9 pediatric patients were also recapitulated when CD34+ CB cells transduced with MLL-AF9 fusion oncogene were injected into huBM-sc mice. Compared to the murine BM environment of NOD-SCID/NSG mice, a higher percentage of myeloid clones could be generated in the huBM-sc model. These observations strengthen the notion that the lineage fate of MLL-AF9 LSCs is dictated primarily by microenvironmental cues and cell of origin, rather than by a potential instructive role of the fusion partner gene. Similarly to the NOD-SCID/NSG model, huBM-sc derived B-ALL clones could easily serially transplant, but secondary engraftment of myeloid clones was more difficult to
achieve. These data indicate that, even though myeloid engraftment was increased in the primary mice, factors that permit robust maintenance of myeloid-permissive LSCs are still lacking, at least for lentiviral CB MLL-AF9-transduced models.
Besides lentiviral models for MLL-AF9, we also showed that primary AML MLL-AF9 derived from an adult patient could engraft efficiently in the huBM-sc. Lastly, we show that the huBM-sc model could also be used as a model for drug testing. Four weeks after injection of the CB B-ALL MLL-AF9 cells generated in huBM-sc mice we performed daily i.sc. injections with I-BET151 inhibitor. Tumor volume size of each scaffold was measured during the experiment every 3 days, allowing us to monitor the progression of the disease over time, and a reduction in tumor volume was observed.
Overall, in this study we established a huBM-sc xenograft model in which the myeloid and lymphoid features of MLL-AF9 and BCR-ABL leukemias can be studied in detail, both using lenti/retroviral model systems as well as patient samples and in which the efficacy of novel drugs can be evaluated.
In parallel studies by our laboratory, a large cohort of patient samples covering all important genetic and risk subgroups successfully engrafted in the huBM-sc model, and stem cell self-renewal properties were better maintained compared to murine niches, as determined by serial transplantation assays and genome-wide transcriptome studies 9.
The presence of an ectopic human BM niche presented clear advantages compared with normal NSG mice, but some key issues remained. For example, transcriptome studies of primary human BM-derived MSCs revealed that a variety of cytokines and growth factors are produced, but some critically important cytokines, such as IL-3
and TPO, are not. In chapter 3, we investigated whether we could genetically engineer MSCs to produce such factors and evaluated these modified MSCs in vitro and in vivo in humanized scaffold xenograft models. We specifically focused on IL3 to favor myeloid transformation and we included TPO which is among other factors like ANGPT1, TGFb and IL1 important in maintaining LSC self-renewal 10–15. Prior to the in vivo experiments, we showed that these IL-3- and TPO-producing MSCs were superior in expanding human CB CD34+ hematopoietic stem/progenitor cells when co-cultured in vitro. Moreover, MLL-AF9-transduced CB CD34+ cells could be transformed efficiently along myeloid or lymphoid lineages, in the absence of exogenous cytokines. These in vitro studies gave us also an indication about the ratio of cytokine-producing MSCs to choose when coating huBM-sc in vivo in order to reach concentrations of secreted IL3 and TPO between 10-50 pg/ml. Next, we assessed the ability of IL-3-MSCs and TPO-MSCs to differentiate in vivo and immunohistochemistry analysis performed six weeks after implantation showed no differences in the ability of forming bone, fat tissue, or other stromal components when compared with wild type MSCs. We then tested three primary AML, B-ALL, and biphenotypic acute leukemia (BAL) patient samples and they all efficiently engrafted in the IL-3/TPO-BM-sc model, with latencies and immunophenotypes that did not significantly differ from what we had observed in our huBM-sc model previously. The main differences were that immature CD34+ cells for the B-ALL sample were better preserved in our new IL-3/TPO-BM-sc model, and the same was true for the myeloid CD33+ immune phenotype for the BAL patient sample. It is currently unclear why the B-ALL patient sample in particular benefitted the most from the IL-3/TPO niche, and it will be interesting to test a larger cohort of patient samples to evaluate which leukemia might benefit the most of this humanized cytokine-producing niche.
Furthermore, engraftment of MLL-AF9-transduced CB cells was evaluated in the new IL-3/TPO-BM-sc model and compared to the data with our previous huBM-sc and intravenously injected NSG models. Injection of MLL-AF9-transduced CB cells into IL-3/TPO-BM-sc models resulted in the development of a fatal leukemia in all experimental animals. Unexpectedly, the efficiency of tumor formation on the injected scaffolds themselves was reduced by ∼55% in the IL-3/TPO-BM-sc model compared with the previous huBM-sc model. Possible explanations for this might be that the local concentrations of IL-3 and TPO produced by the genetically engineered MSCs would be non-physiological, resulting either in the differentiation or the loss of self-renewal properties or, potentially, in the migration of leukemic cells to other mouse niches where the cytokine concentrations would be less high. In fact, for the first time, we detected full myeloid AML clones in PB, BM, spleen, and liver. As already observed in the IV NSG and ISC huBM-sc models, CD33+-sorted myeloid clones failed to self-renew in secondary recipients also in the IL-3/TPO-BM-sc model, whereas B-ALL clones could readily engraft and give rise to secondary leukemia. This finding was somewhat unexpected and suggests that the presence of a modified human microenvironment that overexpresses IL-3 and TPO does not allow myeloid clones to self-renew properly, but a further fine-tuning of the levels of cytokines produced in the humanized niche might be required to solve these issues.
