Emergence of the haematopoietic system in the human embryo and foetus MANUELA TAVIAN,* FERNANDO CORTÉS,* PIERRE CHARBORD,° MARIE-CLAUDE LABASTIE,* BRUNO PÉAULT**Institut d’Embryologie Cellulaire et Moléculaire, CNRS UPR 9064, Nogent-sur-Marne; °Laboratoire d’Étude de l’Hé-matopoïèse, Etablissement de Transfusion Sanguine de Franche-Comté, Besançon, France

Haematologica 1999; 84:(EHA-4 educational book):1-3

Correspondence: Bruno Péault, Unité 506 INSERM, Bâtiment Lavoisier;Groupe Hospitalier Paul Brousse, 12, avenue Paul Vaillant-Couturier,94807 Villejiuf Cedex, France. Tel. international +33-1-45595263 –Fax: international +33.1.45595268 – E-mail: peault@infobiogen.fr

The first haematopoietic cells are observed inthe third week of human development in theextraembryonic yolk sac. Recent observationshave indicated that intraembryonic haematopoiesisoccurs first at one month when numerous clusteredCD34+ Lin– haematopoietic cells have been identi-fied in the ventral aspect of the aorta and vitellineartery. These emerging progenitors express tran-scription factors and growth factor receptors knownto be acting at the earliest stages of haematopoiesis,and display high proliferative potential in culture.Converging results obtained in animal embryos sug-gest that haematopoietic stem cells derived from thepara-aortic mesoderm – in which presumptive endo-thelium and blood-forming activity could be detect-ed as early as 3 weeks in the human embryo by dif-ferential expression of the CD34 and Flk-1/KDR genes– play an essential role in the foundation of definitivehaematopoiesis. Aorta-associated CD34+ cells alsorepresent a unique localised accumulation of primi-tive haematopoietic stem cells worthy of in-depthmolecular characterisation. Differential screening ofa cDNA library has already revealed the expression ofnovel genes in this population, one of which appearsto be involved in the development of both haema-topoietic and endothelial cells. Active blood forma-tion is observed in the liver and bone marrow by theend of the first trimester. Inception of haemato-poiesis occurs earlier in the liver, where CD34+ cellsare detected as early as 30 days, than in the marrow,where haematopoietic cells are not observed beforeweek 11.

Current interest in early human blood cell ontoge-ny may be partly related to the growing use of foetalstem cells for transplantation at postnatal stages,and to emerging cell and/or gene therapies of theblood system in utero, which justify a thorough char-acterisation of embryonic and foetal human haema-topoiesis. In addition, the prenatal haematopoieticsystem is characterised by an outstandingly high rateof progenitor cell expansion, migration and differ-entiation and hence can be seen as a privileged mod-el to identify novel factors involved in these process-

es. In this setting, the recent identification in animalbut also in human embryos of unique intraembryonicsites of haematopoietic stem cell emergence and pro-liferation could be of particular interest.

We shall briefly review here the successive steps ofhuman haematopoietic development, emphasisingthe recent progresses made in our understanding ofthe origin and identity of human embryonic and fetalstem cells.

Primary haematopoiesis in the humanembryo and foetus

As is the case in other mammals, human haema-topoiesis starts outside the embryo, in the yolk sac,then proceeds transiently in the liver before gettingstabilised until adult life in the bone marrow. Only Tlymphocytes are produced in the same tissue atembryonic, foetal and postnatal stages.

The yolk sacIt is at about 18.5 days of development (early head

process) that primitive haematopoietic cells appearinside forming blood vessels in the intermediatemesodermal cell layer of the human yolk sac wall.Studies on human haematopoietic cell emergence atthese early stages are scarce, but our own observa-tions1,2 suggest that the sequence described in animalmodels also applies to the human yolk sac: meso-derm-derived clusters of primitive haematopoieticstem cells – the blood islands – develop in close asso-ciation with the endothelium of emerging blood ves-sels, possibly from a common ancestor cell or hae-mangioblast. The coexpression of the CD34 surfacemolecule by haematopoietic precursor cells andendothelial cells can be traced back to these initialstages, which may support the hypothesis of theircommon origin.1,2 Migliaccio et al.3 described severalgenerations of clonogenic progenitors in the humanyolk sac from 4.5 weeks of development, includingpluripotential (CFU-GEMM), granulomonocytic(CFU-GM) and erythroblastic progenitors (BFU-Eand CFU-E). The human yolk sac starts regressing atabout 45-50 days post-ovulation and virtually allclonogenic progenitors have disappeared from thattissue by week six.

The liverThe liver emerges during the 4th week of develop-

ment when the hepatic bud, an endodermal out-

Educational Session 1Chairman: W.E. Fibbe

Biology of normal and neoplastic progenitor cells

growth of the foregut, invades the adjacent mesoder-mal septum transversum. These two tissues con-tribute hepatocyte cords and vascular sinuses, respec-tively. We have detected CD45+CD34– haematopoi-etic cells from day 23 of development in the liveranlage while the first CD34+ haematopoietic progen-itors could be recognised on day 30.1 In vitro colony-forming cells, i.e. BFU-E, CFU-GM and, slightly later,CFU-E have been indeed detected at 4.5-5 weeks inthe liver rudiment, where their frequency then increas-es dramatically, paralleling their sharp decline in theyolk sac.3 At the end of the first trimester, andonwards, more primitive progenitors – CFU-GEMMand HPP-CFC – have also been detected in the liver.Earlier studies, confirmed by more recent immuno-histochemical approaches, have documented theextensive erythro-myeloid haematopoiesis that takesplace extravascularly in the human embryonic andfoetal liver, and have stressed the prominence of ery-thropoiesis therein (reviewed in ref. #4). Othermyeloid cells present in the haematopoietic liver aregranulocytes, macrophages and rare megakaryocytes.B-lymphopoiesis has been traced in the liver fromabout 9 weeks of gestation by detection of surfaceIgM+ cells.

