Thomas-2017-Biology and relevance of human acu.pdf

Review SeriesLEUKEMIC STEM CELLSBiology and relevance of human acute myeloid leukemia stem cellsDaniel Thomas and Ravindra MajetiDivision of Hematology,Department of Medicine,Cancer Institute and Institute for Stem Cell Biology and Regenerative Medicine,Stanford University Schoolof Medicine,Stanford,CAEvidence of human acute myeloid leuke-mia stem cells(AML LSCs)was firstreported nearly 2 decades ago throughthe identification of rare subpopulationsof engrafting cells in xenotransplantationassays.These AML LSCs were shown toreside at the apex of a cellular hierarchythat initiates and maintains the disease,exhibiting properties of self-renewal,cellcycle quiescence,and chemoresistance.This cancer stem cell model offers anexplanation for chemotherapy resistanceand disease relapse and implies thatapproaches to treatment must eradicateLSCs for cure.More recently,a number ofstudies have both refined and expandedour understanding of LSCs and intrapatientheterogeneity in AML using improvedxenotransplantmodels,genome-scaleanal-yses,and experimental manipulation ofprimary patient cells.Here,we review thesestudies with a focus on the immunopheno-type,biological properties,epigenetics,ge-netics,and clinical associations of humanAML LSCs and discuss critical questionsthatneedtobeaddressedinfutureresearch.(Blood.2017;129(12):1577-1585)IntroductionAcutemyeloidleukemia(AML)isarapidlyprogressinghematopoieticmalignancy,characterized by the accumulation of clonal myeloidprogenitorcellsarrestedintheirabilitytodifferentiateintomaturebloodcells that is accompanied by multilineage cytopenias.1Indeed,it isthe lack of normal cells,including red blood cells,platelets,andgranulocytes,that eventually contributes to morbidity in this disease.While reductions in leukemic blasts can be achieved initially withcytarabine andanthracycline chemotherapyinthemajorityofpatients,long-term outcomes have not improved significantly for over3 decades,with 5-year overall survival rates for patients,60 yearsranging from 35%to 40%and median overall survival of;1 year.1,2ManysamplesfrompatientswithAMLshowevidenceofahierarchicalcellular organization,with a minor fraction of self-renewing leukemiastem cells(LSCs)at the apex of this hierarchy that can maintain thedisease long-term in immunodeficient mice.LSCs are defined as cellsthat are capable of initiating the disease when transplanted intoimmunodeficientanimalsandcanself-renewbygivingrisetoleukemiain serial transplantations and also partially differentiate into non-LSCbulk blasts that resemble the original disease but are unable to self-renew.The earliest conceptual idea of leukemia being organized ina hierarchical manner traces back to studies performed to identifyclonogenic AML progenitors in vitro.3,4Dick and colleagues laterdemonstratedthatsomewhatsimilartonormalhematopoiesis,AMLisorganizedinthishierarchicalfashioninvivo,beingdrivenlong-termbyLSCs.5-7Fromaclinicalperspective,thecancerstemcellmodelimpliesthatin order to eradicate the disease and achieve long-term remissions,treatment courses must eliminate the LSC population.8While thisultimate goal is yet to be realized,detailed characterization of AMLLSCs has demonstrated their properties of self-renewal,relativequiescence,resistance to apoptosis,and increased drug efflux thatlikely render them less susceptible to conventional therapies aimed atthe bulk proliferative disease.9Recently,high-throughput genomesequencing and DNA methylation profiling has increased ourknowledge of the genomic and epigenomic landscape of thisdisease,10,11but merging these data with in vivo biology of LSCsis still in its infancy.The application of genomic methodologies hasalso led to the identification of preleukemic hematopoietic stemcells(HSCs)in AML patients that do not generate leukemia in vivobut carry early competitive driver mutations,typically in epigeneticregulator genes.12-14In this article,we review the immunophenotype,biologicalproperties,epigenetics,genetics,andclinicalassociationsofhuman AML LSCs and highlight themes for further investigation.Past perspectives on leukemia heterogeneityand stem cellsWhile hematopathologists had long noted morphological differencesin leukemic cells both between and within patients,the first in vivoevidence of intraleukemic heterogeneity came directly from AMLpatients injected with tritiated thymidine.15Marked differences in theproliferation kinetics of individual leukemic cells could be distin-guished,anticipating the existence of a rare leukemic population thatcycled very slowly and showed resistance to antiproliferativetherapies.16Unexpectedly,AML cells did not proliferate moreextensivelythannormalhematopoieticcellsinpatients,andaconsistentpopulation of quiescent cells could be distinguished.These observa-tions were later followed by efforts to identify AML progenitorsin vitro3,4,17,18and inspired an interest in conceptualizing AML in ahierarchical organization,akin to what was being observed