This study indicates that the humanized scaffold xenograft model allows for relatively simple genetic engineering of the BM microenvironment. Therefore, this approach will be useful for functional study of the importance of niche factors for normal and malignant human hematopoiesis.
Fig.1: Schematic representation of the different approaches used in modelling human leukemia in NSG mice
Establishing in vitro and in vivo leukemia models is not only instrumental to unravel the molecular mechanisms of the disease but also very useful to identify druggable targets and test new drugs. Over the last years, a large body of reports has highlighted the important role of epigenetic regulators in the pathogenesis of various types of cancer, including leukemias. One example of the latter is represented by MLL-fusion leukemias that, although characterized by an adverse prognosis, possess a relatively simple genetic background where the MLL-fusion gene is in most cases the only driver mutation. Therefore, MLL-rearranged leukemias have been studied extensively to evaluate the influence of epigenetic regulators on the oncogenic transcription programs. In chapter 4, we took advantage of the previously developed in vitro and in vivo models to identify targetable signaling networks in human MLL-AF9 leukemias. First, we show that MLL-MLL-AF9 cells critically depend on FLT3 ligand-induced pathways for their survival. Both the initiation of transformation as well as the maintenance of transformed leukemic cells critically requires the presence of FLT3-ligand. We also find that the FLT3-receptor is strongly upregulated by MLL-AF9, coinciding with a direct binding of MLL-AF9 to the FLT3 locus, thereby facilitating
FLT3 signaling. Downstream, we found that CB MLL-AF9 cells heavily depend on the activity of the serine/threonine kinase TAK1 and NF-kB. Here we show that FLT3-ligand stimulated MLL-AF9 cells also do not tolerate inhibition of either TAK1 of NF-kB activity. Transcriptome studies revealed that deprivation of FLT3 ligand resulted in loss of genes that were enriched for the GO terms DNA replication and G1/S cell cycle transition, coinciding with a loss of S/G2/M and accumulation of cells in G1 in cell cycle analysis studies, and also expression of the survival gene BCL2 depended on the presence of FLT3 ligand.
These data show strong overlap with a dependency of MLL-AF9 cells on BRD3/4 activity. We evaluated the in vitro and in vivo efficacy of the BRD3/4 inhibitor I-BET151 in various human MLL-AF9 (primary) models and patient samples and in line with what was published previously 16,17 we find good efficacy upon treatment with I-BET151. In both myeloid- and lymphoid-transformed CB MLL-AF9 cells we found a strong dependence of BRD3/4 activity. ChIP-seq experiments revealed that many of the genes regulated by BRD3/4 were not directly bound by MLL-AF9, indicating that BRD3/4 does not directly control MLL-AF9 chromatin recruitment but rather suggest that it controls independent pathways for the survival of MLL-AF9 cells. While it was initially proposed that BRD4 would directly recruit MLL-fusion proteins to the chromatin 17 it has now become clear that MLL-fusion proteins/DOT1L and BRD4 act independently, whereby DOT1L results in methylation of H3K79 followed by transcription factor-mediated recruitment of EP300 that deposits an acetyl mark on H4K5 18. This in turn forms a docking site for BRD4 thereby facilitating the target gene expression and co-inhibition of DOT1L and BRD4 has been shown to act synergistically in targeting MLL-rearranged leukemias 18. Another interesting finding is that, although MLL-AF9 cells were sensitive to I-BET151, the sensitivity under
liquid culture conditions was much stronger compared to the sensitivity observed under stromal co-culture conditions. Although the exact underlying mechanisms are not yet clear, it is quite possible that the bone marrow niche can provide protective signals for leukemic cells. For instance, survival pathways controlling BCL2 can also be activated by various (secreted) factors arising from stromal cells, a phenomenon that will certainly be further investigated in our future studies. Possibly, this also explains why we do not observe a complete eradication of MLL-AF9 cells in our in vivo humanized niche xenograft model in which leukemias are grown in the presence of a microenvironment composed of human mesenchymal stromal cells.
In conclusion, a concept emerging from our experiments is that although MLL-AF9 drives an aberrant gene expression program regulating for instance the HOXA cluster, it still critically relies on non-mutated transcription factors or tyrosine kinases. We show that BRD3/4 and the FLT3-TAK1/NF-kB pathways collectively control a set of targets that are critically important for the survival of human MLL-AF9 cells.