The bone marrowA cartilaginous presumptive skeleton is present in

the 6-8-week human embryo. Bone rudiments arethen surrounded by a dense network of CD34+ capil-laries, by CD68+ monocytes and by osteoblast pre-cursors which all invade the diaphyseal cartilage at8.5-9 weeks. Incoming macrophages rapidly digestthe cartilage, leaving only intact small islets of chon-drocytes that soon become surrounded by osteo-blasts, from which ossification proceeds in a typical-ly endochondral manner. In-between ossifying tra-beculae, large vascular sinuses develop leading to thecompletion, at about 10 weeks, of bone marrow cav-ities.5 Marrow haematopoiesis starts during the 11thweek of development in specialised mesodermalstructures or primary logettes, constituted by a loosenetwork of mesenchymal cells supported by dense fib-rillar material and surrounding a central artery, insidewhich CD15+ granulocytes appear first, closely fol-lowed by erythroid cells. Haematopoiesis then devel-ops dramatically in rapidly enlarging logettes which byweek 15 are densely packed with cells of the erythroidand granulocytic series.5

Haematopoietic stem cell emergence inearly human development

As mentioned above, the emergence of the CD34cell surface antigen in ontogeny seems to be contem-porary with the earliest commitment of mesodermalcells in the 3-week yolk sac to haematopoiesis andvasculogenesis. CD34 expression is then consistentlydetected on the surface of haematogenous cells in theliver and bone marrow.

Concepts on the filiation of the stem cells thatfound definitive haematopoiesis have changed in thepast few years with the demonstration that, in ani-mals, these emerge inside the embryo, and not in theyolk sac as previously believed. In mice and birds theoriginal blood-forming territory develops intrinsicallyin the para-aortic splanchnopleural mesoderm con-stituting the presumptive aorta-gonad-mesonephros(AGM) region of the embryo and contributes, at pre-liver stages, multilineage haematopoietic progenitorsand eventually long-term reconstituting true haema-topoietic stem cells. The existence in the humanembryo of an equivalent site of haematopoietic stemcell generation has been suggested by the identifica-tion, at 4-6 weeks of development, of numerous clus-tered CD34+ haematopoietic cells on the ventralendothelium of the aorta and vitelline artery.1,6 Thesecells express surface antigens that typify early bloodcell progenitors, being CD45+, CD34++, CD31+,CD43+, CD44+, CD164+, but display no CD38 or lin-eage-specific markers. In situ hybridisation on embryosections and screening of cDNA libraries preparedfrom these sorted aorta-adherent progenitors havealso revealed that they express genes known to beassociated with the early steps of haematopoieticdevelopment, such as Tal1/SCL, c-myb, GATA-2,GATA-3, flk-1/VGEFR2 and c-kit.7 When directlyassayed in methylcellulose, human intraembryonicaorta-associated CD34+ cells exhibited negligibleclonogenic potential. In contrast, following a 4-10-day co-culture on murine bone marrow stromal cells(MS-5 cell line), they generated about six times moreprogenitors, which yielded large multilineage coloniesin methylcellulose, than the liver rudiment.6 Of note,the para-aortic splanchnopleura – but no other intra-embryonic tissue – exhibited dramatic haematopoieticpotential in culture as early as day 23 of development,i.e. several days before CD34+ stem cells can actuallybe identified on the aortic wall. This result, as well asprovocative semi-thin section histology pictures, sug-gest that haematopoietic stem cells emerge frommesoderm in that territory, and not merely migratethere from another location.6

Differential screening of a cDNA library built fromsorted embryonic aorta-associated CD34+ cells withprobes prepared from embryonic liver and foetal bonemarrow CD34+ stem cells is in progress. Several dif-ferentially expressed genes have already been found,one of which encodes a serine-threonine kinase which,interestingly, is co-expressed in all developing endo-thelial and haematopoietic stem cells.7 This arguesfor the existence of haemangioblasts, i.e. common prog-enitors for vascular and blood cells.

We have reported that KDR, the human homologueof VEGFR2/Flk1, is strongly expressed in the humanembryo by endothelial cells but barely detectable inthe first haematopoietic stem cells arising in the wallof the aorta.7 Conversely, the CD34 protein is detect-ed from early stages at the surface of both cell types.

M. Tavian et al.2

We took advantage of this differential expression pat-tern to trace the emergence of putative human hae-mangioblasts and their segregation into endothelialand haematopoietic lineages. A population of KDR+CD34– mesoderm cells emerges in early-somatichuman embryos, by the beginning of the 4th week ofgestation. During blood vessel formation these KDR+CD34– cells gradually co-express increasing levels ofCD34. Simultaneously, in the yolk sac, the solehaematopoietic tissue at that stage, most haemato-poietic progenitors exhibit a KDR– CD34+ phenotype.Remarkably, as development proceeds, a KDR+ CD34–compartment persists in the splanchnopleura untiljust prior to the emergence of aorta-associatedhaematopoietic cell clusters. This cell compartmentmay include the putative haemangioblastic precursorof human haematopoietic and endothelial lineages.8

ConclusionsThe localised accumulations of haematopoietic

stem cells observed along intra-embryonic artery wallsand in the yolk sac in the third-fifth weeks of gestationprobably reflect phases of progenitor cell emergenceand amplification that do not occur any later in devel-opment. These unique, transient haematogenous ter-ritories are presently being actively studied: decipher-ing the molecular control of mesoderm commitmenttowards blood cell lineages is of prime interest fordevelopmental biologists, while haematologists sus-pect that novel factors are to be identified at theseearly stages that could be used to manipulate the sur-vival/renewal/proliferation of adult haematopoieticcells. Experiments to that end are based on in vitro cellor organ culture of the mouse or human yolk sac,para-aortic splanchnopleura and derived AGM tis-sues, the ability of which to drive the expansion anddifferentiation of co-cultured stem cells is tested inhomospecific or xenogeneic combinations. Muchemphasis is also being put on subtractive cloning ofnovel genes whose function can be tested in vitro and

in living models of overexpression and inactivationsuch as mouse ES cells, transgenic mice and zebrafish.