in normalhematopoiesisatthattime.Isolatedprimarycellsthatcouldgiverisetoleukemic blast colonies(;1%)in semisolid media were termed AMLcolony-forming cells(CFCs)and showed wide variation betweenpatients;however,they did not correlate with clinical outcome andSubmitted 30 September 2016;accepted 22 November 2016.Prepublishedonline as Blood First Edition paper,3 February 2017;DOI 10.1182/blood-2016-10-696054.2017 by The American Society of HematologyBLOOD,23 MARCH 2017xVOLUME 129,NUMBER 121577For personal use only.on December 17,2018.by guest www.bloodjournal.orgFrom exhibited very limited replating efficiency.18Improvements in cultureconditions using normal human bone marrow or mouse fibroblastfeeder cell layers and recombinant human cytokines allowed measure-ment of rarer cells that could give rise to AML CFCs for more than5 weeks in vitro,termed long-term culture-initiating cells(LTC-IC),withafrequency5-to300-foldlowerthanCFCs.19InvivoevidenceforLSCswaslaterdemonstratedbyserialxenotransplantationexperimentsinseverecombinedimmunodeficiency(SCID)mice,6andthenlaterinnonobese diabetic(NOD)/SCID mice,7using fluorescence-activatedcellsorting(FACS)toisolateengraftingsubpopulationsfromanumberof AML samples.These robust but scarce leukemic cells,termedleukemia-initiating cells(LICs),comprised roughly 1 per 1 3 106leukemic blasts and had 2 important properties:(1)they gave rise toleukemic engraftment that could be propagated for multiple serialtransplants(self-renewal),and(2)they produced non-LSC progenyunable to engraft.In the most rigorous definition,cells that meetboth these criteria are termed LSCs,although not all studies in theliteraturethatusethisdesignationperformfunctionalinvivoorserialtransplantation experiments.20Immunophenotype of LSCs in human AMLThe critical step in the identification of human AML LSCs is theestablishmentofaxenotransplantationassaythatfacilitatesengraftmentofpatientsamples.Overtheyears,anumberofimmunodeficientmousestrains have been developed to facilitate engraftment of normal andmalignant human hematopoietic cells,including AML(reviewed inGoyama et al21).However,what is often not apparent from reportedstudies is that many primary AML samples do not engraft at all whentransplanted into these models.Engraftment has historically beendefined as human leukemic chimerism.0.1%in the mouse bonemarrow,a notably low bar compared with clinical disease.Using thiscriterion,a review of historical studies of diagnostic samples of AMLreporting both engrafting and nonengrafting samples found that inaggregate,23 out of 49 samples(47%)engrafted.22-25A recent studyreported results of engraftment of 56 AML samples,and found that45(80%)engrafted in xenotransplantation assays,although it shouldbe noted that more than half of these were from relapsed or refractorypatients.26Recently,a humanized ossicle xenotransplantation modelwas reported in which 13 out of 15 diagnostic AML samples(87%)engrafted to high levels of bone marrow chimerism.27Apart fromengraftment,the frequency of LICs varied greatly between samplesand models,indicating that identification and frequency of LSCs ishighlydependentonthemodelused(Figure1).Indeed,towhatextentimmunological and interspeciesdifferences inspecific xenotransplantmodels havedictated the observed biologicalpropertiesof LSCsisanimportant and open question.CD341AMLThe majority of AML samples(;75%)are positive for expression ofCD34,defined as present on.10%of blasts,and nearly all studies ofAMLLSCshavefocusedonthissubgroup.Intheinitialstudies,LSCswere shown to be positive for expression of CD34 and negative forexpression of CD38 as demonstrated by engraftment in NOD/SCIDmice7and LTC-IC activity in vitro.19,28Further work by multiplelaboratories,22,23,29,30using both intrafemoral and intravenous trans-plantationandmorepermissivemodels,includingNOD/SCID/IL2R-gnull(NSG)mice,showed that for CD341AML,LICs predominantlyresideintheCD341CD382fraction.However,inmorethanhalfofthesamples,LSCs were also present in at least one other subpopulation,usually the CD341CD381fraction and sometimes in the CD342fraction.20To further examine the relationship and plasticity betweenthese subpopulations,one group performed detailed immunopheno-typicanalysisof100CD341samplescoupledwithengraftmentstudiesin 12 cases.Consistent with prior work,they demonstrated thecoexistence of at least 2 distinct CD341LSC populations in thesepatients:a CD341CD382fraction resembling normal lymphoid-primed multipotent progenitors(LMPP-like LSCs;Lin2CD341CD382CD902CD45RA1)and a CD341CD381fraction resemblinggranulocyte-macrophage progenitors(GMP-like LSCs;Lin2CD341CD381CD1231CD45RA1)(Figure 2A).22Interestingly,in 80%of patients studied,both populations coexisted with evidence of ahierarchical relationship in which LMPP-like LSCs gave rise to theGMP-like population,but not the converse(Figure 3).It will beimportant to see if such a hierarchical relationship between these2 LSC populations is consistently recapitulated in more