The selective BCL2 inhibitor, ABT199 (venetoclax), has shown promising results in clinical trials against chronic lymphocytic leukemia and other hematopoietic malignancies 20, but recently it was also shown to be effective both in vitro and in xenograft models against MLL rearranged (MLL-r) leukemias in combination with chemotherapies and DOT1L inhibitor 21,22. These data suggest that targeted agents against BCL2 may prove useful adjuncts to therapy in MLL-r leukemias.
Furthermore, clinical trials involving FLT3 inhibitors such as Quizartinib 23 have been recently started in both adult and pediatric leukemias. One study was a phase 1 trial evaluating the safety of quizartinib (AC220) in combination with high-intensity chemotherapy for relapsed childhood leukemia 24. The first results showed that the
inhibitor exert near-maximal suppression of FLT3 phosphorylation in the patients, but did not impact on overall survival. The study was not powered to make definitive conclusions about the efficacy of the treatment, and the outcome of other clinical trials currently ongoing will help to understand whether FLT3 inhibition is of therapeutic value in MLL-r leukemia.
In the final experimental chapter of this thesis (chapter 5), we took advantage of our in vitro and in vivo models to investigate the rare phenomenon of switches between lymphoid and myeloid lineages that can occur during treatment or during relapse of pediatric leukemia patients bearing MLL-translocations. Conversion of the leukemic cell lineage is almost exclusively associated with MLL-r pediatric leukemia and in the clinic most cases involve the conversion of B-ALL to AML. Our data show that upon environmental changes it is possible to convert CD19+ B-ALL cells to a complete myeloid phenotype, in a CB transduced MLLAF9 model as well as primary MLL-AF4 pediatric patient sample. Clonality studies and RNA-sequencing performed before and after the switch confirmed the same clonal origin and a complete switch from a lymphoid to myeloid gene signature. Conversely, CD33+ AML cells driven my MLL-AF9 were able to convert to a lymphoid phenotype only within the first two weeks of culture, while cells cultured for longer times did not generate CD19+. Conversion of CD19-CD33+ to lymphoid lineage was not achieved also in primary MLL-AF4 pediatric cells and adult MLL-AF9 cells. Overall, these findings highlight the importance of the microenvironmental signals in fate decision of MLL-r leukemia, but also point to the established notion that these leukemias can originate from different cells, and several models are proposed in chapter 5. In case the initial oncogenic event occurs at the level of HSCs or multipotent CD34+ progenitors the immortalization by MLL-AF9 can lead to ALL, AML or BAL. The LSCs which maintain
these leukemias are multipotent and lineage output can be influenced by microenvironmental cues. AML driven by MLL-AF9 instead can also be initiated by a multipotent LSC, but as the disease progress in time these LSCs convert to more myeloid-restricted leukemic GMPs that maintain the AML which lose their ability to produce lymphoid output. An alternative third scenario is that the initial oncogenic event does not occur in the immature multipotent HSC but occurs at the level of more committed hematopoietic cells such as GMPs. In this case, the LSC is natively lineage-restricted and the immortalization by MLL-AF9 can lead only to AML, excluding the possibility of conversion to the lymphoid lineage.
Lineage switching offers an interesting example of the lineage heterogeneity observed in acute leukemias and progress in the available sequencing technologies, in vitro and in vivo models and knowledge of leukemia pathogenesis will help to understand the exact mechanisms driving lineage switching during treatment or relapse.
Discussion and future perspectives
In this thesis we aimed to improve the current in vitro and in vivo models to study human leukemia in order to establish more physiologically and clinically relevant systems. We mainly focused on MLL-AF9 leukemia because of its high incidence in pediatric cases and also because of its rather unique biological features. In the next paragraphs, we will discuss remaining questions and suggest future directions of research.
A bone to pick: future directions in the development of BM humanized xenograft mice models
We showed that the implantation of an ectopic human bone marrow niche supports the engraftment and growth of leukemic CB models and patient samples, where it better preserves the leukemic stem cell self-renewal properties 9,25. Our approach relies on calcium phosphate (BCP) particles seeded in vitro with human MSCs, where differentiation into different lineages, including endochondral ossification, can be primed. Subsequently, these humanized BM scaffolds can be surgically implanted subcutaneously in the back of NSG mice.