A better understanding of the microenvironment ofthese different haematopoietic tissues is of theoreti-cal and practical interest in order to unravel similari-ties with or differences from adult bone marrow stro-ma in terms of critical cells and mediators;8 progressin this domain should be obtained via the generationof immortalised stromal cell lines.

References

1. Tavian M, Hallais MF, Péault B. Emergence ofintraembryonic hematopoietic precursors in the pre-liver human embryo. Development 1999; 126:793-803.

2. Cortés F, Debacker C, Péault B, Labastie MC. Differ-ential expression of KDR/VEGFR-2 and CD34 definesdistinct stages of endothelial and hematopoietic devel-opment in early human embryos. Mech Dev 1999; inpress.

3. Migliaccio G, Migliaccio AR, Petti S, et al. Humanembryonic hemopoiesis. Kinetics of progenitors andprecursors underlying the yolk sac->liver transition. JClin Invest 1986; 78:51-60.

4. Péault B, Touraine JL, Charbord P. Haematopoieticstem cell emergence and development in the humanembryo and fetus; perspectives for blood cell therapiesin utero. Semin Neonatol 1999; in press.

5. Charbord P, Tavian M, Humeau L and Péault B. Ear-ly ontogeny of the human marrow from long bones:an immunohistochemical study of hematopoiesis andits microenvironment. Blood 1996; 87:4109-19.

6. Tavian M, Coulombel L, Luton D, San Clemente H,Dieterlen-Lièvre F, Péault B. Aorta-associated CD34+hematopoietic cells in the early human embryo. Blood1996; 87:67-72.

7. Labastie MC, Cortés F, Roméo PH, Dulac C, Péault B.Molecular identity of hematopoietic precursor cellsemerging in the human embryo. Blood 1998; 92:3624-35.

8. Cortés F, Deschazeaux F, Uchida N, et al. HCA, animmunoglobulin-like adhesion molecule present onthe earliest human hematopoietic precursor cells, isalso expressed by stromal cell in blood-forming tis-sues. Blood 1999; 93:826-37.

Session #1 – Biology of normal and neoplastic progenitor cells 3

Characterisation and biology of normal human haematopoieticstem cells ROB E. PLOEMACHERDepartment of Haematology, Erasmus University, Rotterdam, The Netherlands

Correspondence: Dr. Rob E. Ploemacher, Dept. of Haematology, Eras-mus University, P.O. Box 1738, 3000 DR, Rotterdam, The Netherlands.Phone: +31-10-408.7603; Fax: +31-10-436.2315; Email: ploemach-er@hema.fgg.eur.nl.

ate life-long blood cells of a multiplicity of lineages.In the laboratory, these properties are often project-ed on individual HSC and the assays used are tunedto reveal one or more of their abilities. From physicalsorting of various haemopoietic grafts it has becomeapparent that HSC assays should at least test formultilineage cell production over extended periodsof time (i.e. months rather than weeks), in vitro butpreferentially in vivo. Preferentially such assays shouldbe able to allow HSC frequency analysis as well asassessment of the proliferative ability of (individual)HSC subsets. Assays that allow such a read-outinclude (a) analysis of long-term stroma-supportedhaemopoiesis and generation of progenitors (long-term culture, LTC) and frequency analysis usingminiaturised LTC’s in limiting dilution (cobblestonearea forming cell or CAFC assay; LTC initiating cell orLTC-IC assay);2,3 (b) assessment of human multi-lin-eage haemopoiesis in the bone marrow of immun-odeficient mice, e.g. in NOD/SCID mice (Scid repop-ulating cell or SRC assay),4 or (c) in sheep, where thehuman cells are infused during the pre-immune foetalstage (human-foetal sheep model).5 The LTC, whichis performed in flask cultures, gives an insight intothe capacity of a graft to produce progenitor cellswhile it does not reveal how many HSC and progen-itors contribute to it. The latter issue is highly relevantwhen differences are anticipated in the progenitorcell generating capacity of HSC e.g. on the basis ofdisease, chemotherapy or in vitro methods for physi-cal or chemical purging or selection. As frequencyanalysis requires a limiting dilution set-up of the assayit will be evident that animal assays are unfit to enu-merate HSC on a routine basis, but nevertheless,probably provide the investigator with a more rele-vant estimation of the total graft ability than do in vit-ro assays. Typical drawbacks of all these assays arethat they take a long time to complete and are oftentedious to quantify.

Extensive studies using physically sorted murine BMhave allowed regression analysis on the applicabilityof the murine CAFC assay as an in vitro equivalent ofassays for a series of HSC subsets in vivo. These stud-ies have indicated that, both in vivo and in vitro, early-developing, short-lived clones are initiated by thetransient repopulating spleen colony-forming cells(CFU-S day-12), whereas later developing and morepermanent clones are descendants of more primitive,long-term repopulating HSC, that induce stable chi-

Haematopoiesis is a life-long process respon-sible for replenishing both haematopoieticprogenitor cells and mature blood cells froma pool of pluripotent, long-term reconstitutinghaemopoietic stem cells (HSC). The daily turnover ina normal adult of approximately 1012 blood cells istightly regulated, involving, in part, a complex inter-action between soluble and membrane-bound stim-ulatory and inhibitory cytokines and their corre-sponding receptors. The molecular cloning of thesehaemopoietic growth factors and their receptors hasbeen instrumental in the partial delineation of thepathways that lead from a single HSC to the variousterminally differentiated cells in the haemopoieticsystem.