humanizedmousemodels.Finally,in10%to15%ofcases,adominantpopulationresembling multipotent progenitors was observed(MPP-like,Lin2CD341CD382CD902CD45RA2).In this and other studies,theCD341CD382AML populations exhibited a higher LSC frequencycompared with the CD341CD381fraction,consistent with higherself-renewal potential of the more immature cells.29,30Global geneexpression profiling showed that these LSC populations weremolecularly distinct and,in keeping with their immunophenotype,resembled normal progenitors more than normal stem cells.22Thissuggests,but does not prove,that LSCs arise from a progenitor that hasacquiredself-renewalpropertiesratherthanadirectHSCoriginforAML.DetailedcharacterizationofthesurfaceimmunophenotypeofAMLLSCsandmappingthembacktotheirnormalcounterpartshavebeenofgreatinterest,inparticulartheidentificationofcellsurfacemarkersthat19952000200520102015SCID6EngraftingTimelineNK,+8,multipleMLL rearrangedCD34-,NPM155,56Inv1660,61APL30,61AML SubtypeNOD-SCID7NS-B2M55MISTRG60Humanized ossicle30NSG+anti-CD12225NSG-S28NSG32Figure 1.Timeline of human AML engraftmentmodels.Historical timeline is indicated with immunode-ficient mouse models used for AML engraftment.Thepanel below indicates the subtypes of AML exhibitingengraftment in the models along with the relevantreferences.Note that as the immunodeficient mousemodels improved the engraftment of more AML sub-types was observed.APL,acute promyelocytic leuke-mia;NK,normal karyotype;MISTR-G,RAG2-IL2R-gnull with human M-CSF,IL-3,GM-CSF,SIRPA,andTPO;MLL,mixed-lineage leukemia;NSG,NOD-SCID-IL2R-g null;NS-B2M,NOD-SCID-beta2-microglobulinnull;NSG-S,NSG expressing human IL-3,GM-CSF,and stem cell factor.1578THOMAS and MAJETIBLOOD,23 MARCH 2017xVOLUME 129,NUMBER 12For personal use only.on December 17,2018.by guest www.bloodjournal.orgFrom are differentially upregulated on LSCs compared with normal HSCs.A number of such markers have been identified,but thus far,no oneunique marker has been discovered that is universally expressed onCD341CD382LSCs across AML patients,but not on bulk leukemicblasts or on normal hematopoietic stem or progenitor cells(HSPCs).This challenge is partially due to the intrinsic heterogeneity ofAML,both between patients and within an individual sample,but italso arises from the fact that LSCs immunophenotypically resemblecertainnormalhematopoieticprogenitorpopulations.Inaddition,manylymphoidandmyeloidantigensareaberrantlyexpressedinAML(suchas CD7 or CD11b),giving rise to complex leukemia-associatedphenotypes that may can change at the time of relapse.Despite these difficulties,a number of cell surface markershave been identified that are upregulated on CD341CD382LSCscompared with normal CD341CD382HSPCs,including CD123,31CD44,32CD47,33,34TIM3,35,36CD96,37CD99,13,38,39CLL-1,40,41CD32,42CD25,42IL1RAP,43,44GPR56,45and CD93.46However,mostofthesemarkershavenotbeenstudiedatrelapse,sotheirstabilityin this setting is not known.In the case of CD123,the frequency ofCD1231CD341CD382cells consistently increased at relapse.47,48CD123 also cosegregates with FLT3-internal tandem duplication(ITD)mutationpositivecellswithintheCD341CD382population,49suggesting it may be a robust LSC marker in FLT3-ITDmutatedAML.Notably,in only a handful of cases have any of these markersbeen shown to segregate LSC activity between marker-positive andmarker-negative leukemic cells.Antibody and cell therapies directed against LSC antigens.Apart from the refinement and purification of LSCs,these antigensserve as potential targets for immunotherapeutic approaches for thetreatment of AML.Indeed,monoclonal antibodies that bind and/orblock the function of these antigens have demonstrated antileukemicresponsesand,insomecases,anti-LSCactivityinpreclinicalmodelsofhuman AML.50Monoclonal antibodies or biologics against 3 ofthese targets have recently entered clinical trials:CD123(for AMLwith decitabine;NCT02472145),CD47(for AML as monotherapy;NCT02678338,NCT02641002,NCT02663518),and IL1RAP(forCML;NCT02842320).Engineered T cells directed against CD123,including T cells with chimeric antigen receptors,have recently enteredclinical trials(NCT02623582 and NCT02159495),but toxicity againstCD123hinormalHSPCsmaybedifficulttocircumvent.51UnlikehumanB-cellpopulations,inanonmyeloablativesetting,theultimatesuccessofsuch approaches depends on stable expression of leukemia-specificantigensthatarenotpresentonnormalregeneratingmyeloidprogenitors.CD342AMLA minor proportion of AML cases(;25%)lack expression of CD34,defined as present on,10%of blasts,and such cases are enriched forNPM1 mutations52,53and possibly TET2 mutations.54Although trans-plantable LSCs can be found,the majority reside in a CD342compartment,with some detectable in a smaller CD341compartment,indicating a striking difference from the CD341AML cases.29,53,55Rather than one population hierarchically giving rise to the other,bothCD341andCD342LSCpopulationsessentiallyrepresentthesamecells(with aberrant plasticity of CD34 expression)based on transcriptionalprofiling