Contemporary to our studies, other investigators have generated similar ectopic humanized marrow organs using different protocols. Reinisch et al. developed a humanized niche model where by contrast, in vitro-expanded MSCs mixed with Matrigel are simply injected into the subcutaneous space where they undergo differentiation in situ, and form a humanized BM microenvironment. Due to the absence of a substrate such BCP to induce efficient bone formation, daily injections of human 1-34 parathyroid hormone (PTH) are necessary for a period of 4 weeks. The administration of PTH is necessary to increase ossicle weight and marrow cavity
size, but it is quite laborious, costly and still does not reach the same bone formation efficiency of the BCP scaffold-based model. Using this approach, the authors showed that AML samples could engraft efficiently including good risk PML-RARa samples, which are known to be difficult to engraft in NSG mice 26. Furthermore, they also observed robust engraftment of myelofibrosis (MF) samples, and identified leukemia-initiating cell (LICs) in these malignancies. This method supports the hypothesis that a human BM microenvironment is beneficial for the growth of leukemia cells, and presents some advantages such as avoiding surgical implantation and allows the formation of larger niche sizes. Current studies in the Experimental Hematology lab (UMCG) now combine small BCP particles (that can be injected by a 16 gauge needle) mixed with Matrigel, in order to avoid surgery and generate larger niche sizes but still having the advantage of particles on which bone can be deposited. The mix of Matrigel and BCP particles should also favor a superior level of vascularization.
In the models discussed thus far, the vasculature of the ectopic human BM niches are murine, and experimental evidence suggests that also human endothelial cells can secrete species-specific factors that might have regulatory functions. Chen et al. have already described a BM niche model in which MSC and endothelial colony-forming cells are implanted subcutaneously into immunodeficient mice 27. Further studies are needed to determine whether the addition of human EPCs further enhances the humanized niche model.
Besides offering a favorable environment for the engraftment of human hematopoietic cells (normal or malignant), the humanized BM xenotransplant models offer other substantial advantages. For instance, these methods minimize the loss of
cells generally experienced with intravenous injections and therefore will allow the transplantation of lower cell numbers to achieve robust engraftment (an issue that is of course also circumvented by intrafemoral injections in mice, but these normal mice do not possess a human BM environment). Furthermore, injecting cells into multiple scaffolds in the same mouse increases the chances of successful engraftment and reduces the number of mice needed for experiments, as leukemia engraftment in every scaffold can be studied. The humanized BM xenotransplant model can also be used to investigate homing and migration of transplanted cells between scaffolds in case not all scaffolds are directly injected.
Finally, a number of further improvements of the humanized microenvironment implanted in xenograft mice models can be envisioned. The study presented in Chapter 3 of this thesis indicates that the humanized scaffold xenograft model allows for relatively simple genetic engineering. Therefore, this approach paves the way for further functional studies on the importance of niche factors for normal and malignant human hematopoiesis, with the ultimate goal of establishing tailored humanized mouse models that might be used for specific disease questions. Lastly, another approach would consist of using MSCs directly derived from the patient. This approach could be of great relevance for those malignancies where it has been shown that the MCS directly contribute to the pathogenesis, possibly due to (epi)genetic alterations that have occurred.
Collectively, these data demonstrate that the presence of a human BM microenvironment improves the efficacy of in vivo models to study leukemia, and further refinements of these protocols will hopefully extend the use of these models for all the different leukemic subtypes
Reconstitution of human immune function in xenograft models
The recent success of immunotherapy such PD-L1 inhibitors and CAR-T cell therapy constitute direct clinical proof that cancer can be treated through the modulation of immunity, a result that was first predicted on the basis of mouse models of cancer 28. On the basis of evidences that the interaction between the tumor and the immune system determines patient fate, it is crucially important to investigate cancer and its treatment in realistic mouse models. The highest goal of humanization consists in generating mice with a fully competent human immune system that are able to exhibit anticancer immune responses and thus enable more realistic predictions of the clinical response. To achieve this objective, it is necessary to introduce malignant and immune cells, ideally from the same donor, into mice while creating an environment that guarantees full compatibility between the graft and the host.