The HSC compartment is distributed through theseparated bone marrow (BM) locations and HSC areknown to traverse in low numbers via the peripheralblood (PB) between these compartments. The HSCcompartment is heterogeneous in various aspects,possibly as a result of different mitotic histories ofthe individual HSC caused by stochastic mechanismsthat regulate cycling activity in a predominantly qui-escent HSC population. In analogy to the situation inthe mouse, the human HSC compartment probablyrepresents a hierarchy of primitive cells on the basisof decreasing ability to generate new HSC, decreas-ing pluripotentiality and proliferative potential, andincreasing turn-over rate. In a transplant setting thisheterogeneity may be reflected in the different timeperiods that individual HSC clones contribute tolong-term reconstitution of a conditioned host, irre-spectively of whether this reconstitution includes allhaemopoietic lineages, and whether these lineagescan be detected at the same moment in time.1

HSC assays HSC cannot be detected in bone marrow smears

because they occur with low frequencies and have nounique features that allows their enumeration.Depending on the HSC assay used, HSC frequenciesin normal BM are estimated to be between 1 in 104to 106 cells. The HSC compartment is typically char-acterised by its ability to self-replicate and to gener-

Haematologica 1999; 84:(EHA-4 educational book):4-7Educational Session 1

Chairman: W.E. FibbeBiology of normal and

neoplastic progenitor cells

maerism for more than a year. The CAFC assay, there-fore, seems fit to quantify HSC subsets with differentrepopulation potentials.

In order to facilitate more rapid enumeration ofHSC and progenitor cells other methods have beenand are being investigated, including colony forma-tion and phenotypic analysis. Although stem cells mayform (multi-lineage) colonies in vitro, they are greatlyoutnumbered by the more mature progenitor cellswhich therefore renders colony formation unsuitablefor HSC enumeration in a graft. Determination of thereplating efficiency of individual colonies may givemore information on the proliferative ability ofcolony-forming cells, but, again requires a large timeexpenditure.

Data from many laboratory and clinical investiga-tions indicate that almost all lymphohaemopoieticstem cells and all their progenitor cells are containedin approximately 1% of human BM mononuclear cellsthat express the surface marker CD34. Because stemcells are a rare cell type in the CD34+ cell population,investigators have subdivided the CD34+ cell popula-tion in order to enrich stem cells further. The CD34+/CD38– cell subset comprises less than 10% of humanCD34+ adult BM cells (equivalent to < 0.1% of mono-nuclear cells), lacks lineage (i.e. it is Lin–) antigens,contains cells with in vitro replating and long-termhaemopoietic culture capacity, and is predicted to behighly enriched for in vivo repopulating stem cells. BM-derived CD34+/CD38– cells also contain SRC and gen-erate long-term, multilineage human haemopoiesis inthe human-foetal sheep in vivo model, whereas trans-fer of CD34+/ CD38+ cells generates only short-termhuman blood cells.

Another interesting glycoprotein antigen is AC133,which is selectively expressed on CD34bright HSC andprogenitor cells derived from human foetal liver andBM, and PB. AC133-selected cells, which also includesome CD34– cells, engraft successfully in a foetalsheep transplantation model, and human cells har-vested from chimaeric foetal sheep BM have beenshown to engraft secondary recipients successfully,providing evidence for the long-term repopulatingpotential of AC133+ cells.

It should be noted that enumeration of HSC usingphenotypic analysis may, or may not, generate simi-lar data as those from in vitro or in vivo functionalassays. Firstly, it would require great skill, superbprobes and instrumentation to allow rare event detec-tion for enumeration of HSC in unseparated graftsusing the presently available technology. Thus,although we may find that the bulk of HSC activity iscontained in a CD34+/CD38– fraction, not everyCD34+/CD38– is necessarily a HSC. Moreover, someHSC that do not meet these criteria may be over-looked as was demonstrated by the recent findingthat perhaps the most potent, long-term repopulat-ing HSC do not express the CD34+ antigen but areLin–CD34–.4,6 A second issue with phenotypic analysis

is that the markers that we use may be promiscuous,or may not relate to function following manipulationof HSC. Ex vivo HSC manipulation protocols, e.g. forexpansion of HSC from umbilical cord blood cells(UCB), somatic gene transfer or tumour cell purging,include either stimulation with a variety of cytokines,or use of selection procedures, or both. It has beenfrequently observed that large numbers of CD34+ cellsand progenitors can be generated ex vivo, however,without a substantial maintenance or increase of HSCas tested by functional assays (e.g. SRC, LTC-IC,CAFC). We have recently found over 105-fold expan-sion of the CD34+/ CD38– cells in 11 week cultures ofhuman umbilical cord blood CD34+ cells, whereasLTC-IC/CAFC were only modestly (1-100-fold)expanded and although SRC were initially expanded,no SRC activity was detected at 11 weeks of culture.While the freshly uncultured CD34+/ CD38– cells wereLin–, it could be shown that most of the cultured andresorted CD34+/CD38– cells, although still containingall CAFC activity, expressed lineage markers in thatfraction. These data show that one should beextremely careful in extrapolating a function-pheno-type relationship as defined in fresh specimen to oth-er circumstances.

HSC quiescence, chemosensitivity andradiosensitivity

Virtually none of the stem cells from the mobilisedperipheral blood (MPB) and BM is cycling. Indeed,LTC-IC activity is higher in CD34+ cells isolated in G0than in those residing in G1. G0 phase HSC showlonger persistence of CD34 expression in suspensionculture than do G1 phase HSC, and maintain in vitrohaemopoiesis for longer periods. The deep quiescenceof most HSC explains why they are refractive to a sin-gle treatment with cycle-specific drugs (e.g. 5-fluo-rouracil, cytosine arabinoside, hydroxyurea, vincri-stine), whereas repeated chemotherapy may lead toHSC activation and their recruitment into a drug-sen-sitive state. In contrast, some alkylating drugs, e.g.busulphan, preferentially kill the most primitive CAFC(week 6) while sparing most of the transiently repop-ulating CAFC (week 2).7 Multiple rounds of chemo-therapy may, therefore, not only result in decreasednumbers of HSC in the BM and PB, but may also leadto loss of the ability of the individual HSC to produceprogeny.