As extensively discussed in this thesis, NSG mice are the most widely used immunodeficient host for xenotransplantation. In the NSG recipients, the differentiation of human HSC into functional myeloid cells and NK cells is limited 29, B cells do not undergo sufficient maturation to become memory and antibody-producing cells 30 and maturation of T cells in adult animals is disturbed due to the lack of thymic support. Furthermore, the ratio of B- to T-lymphocytes is strongly biased towards B-cell differentiation and does not mimic the human system. It has been hypothesized that the limitations are attributable to a lack of cross-reactive mouse cytokines and possibly other chemotactic factors. To provide human-specific growth factors, transgenic mice have been generated such as the NSG-SGM3 mouse expressing human IL3, GM-CSF and stem cell factor (SCF), which proved to be a significantly better host for at least a subset of AML samples 31,32. However,
since these are knock-in mice in which artificial promoters drive cytokine expression, it has been noted that the non-physiological levels of the human factors can drive the mobilization of hematopoietic stem cells and limit the long-term engraftment. To achieve a physiological expression of human factors, in 2014 Rongvaux et al. published a paper in which they described the generation of two new mouse strains, the MITRG and MISTRG. Using Rag2-/-Il2rg-/- mice as background, the researchers have knocked-in 4 genes encoding for human M-CSF, IL-3, GM-CSF and TPO in their respective mouse loci, these mice are known as MITRG mice. Another strain, the MISTRG mice, together with these 4 human cytokines also carries a BAC-transgene encoding for human SIRPα. Binding of SIRP1α to CD47, a protein expressed on (nearly all) human cells including leukemic cells, suppresses phagocytosis. Thus, the expression of human SIRPα enables mouse phagocytes to ‘tolerate’ and not engulf engrafted human cells. The humanization of this gene is necessary since these mice have a Rag2-/-Il2rg-/- background, but is not required in all the NOD strains 2. In NOD mice the allele of Sirpa is very similar to the human SIRPA allele and NOD-SIRPα binds to human CD47. The advantage of MITRG/MISTRG mouse models is represented by the fact that human cytokine genes are replacing their mouse counterparts, providing a correct context-specific expression. Furthermore, the combined effect of multiple cytokines allows the recapitulation of human myeloid development and functions in the mouse, in particular of the monocyte and macrophage compartment which is especially important in tissue homeostasis, inflammation, tumorigenesis and in the response to infectious agents. Additionally, the expression of the IL15/IL15Rα complex on human monocytes has been shown to support the development and function of human NK cells 33. The success of this strain potentially lies in the mix of specific human
cytokines and the fact that these cytokines are knocked in to the endogenous mouse loci, allowing proper regulation and contextual control overexpression that is not possible in other mice.
Similarly to our approach of transplanting a human BM environment in NSG mice, researchers have also attempted to co-transplant parts of a fetal human thymus to provide support to the development of human T cells, as well as their selection on human MHC molecules. In this model, named “BLT” (bone marrow-liver-thymus), fragments of human fetal liver and thymus are co-transplanted under the mouse kidney capsule 34. BLT mice sustain efficient human thymic lymphopoiesis, and T cells are the main component of the human graft. However, adaptive immunity in BLT mice is still insufficient and not very different when compared to transplantation of human CD34+ cells in NSG mice 35. Recently, it was shown that the transplantation of human embryonic stem cell-derived thymic epithelial progenitors into thymus-deficient nude mice supports the development of functional human T cells 36. Although potentially very useful, these approaches have the major limitation of being not accessible to most investigators due to the limited availability of embryonic and fetal tissues. Moreover, this model posed a new challenge in the field since there are experimental evidences that the human immune system is not entirely tolerant for the mouse host 37. Thus, future development of humanized mice also needs to focus on inducing tolerance of the human immune system for the mouse host.
Further refinement of mouse models and humanized mice will contribute significantly to disease understanding and development of novel therapeutic approaches. In future humanized models, ulterior improvements are needed to strengthen the development of human B and T cells, their education and maturation, the
homeostasis of those cells in the periphery and the appropriate formation of secondary lymphoid structures (such as lymph nodes with functional germinal centers). These models will allow studying the interaction between human cancer and human immune effectors in greater detail and hopefully will improve treatment for all types of cancer.
In vivo models to study leukemia clonal heterogeneity
Acute leukemia is characterized by an uncontrolled clonal proliferation of abnormal progenitor cells in the bone marrow and blood. Advances in genome sequence technologies have revealed the spectrum of somatic mutations that give rise to human leukemia and drawn attention to its molecular evolution and clonal architecture. This, in combination with the use of in vivo models, has greatly contributed to our understanding of the evolution and pathogenesis of acute leukemia. It is now evident that most leukemias harbor small numbers of mutations, which are acquired in a stepwise manner. This characteristic makes acute leukemia an excellent model for understanding the principles of cancer evolution in general 38.
Studies on AML and myelodysplastic syndromes (MDS) demonstrated that these myeloid disorders exhibit clonal heterogeneity that evolves upon disease progression and/or relapse 39–41. Besides genetic heterogeneity, functional heterogeneity also exists and it has been studied in the context of identifying initiating tumors cells into immunodeficient mice. However, the relationship of these initiating cells (or cancer stem cells) to the clonal organization of a tumor is not yet clear.
In a recent study, the group of Timothy Ley assessed the relationship between functional and genetic heterogeneity by following genetically-defined subclones under different experimental and biological conditions in de novo AML samples 42.
The authors found that individual cell populations with well-established cellular phenotypes in the peripheral blood at the time of AML diagnosis were derived from the AML founding clone; in some cases, genetically defined subclones corresponded to distinct cell populations that could be identified by cell surface markers. When cells were injected in immunodeficient mice only one subclone engrafted in most mice even though multiple subclones were present in the sample that was injected. In some instances, the engrafting subclone represented only a small fraction of the injected cells (<10%), implying that some subclones have a cell-intrinsic advantage after transplantation. These experimental evidence suggests that conventional NSG mouse xenograft assays do not fully recapitulate the clonal heterogeneity of patient samples, resulting in engrafting subclones that may be distinct from the dominant founding or relapse clone, presumably due to selective pressure imparted by the mouse BM microenvironment. In the same study, the authors also tested AML engraftment also in the previously discussed NSG-SGM3 model. This recipient preferentially support certain AML sub-clones that are probably more sensitive to regulation and support by human cytokines produced in vivo from the transgene. Intriguingly, the most abundant clone found in NSG-SGM3 mice was not the most abundant clone found in the AML sample injected. This suggests that minor sub-clones present in the injected sample can become dominant depending on the different driving or selecting milieu present in the mouse.
In our and Reinish humanized BM scaffold models, similar genetic approaches were used to evaluate clonal heterogeneity and clonal drift in the injected samples in comparison to a murine environment 9,26. In both studies, the authors showed that within the humanized BM niches, the clonal heterogeneity of the original patient blasts was more closely recapitulated, whereas engrafted cells in the mouse BM after
intravenous transplantation tended to show variable variant allele frequencies. Together, these data suggest that the presence of a humanized niche might be better to maintain clonal heterogeneity as observed at diagnosis.
Thus, xenograft models for AML not only define leukemic stem cells and are preclinical therapy models, but also represent a tool allowing dissection of the clonal architecture of AML and the functional properties of AML sub-clones that could become targets for therapy. Mice that produce “myeloid” cytokines, (such as the cutting-edge MISTRG model) and humanized BM models seem to move the field in the right direction, but the selective and preferential support of hematopoietic sub-clones need to be taken into consideration.
Intrinsic and extrinsic cues in lineage fate decision of MLL rearranged leukemias
As pointed out earlier in this thesis, MLL rearrangements have the capacity of giving rise to both acute myeloid leukemia and acute lymphoid leukemia. Intriguingly, certain fusion partners are associated more frequently with AML than with ALL and vice versa 43. For example, the t(4;11) translocation, which result in MLL-AF4, is most frequently found in ALL, while t(9;11) that generates the MLL-AF9 fusion is the most common genetic alterations of MLL in AML. It is thus tempting to speculate that the MLL fusion partner gene may play a crucial instructive role in determining the lineage fate of the leukemia. Studies in mice using knock-in technology or bone marrow retroviral transduction posed a challenge to this theory, since depending on the combination of MLL fusion partner and mouse model used an exclusive AML or ALL disease phenotype was observed.
took advantage of the transgenic NSG-SGM3 mice, and found that when MLL-AF9 cells from the myeloid or lymphoid cultures were injected, only aggressive myelomonocytic leukemia developed. This was not the case when the same cells were injected into NSG mice, where injection of MLL-AF9 cells resulted in a mix of AML, acute B-lymphocytic leukemia (B-ALL) and acute biphenotypic leukemia (BAL). While B-ALL and BAL cells were present in the injected mice, a distinct population of AML was also detected, implying that separate LSCs had proliferated in the mice or that a single LSC could generate both AML and ALL. Thus, xenograft models revealed that the cytokine milieu of leukemic stem cells profoundly affects their lineage commitment 44. In support of these findings, we also found that when CB MLL-AF9 cells were injected in our huBM-sc and cytoBM-sc model, significantly more myeloid leukemia was observed compared to the normal NSGs that exclusively generate lymphoid.
Recently, researchers took once again advantage of the NSG-SGM3 to test the susceptibility to microenvironmental cues of MLL-AF4 transduced cells. As previously mentioned, MLL-AF4 is almost exclusively associated with ALL in patients 45. When MLL-AF4 cells were injected in normal NSG mice, a full lymphoid transformation was observed. But when cells were injected in the myeloid microenvironment of the NSG-SGM3 mice instead, cells displayed a compromised transformation capacity. AML MLLAF4 cells generated in NSG-SGM3 mice in fact were unable to self-renew in secondary recipients or in some cases they underwent a lineage switch back to ALL. Gene set enrichment analysis showed that the HSC gene signatures were underrepresented in MLL-AF4 myeloid cells compared to MLL-AF9 myeloid cells, potentially accounting for the diminished AML transformation of MLL-AF4 cells. In contrast, this incomplete activation of the HSC program by MLL-AF4 was not seen in
ALL cells. The authors found that MLL-AF4 activated a self-renewal program in a lineage-dependent manner, showing that the leukemogenic activity of MLL-AF4 was interlinked with lymphoid lineage commitment. These data suggest that the defective self-renewal ability and leukemogenesis of MLL-AF4 myeloid cells could contribute to the strong B ALL association of MLL-AF4 leukemia observed in patients. It could also be argued that in B-ALL the cell of maintenance is a multipotent HSC while the cell of maintenance in AML is a myeloid restricted leukemic-GMP with much less expression of stem cell signatures.