Ample evidence in the murine model has shownthat the most primitive, long-term repopulating HSCand CAFC-week 6 have the lowest sensitivity for ion-ising radiation (gamma, X-ray and fission neutrons)and display unexpected high sub-lethal damage (SLD)repair. In contrast, transiently repopulating HSC,CAFC-week 2 and cells that form spleen colonies inirradiated recipients (CFU-S) are highly sensitive andlack SLD repair.8

The regulated localisation, conservation, commit-ment and terminal differentiation of undifferentiated

5Session #1 – Biology of normal and neoplastic progenitor cells

HSC is believed to occur in niches or local area net-works in BM stromal microenvironments. This resultsin preservation of the stem cell pool while permittingcontrolled cell proliferation and differentiation. Theexact nature of such niches is only slowly emergingfrom many experiments. It is clear that complex inter-actions between stromal cells and haemopoietic stemand progenitor cells involve cell adhesion molecules,and extracellular matrix molecules that may bind andpresent elaborated cytokines and chemokines. Stro-mal cells display membrane bound cytokines (e.g.stem cell factor) and their receptors (e.g. c-kit), andspecific heparan sulphate proteoglycans containinghigh 6-O-sulphation on the glucosamine residues.These interactions lead to specific docking of haemo-poietic cells where they co-localise with regulatorymolecules in an as yet unsufficiently characterised con-text. Specific niches might exist that induce conserva-tion and maintenance of primitive progenitors andother niches that promote proliferation and differen-tiation, depending on the specific cytokines and matrixcomponents present within.

In vitro, primitive HSC require combined stimula-tion by multiple cytokines for growth, but somecytokines selectively promote viability rather thangrowth when acting individually. These cytokines, e.g.Interleukin-3 (IL-3), the ligands for c-kit (KL) and flt3(FL), but especially Tpo, exert direct and selective via-bility-promoting effects on a small fraction of CD34+CD38–, but not CD34+CD38+, human bone marrowprogenitor cells at the single-cell level. Tpo mRNA isexpressed in many stromal cell cultures, while there isvariable expression of KL and FL.

Mobilisation and homing of HSCAlthough in the steady state low HSC numbers can

be detected in the PB, numerous agents, varying fromlipopolysachharides to specific cytokines, are able toincrease the numbers of HSC in the blood dramati-cally. The probable underlying drive of the haemo-poietic system is (a threatening) depletion of thebody’s HSC reserve due to infection, blood loss ortreatment with antineoplastic agents. It is clinicallyuseful that repeated infusions of some cytokines (e.g.G-CSF), in combination or not with chemotherapy,mobilise large numbers of quiescent HSC into theblood that can be harvested by leukapheresis andused for autologous or allogeneic transplantation.This method is an advantage for the stem cell donorswho would otherwise have to undergo anaesthesiaand repeated bone marrow punctures.

Following their transplantation, the HSC have tofind the BM niches that guarantee their life-long reg-ulated preservation and outgrowth according to thebody’s demands. Homing of transplanted HSCs inthe recipient BM cavity is thus a critical step in theestablishment of long term haemopoiesis after BMT,as only the cells that home to the marrow in a murinemodel are capable of reconstituting lethally irradiat-

ed secondary hosts long-term. Quiescent HSC, espe-cially those in G0, have been demonstrated to showmore effective seeding in the BM than do G1 or cyclingHSC. What is more, using PKH26-labelled murinecells that were enriched for either transient or long-term in vivo repopulation suggestive evidence wasfound that even at 48 hours after transplant long-term repopulating cells are still quiescent in the bonemarrow (BM). Short-term ex vivo cycle progression ofHSC, in the absence of cell division, appears to reducethe seeding efficiency and long-term engraftmentcapacity of both human and murine HSC.9

From murine studies it appears that only about 1 of4 infused HSC homes to the total BM of a condi-tioned recipient, although other studies have sug-gested a 100 percent seeding efficiency. The homingof human CAFC in the NOD/SCID model is extreme-ly low and less than 1 percent of infused human BM-derived HSC can be recovered 24 hrs after injectionfrom the total recipient’s BM (Van Hennik et al., unpub-lished results).

A wide variety of adhesion molecules and other lig-ands that mediate cell-to-matrix and cell-to-cell inter-actions have been implicated in HSC adherence to vas-cular endothelium in the BM and subsequent trans-migration of haemopoietic progenitor cells across it.Specific HSC homing is probably a complex multistepprocess of rolling, crawling and nesting of the HSCs.There is evidence that a family of selectins (L, P and E)can mediate initial tethering, rolling and subsequentadhesion of HSCs to endothelial cells. In fact, all of thedifferent classes of adhesion molecules appear to playroles in anchoring HSCs within the BM or the promo-tion of differentiation. Intercellular adhesion molecule(ICAM-1) in the Ig family, very late antigen 4 (VLA-4),an integrin, L-selectin and CD44 are examples of suchimportant molecules. Another important considera-tion is the functional activity of these adhesion recep-tors. The integrins can be activated by different cyto-kines, including GM-CSF, IL-3 and KL. The chemokinestromal cell-derived factor-1 (SDF-1), too, was foundto be critical for bone marrow engraftment. SDF-1binds to its receptor CXCR4, which is expressed onmany cell types including some CD34+CD38– cells.SDF-1 attracts CD34+CXCR4+ HSCs and its importantrole in homing is illustrated by the absence of haemo-poiesis in the bone marrow of mice that lack SDF-1 ordo not express CXCR4. Recently, Peled et al.10 demon-strated that KL and IL-6 induce CXCR4 expression onhuman CD34+ cells. CXCR4 expression potentiatesmigration to SDF-1 and engraftment in primary andsecondary transplanted NOD/SCID mice. Moreover,anti CXCR4 antibody completely abrogated stem cellengraftment in this model.

The use of HSC in clinicsPrimitive HSC and progenitors from BM, MPB and

recently also UCB are targets for high-dose chemo-therapy or radiotherapy. With the discovery of the

R. E. Ploemacher6

Session #1 – Biology of normal and neoplastic progenitor cells 7

currently known cytokines and chemokines and theimproved definition of culture ingredients, liquid cul-tures of unmanipulated or physically sorted humanBM, MPB, UBC and foetal liver have been and will beincreasingly used in an attempt to expand repopulat-ing HSC, purge malignant cells and permit somaticgene therapy. Other potential clinical applications ofthese ex vivo graft manipulations include T-cell deple-tion for allogeneic HSC cell transplantation, adoptiveimmunotherapy via T-lymphocytes that are grownand educated in culture, and haemopoietic supportfor haemopoietically compromised patients. Stillmore extensive study is required in these areas to over-come issues such as loss of repopulating stem cellsdue to manipulation, and inefficient gene transfer andexpression in human HSC and their progeny.