Similarly to MLL-AF9, lineage switches between ALL and AML after CD19-directed therapy were reported for two t(4;11) ALL patients 45. These two patients then received myeloid-directed therapy followed by allogeneic hematopoietic stem cell transplantation. While one infant patient achieved complete remission, the adult patient died from AML relapse. The ALL-to-AML switch can also occur during chemotherapy and is usually associated with poor outcome. Acquisition of additional genetic events enhancing the self-renewal program may also give rise to the occasional MLL-AF4 AML in t(4;11) AML patients.
It appears increasingly clear that lineage fate in the context of MLL-rearrangement leukemia imply a complex interplay of oncogene, intra- and extra-cellular microenvironmental cues. In the future, we hope that the combined used of xenograft models together with FACS and in-depth genetic analysis may help to identify the critical factors underlying the MLL-rearrangements transformation process and identify therapeutic targets for this disease.
Reference list
1. Vargaftig, J. et al. Frequency of leukemic initiating cells does not depend on the xenotransplantation model used. Leukemia 26, 858–860 (2012).
2. Rongvaux, A. et al. Human hemato-lymphoid system mice: current use and future potential for medicine. Annu. Rev. Immunol. 31, 635–674 (2013). 3. Rizo, A. et al. BMI1 collaborates with BCR-ABL in leukemic transformation of
human CD34+ cells. Blood 116, 4621–30 (2010).
4. Rizo, A. et al. BMI1 collaborates with BCR-ABL in leukemic transformation of human CD34+ cells. Blood 116, 4621–30 (2010).
5. Yang, J. et al. Bmi-1 is a target gene for SALL4 in hematopoietic and leukemic cells. Proc. Natl. Acad. Sci. 104, 10494–10499 (2007).
6. Dierks, C. et al. Expansion of Bcr-Abl-positive leukemic stem cells is
dependent on Hedgehog pathway activation. Cancer Cell 14, 238–49 (2008). 7. Zhao, C. et al. Hedgehog signalling is essential for maintenance of cancer
stem cells in myeloid leukaemia. Nature 458, 776–9 (2009).
8. Wang, X. et al. Sonic hedgehog regulates Bmi1 in human medulloblastoma brain tumor-initiating cells. Oncogene 31, 187–99 (2012).
9. Antonelli, A. et al. Establishing human leukemia xenograft mouse models by implanting human bone marrow-like scaffold-based niches. Blood 128, 2949– 2959 (2016).
10. Kimura, S., Roberts, A. W., Metcalf, D. & Alexander, W. S. Hematopoietic stem cell deficiencies in mice lacking c-Mpl, the receptor for thrombopoietin. Proc. Natl. Acad. Sci. U. S. A. 95, 1195–200 (1998).
11. Arai, F. et al. Tie2/angiopoietin-1 signaling regulates hematopoietic stem cell quiescence in the bone marrow niche. Cell 118, 149–61 (2004).
12. Sacchetti, B. et al. Self-renewing osteoprogenitors in bone marrow sinusoids can organize a hematopoietic microenvironment. Cell 131, 324–36 (2007). 13. de Graaf, C. A. & Metcalf, D. Thrombopoietin and hematopoietic stem cells.
Cell Cycle 10, 1582–9 (2011).
14. Schepers, K. et al. Myeloproliferative neoplasia remodels the endosteal bone marrow niche into a self-reinforcing leukemic niche. Cell Stem Cell 13, 285–99 (2013).
15. Arranz, L. et al. Neuropathy of haematopoietic stem cell niche is essential for myeloproliferative neoplasms. Nature 512, 78–81 (2014).
16. Zuber, J. et al. RNAi screen identifies Brd4 as a therapeutic target in acute myeloid leukaemia. Nature 478, 524–528 (2011).
treatment for MLL-fusion leukaemia. Nature 478, 529–533 (2011). 18. Gilan, O. et al. Functional interdependence of BRD4 and DOT1L in MLL
leukemia. Nat. Struct. Mol. Biol. 23, 673–681 (2016).
19. Lu, C. & Thompson, C. B. Metabolic regulation of epigenetics. Cell Metab. 16, 9–17 (2012).
20. Roberts, A. W. et al. Targeting BCL2 with Venetoclax in Relapsed Chronic Lymphocytic Leukemia. N. Engl. J. Med. 374, 311–322 (2016).
21. Godfrey, L. et al. MLL-AF4 binds directly to a BCL-2 specific enhancer and modulates H3K27 acetylation. Exp. Hematol. 47, 64–75 (2017).