References

1. Lemischka IR. What we have learned from retroviralmarking of hematopoietic stem cells. In: Muller-Sieburg C, Torok-Storb B, Visser J, Storb R, eds. Cur-rent topics in microbiology and immunology. Hema-topoietic stem cells, vol. 177. New York: Springer Ver-lag; 1992. p. 59-71.

2. Sutherland HJ, Lansdorp PM, Henkelman DH, EavesAC, Eaves CJ. Functional characterization of individ-ual human hematopoietic stem cells cultured at lim-iting dilution on supportive marrow stromal layers.Proc Natl Acad Sci USA 1990; 87:3584-8.

3. Breems DA, Blokland EAW, Neben S, Ploemacher RE.

Frequency analysis of human primitive haematopoi-etic stem cell subsets using a cobblestone area form-ing cell assay. Leukemia 1994;8:1095-104.

4. Bhatia M, Bonnet D, Murdoch B, Gan OI, Dick JE. Anewly discovered class of human hematopoietic cellswith SCID-repopulating activity. Nat Med 1998; 4:1038-45.

5. Civin CI, Almeida-Porada G, Lee MJ, Olweus J, Ter-stappen LW, Zanjani ED. Sustained, retransplantable,multilineage engraftment of highly purified adulthuman bone marrow stem cells in vivo. Blood 1996;88:4102-9.

6. Zanjani ED, Almeida-Porada G, Livingston AG, FlakeAW, Ogawa M. Human bone marrow CD34- cellsengraft in vivo and undergo multilineage expressionthat includes giving rise to CD34+ cells. Exp Hematol1998; 26:353-60.

7. Down JD, Ploemacher RE. Transient and permanentengraftment potential of murine hemopoietic stemcell subsets: differential effects of host conditioningwith gamma radiation and cytotoxic drugs. ExpHematol 1993; 2:913-21.

8. Down JD, Boudewijn A, Van Os R, Thames HD,Ploemacher RE. Variations in radiation dose survivalamong different bone marrow hemopoietic cell sub-sets following fractionated irradiation. Blood 1995;86:122-7.

9. Van der Loo JCM, Ploemacher RE. Marrow and spleenseeding efficiencies of all murine hematopoietic stemcell subsets are decreased by preincubation withhematopoietic growth factors. Blood 1995; 85:2598-606.

10. Peled A, Petit I, Kollet O, et al. Dependence of humanstem cell engraftment and repopulation of NOD/SCIDmice on CXCR4. Science 1999; 283:845-8.

Gene regulation in normal and leukaemic progenitor/stem cellsRUGGERO DE MARIA,°+ FRANCESCO GRIGNANI,+^ UGO TESTA,+ MAURO VALTIERI,°+

BENEDIKT L. ZIEGLER,* CESARE PESCHLE°+°Kimmel Cancer Institute, T. Jefferson University, Philadelphia, PA, USA; + Dept. of Haematology and Oncology, IstitutoSuperiore di Sanità, Rome, Italy; * Dept. of Haematology and Oncology, University of Tübingen, Tübingen, Germany; ^Istituto di Medicina Interna e Scienze Oncologiche, University of Perugia, Italy

Correspondence: C. Peschle, M.D., Kimmel Cancer Institute, T. Jeffer-son University, B.L.S.B., Room # 902, 233 South 10th Street, Philadel-phia, PA 19107-5541, USA. Tel. international +1-215-5031763; fax.international +1-215-9231098 – e-mail: Cesare.Peschle@mail.tju.edu.

humans and flk-1 in mice) in murine embryonichaematoangiogenesis. Targeted gene disruption stud-ies indicate that flk-1 is required for initiation of prim-itive/definitive haematolymphopoiesis and vasculo-genesis:5 this suggests a role for flk1 in the generationof haemoangioblasts, i.e., hypothetical stem cellsbipotent for haematolymphopoietic and endotheliallineages.

We observed that in human post-natal haemato-poietic tissues [CB, BM, normal or mobilised periph-eral blood (PB, MPB)] CD34+ cells comprise 0.1-0.5%cells expressing KDR. Pluripotent HSCs repopulatingthe haematolymphopoietic tissues (assayed in bothNOD-SCID mice and foetal sheep xenografts) andputative HSCs [assayed in 12-wk extended Dexter typelong-term culture (LTC) as LTC initiating cells (LTC-ICs)6 or cobblestone area forming cells (CAFCs)7], arerestricted to and highly enriched in the CD34+KDR+subset. Conversely, oligo-unipotent HPCs with noself-renewal capacity are restricted to and highly puri-fied in the CD34+KDR– cell fraction. Based on limit-ing dilution analysis, the HSC frequency inCD34+KDR+ cells is 22% in BM by NOD-SCID miceassay and 25-42% in BM, PB, CB by 12-wk LTC assay.The latter enrichment values rises to 53-63% in LTCsupplemented with VEGF, thus suggesting a func-tional role for the VEGF/KDR system in HSCs: thepurification index rises yet further to > 95% in theCD34+KDR+ cell subfraction resistant to prolongedGF starvation in culture.

These results indicate that KDR is a novel function-al marker defining pluripotent repopulating HSCs anddistinguishing them from oligo-unipotent HPCs: thesefindings pave the way to characterisation and func-tional manipulation of HSCs/HSC subsets, as well asinnovative approaches for HSC clinical utilisation.