22. Khaw, S. L. et al. Venetoclax responses of pediatric ALL xenografts reveal sensitivity of MLL-rearranged leukemia. Blood 128, 1382–95 (2016).
23. Zarrinkar, P. P. et al. AC220 is a uniquely potent and selective inhibitor of FLT3 for the treatment of acute myeloid leukemia (AML). Blood 114, 2984–2992 (2009).
24. Cooper, T. M. et al. A Phase I Study of Quizartinib Combined with
Chemotherapy in Relapsed Childhood Leukemia: A Therapeutic Advances in Childhood Leukemia & Lymphoma (TACL) Study. Clin. Cancer Res. 22, 4014–4022 (2016).
25. Sontakke, P. et al. Modeling BCR-ABL and MLL-AF9 leukemia in a human bone marrow-like scaffold-based xenograft model. Leukemia 30, 2064–2073 (2016).
26. Reinisch, A. et al. A humanized bone marrow ossicle xenotransplantation model enables improved engraftment of healthy and leukemic human hematopoietic cells. Nat. Med. 22, 812–821 (2016).
27. Chen, Y. et al. Human extramedullary bone marrow in mice: a novel in vivo model of genetically controlled hematopoietic microenvironment. Blood 119, 4971–4980 (2012).
28. Zitvogel, L., Pitt, J. M., Daillère, R., Smyth, M. J. & Kroemer, G. Mouse models in oncoimmunology. Nat. Rev. Cancer 16, 759–773 (2016).
29. Tanaka, S. et al. Development of Mature and Functional Human Myeloid Subsets in Hematopoietic Stem Cell-Engrafted NOD/SCID/IL2r KO Mice. J. Immunol. 188, 6145–6155 (2012).
30. Lang, J. et al. Studies of Lymphocyte Reconstitution in a Humanized Mouse Model Reveal a Requirement of T Cells for Human B Cell Maturation. J. Immunol. 190, 2090–2101 (2013).
31. Nicolini, F. E., Cashman, J. D., Hogge, D. E., Humphries, R. K. & Eaves, C. J. NOD/SCID mice engineered to express human IL-3, GM-CSF and Steel factor constitutively mobilize engrafted human progenitors and compromise human stem cell regeneration. Leukemia 18, 341–347 (2004).
32. Wunderlich, M. et al. AML xenograft efficiency is significantly improved in NOD/SCID-IL2RG mice constitutively expressing human SCF, GM-CSF and IL-3. Leukemia 24, 1785–8 (2010).
33. Rongvaux, A. et al. Development and function of human innate immune cells in a humanized mouse model. Nat. Biotechnol. 32, 364–72 (2014).
34. Lan, P., Tonomura, N., Shimizu, A., Wang, S. & Yang, Y.-G. Reconstitution of a functional human immune system in immunodeficient mice through combined human fetal thymus/liver and CD34+ cell transplantation. Blood 108, 487–492 (2006).
35. Jaiswal, S. et al. Enhanced humoral and HLA-A2-restricted dengue virus-specific T-cell responses in humanized BLT NSG mice. Immunology 136, 334– 43 (2012).
36. Parent, A. V. et al. Generation of Functional Thymic Epithelium from Human Embryonic Stem Cells that Supports Host T Cell Development. Cell Stem Cell 13, 219–229 (2013).
37. Greenblatt, M. B. et al. Graft versus host disease in the bone marrow, liver and thymus humanized mouse model. PLoS One 7, e44664 (2012).
38. Grove, C. S. & Vassiliou, G. S. Acute myeloid leukaemia: a paradigm for the clonal evolution of cancer? Dis. Model. Mech. 7, 941–51 (2014).
39. Ding, L. et al. Clonal evolution in relapsed acute myeloid leukaemia revealed by whole-genome sequencing. Nature 481, 506–510 (2012).
40. Ley, T. J. et al. DNMT3A Mutations in Acute Myeloid Leukemia. N. Engl. J. Med. 363, 2424–2433 (2010).
41. Welch, J. S. et al. The Origin and Evolution of Mutations in Acute Myeloid Leukemia. Cell 150, 264–278 (2012).
42. Klco, J. M. et al. Functional heterogeneity of genetically defined subclones in acute myeloid leukemia. Cancer Cell 25, 379–92 (2014).
43. Krivtsov, A. V & Armstrong, S. A. MLL translocations, histone modifications and leukaemia stem-cell development. Nat. Rev. Cancer 7, 823–33 (2007).
44. Wei, J. et al. Microenvironment Determines Lineage Fate in a Human Model of MLL-AF9 Leukemia. Cancer Cell 13, 483–495 (2008).
45. Lin, S., Luo, R. T., Shrestha, M., Thirman, M. J. & Mulloy, J. C. The full transforming capacity of MLL-Af4 is interlinked with lymphoid lineage commitment. Blood 130, 903–907 (2017).