Unicellular-unilineage erythropoieticcultures: molecular analysis of regulatory gene expression at siblingcell level

In vitro studies on haematopoietic control mecha-nisms have been hampered by the heterogeneity of theanalysed cell populations, i.e., lack of lineage speci-ficity and developmental stage homogeneity of prog-enitor/precursor cells growing in culture. We devel-oped unicellular culture systems for unilineage differ-

New approaches for analysis of generegulation in normal and leukaemichaematolymphopoietic progenitor/stem cells (HPCs/HSCs)

While recent studies provided insight into gene reg-ulation in normal and leukaemic HPCs/HSCs newapproaches and model systems may be required tofurther our understanding in diverse key areas. Specif-ically, (i) HPCs have been stringently purified andextensively characterised, but isolation and pheno-typing of HSCs is still unsatisfactory; (ii) novelmethodology is required for analysis of gene regula-tion at single HPC/HSC level; (iii) delineation of neg-ative regulatory mechanisms for HPCs/HSCs is stillunsatisfactory: particularly, the role of apoptoticmechanisms in normal haematopoiesis is uncertain;(iv) finally, while leukaemogenetic models have beenestablished in immortalised cell lines and transgenicanimals, an in vitro system for leukaemic transforma-tion of normal HPCs/HSCs is still unavailable. Thisreport reviews recent advances addressing theseaspects.

The KDR receptor allows isolation andcharacterisation of normal HSCs

The haematolymphopoietic hierarchy is defined byfunctional assays. Pluripotent HSCs, endowed withextensive self-renewal capacity, are assayed in vivo onthe basis of their capacity to repopulate the haema-tolymphopoietic system,1 i.e., to xenograft irradiatedNOD-SCID mice2 and pre-immune sheep foetuses.3

The major hurdle in HSC studies has been the lackof a specific positive marker, comparable to CD34 forearly haematopoietic precursors:4 this hamperedpurification, characterisation and utilisation of thisextremely rare cell population. Indeed, HSCs are usu-ally enriched in the CD34+/CD38– fraction, whichcomprises only 0.1-0.2% repopulating HSCs in adultbone marrow (BM) and cord blood (CB).1

While HSC identification is still elusive, recentobservations have suggested a role for vascular endo-thelial growth factor receptor 2 (VEGFR2, KDR in

Haematologica 1999; 84:(EHA-4 educational book):8-10Educational Session 1

Chairman: W.E. FibbeBiology of normal and

neoplastic progenitor cells

entiation of purified HPCs followed by daughter cellanalysis at cellular and molecular levels.8 In the culturesystem reported here (i) the GF stimulus induces CBHPCs to proliferate and differentiate/mature exclu-sively along the erythroid lineage;9 (ii) this erythropoi-etic wave is characterised by < 4% apoptotic cells; (iii)asymmetric divisions are virtually absent, i.e., non-responsive HPCs with no erythropoietic potential areforced into apoptosis; (iv) the system is cell divisioncontrolled, i.e., the number of divisions performed byeach cell is monitored. Single-cell reverse transcriptase(RT)-PCR analysis was applied to this culture system toinvestigate gene expression of diverse receptors, mark-ers of differentiation and transcription factors (EKLF,GATA-1, GATA-2, NF-E2, PU.1, SCL/Tal1) at discretestages of erythropoietic development. Freshly isolatedCD34+ cells expressed CD34, c-kit, PU.1 and GATA-2but did not express CD36, erythropoietin receptor(EpoR), SCL/Tal1, EKLF, NF-E2 or GATA-1 and glyo-phorin A (GPA). In early to intermediate stages of ery-throid differentiation, we monitored the induction ofCD36, Tal1, EKLF, NF-E2 and GATA-1, which preced-ed expression of EpoR. In late stages of erythroid mat-uration, GPA was upregulated, while CD34, c-kit, PU.1and GATA-2 were barely or not detected.

In addition, competitive single-cell RT-PCR wasused to assay CD34 mRNA transcripts in siblingCD34+CD38– cells differentiating in unilineage ery-throid cultures: this analysis allowed quantification ofthe gradual downmodulation of CD34 mRNA fromHPCs through their differentiating erythroid progeny.

It is concluded that this novel culture system, cou-pled with controlled single-cell RT-PCR analysis, mayeliminate the ambiguities intrinsic to molecular stud-ies on heterogeneous populations of haematopoieticprogenitors/precursors growing in culture, particular-ly in the initial stages of development.

Apoptotic role of fas/fas ligand systemin the regulation of erythropoiesis

In the BM, erythropoiesis occurs in discrete anatom-ic units, the erythroblastic islands, consisting of one ortwo macrophages surrounded by one or more rings oferythroblasts at different maturation stages. The innererythroblastic layers contain immature cells, whereasthe more mature cells are at the periphery of theisland.10 This spatial association of mature and imma-ture erythroblasts may play an important role in ery-thropoiesis as homocellular cell-cell interaction seemsto be required for erythroid cell growth and matura-tion.11 We speculated that maturating erythroblastsmight deliver negative signals to neighbouring cells, asa consequence of a decreased requirement for ery-throid cell production. We therefore studied the pos-sible involvement of Fas and Fas ligand (FasL) in theregulation of erythropoiesis.

Immunohistochemistry of normal BM specimensrevealed that several immature erythroblasts undergoapoptosis in vivo. Analysis of BM erythroblasts and

purified HPCs undergoing unilineage erythroid differ-entiation9 showed that Fas is rapidly upregulated inearly erythroblasts and expressed at high levels throughterminal maturation. However, Fas crosslinking waseffective only in less mature erythroblasts, particular-ly at basophilic level, where it induced apoptosisantagonised by high levels of Epo. In contrast, FasLwas selectively induced in late differentiating Fas-insensitive erythroblasts, mostly at the orthochromat-ic stage. FasL is functional in mature erythroblasts, asit was able to kill Fas-sensitive lymphoblast targets ina Fas-dependent manner. Importantly, FasL-bearingmature erythroblasts displayed a Fas-based cytotoxi-city against immature erythroblasts which was abro-gated by high levels of Epo.

These findings suggest the existence of a negativeregulatory feed-back between mature and immatureerythroid cells, whereby the former cell populationmight exert a cytotoxic effect on the latter one in theerythroblastic island. Hypothetically, this negativefeedback operates at low Epo levels to moderate theerythropoietic rate; however, it is gradually inhibited atincreasing Epo concentrations coupled with enhancederythrocyte production. Thus, the interaction of Fasand FasL may represent a major apoptotic mechanismfor erythropoiesis, contributing to the regulation ofred blood cell homeostasis.

A novel in vitro leukaemogenic model:PML-RARa expression in normal HPCsdictates the leukaemic phenotype

While the role of fusion proteins in acute myeloidleukaemia (AML) is well recognised, the leukaemic tar-get cell and the cellular mechanisms generating theAML phenotype are largely unknown. To address thisissue we have established a novel in vitro leukae-mogenic model: highly purified human HPCs/HSCsare transduced with retroviral vectors carrying cDNAsof the fusion protein and the green fluorescent protein,purified to homogeneity12 and induced into multi- orunilineage differentiation by specific GF combina-tions.9

Expression of PML/RARa fusion protein in humanHPCs/HSCs dictates the acute promyelocyticleukaemia (APL) phenotype, largely via previouslyunreported effects: (i) rapid induction of HPC/HSCdifferentiation to the promyelocytic stage: this is fol-lowed by maturation arrest, which is abolished byretinoic acid; (ii) reprogramming of HPC commitmentto preferential granulopoietic differentiation, irre-spectively of the HGF stimulus: transduction of singlesibling HPCs formally demonstrated this effect; (iii)HPC protection from apoptosis induced by HGFdeprivation. A PML/RARa mutated in the co-repres-sor N-CoR/histone deacetylase binding region13 lostthese biological effects, showing that PML/RARaalters the early haematopoietic programme via N-CoR-dependent target gene repression mechanisms.

Session #1 – Biology of normal and neoplastic progenitor cells 9

R. De Maria et al.10

These observations identify the cellular mechanismunderlying development of the APL phenotype, show-ing that the fusion protein directly dictates the specif-ic lineage and differentiation stage of leukaemic cells.

Altogether, this model system allows us to analysethe proliferation/differentiation potential of trans-duced primary HPCs/HSCs starting from the earliestphases after oncogene transfer through differentia-tion/maturation along each haematopoietic pathway.The investigative potential of this approach is shownby the fact that it allowed reproduction, in primaryHPC culture, of the PML/RARa effects detected so farin cell lines (i.e., maturation block, protection fromapoptosis, sensitisation to RA differentiative stimu-lus),14 while unveiling novel biological actions of thefusion protein (i.e., HPC reprogramming and rapiddifferentiation to the promyelocytic stage). We believethat this model system will have a wide impact, in thatit represents a novel experimental tool for studiesaimed at recapitulating in vitro the genetic events lead-ing to haematopoietic neoplasias.

References

1. Ogawa M. Differentiation and proliferation of hema-topoietic stem cells. Blood 1993; 81:2844-53.

2. Bhatia M, Wang JCY, Kapp U, Bonnet D, Dick JE.Purification of primitive human hematopoietic cellscapable of repopulating immune-deficient mice. ProcNatl Acad Sci USA 1997; 94:5320-5.

3. Zanjani ED, Almeida-Porada G, Livingston AG, FlakeAW, Ogawa M. Human bone marrow CD34- cellsengraft in vivo and undergo multilineage expressionthat includes giving rise to CD34+ cells. Exp Hematol1998; 26:353-60.

4. Gabbianelli M, Sargiacomo M, Pelosi E, Testa U, Isac-

chi G, Peschle C. Pure human hematopoietic progen-itors: permissive action of basic fibroblast growth fac-tor. Science 1990; 249:1561-4.

5. Shalaby F, Ho J, Stanford WL, et al. A requirement forFlk1 in primitive and definitive hematopoiesis and vas-culogenesis. Cell 1997; 89:981-90.

6. Hao QL, Thiemann FT, Petersen D, SmogorzewskaEM, Crooks GM. Extended long-term culture revealsa highly quiescent and primitive human hematopoieticprogenitor population. Blood 1996; 88:3306-13.

7. van der Loo JC, Ploemacher RE. Marrow- and spleen-seeding efficiencies of all murine hematopoietic stemcell subsets are decreased by preincubation withhematopoietic growth factors. Blood 1995; 85:2598-606.

8. Ziegler BL, Müller R, Valtieri M, et al. Unicellular-uni-lineage erythropoietic culture: molecular analysis ofregulatory gene expression at sibling cell level. Blood(in press).

9. Ziegler B, Testa U, Condorelli GL, Valtieri M, PeschleC. Unilineage hematopoietic progenitor differentia-tion in bulk and single cell cultures. Stem Cells 1998;16:51-73.

10. Bernard J. The erythroblastic island: past and future.Blood Cells 1991; 17:5-14.

11. Hanspal M. Importance of cell-cell interactions in reg-ulation of erythropoiesis. Curr. Opin. Hematol. 4:142,1997.

12. Grignani F, Kinsella T, Mencarelli A, et al. High effi-ciency gene transfer and selection of human hemato-poietic progenitor cells with a hybrid EBV/retroviralvector expressing the green fluorescence protein. Can-cer Res 1998; 58:14-9.

13. Grignani F, De Matteis S, Nervi C, et al. Fusion pro-teins of the retinoic acid receptor-a recruit histonedeacetylase in promyelocytic leukaemia. Nature (Lond)1998; 391:815-8.

14. Grignani F Jr, Ferrucci PF, Testa U, et al. The acutepromyelocytic leukemia specific PML/RARa proteininhibits differentiation and promotes survival of U937myeloid precursors. Cell 1993; 74:423-431.

15. Rubnitz JE, Look AT. Molecular basis of leukemogen-esis. Curr Opin Hematol 1998; 5:264-70.

DNA vaccines against haematological malignanciesFREDA K. STEVENSON, CATHERINE A. KING, MYFANWY B. SPELLERBERG, DELIN ZHU, JASON RICE,SURINDER SAHOTA, ANDREW THOMPSETT, JIRI RADL, TERRY J. HAMBLINMolecular Immunology Group, Tenovus Laboratory, Southampton University Hospitals Trust, Southampton, UK

Correspondence: Prof. F.K. Stevenson, Molecular Immunology Group,Tenovus Laboratory, Southampton University Hospitals Trust,Southampton, S016 6YD, UK. Tel. international +44-1703-796923 –Fax. international +44-1703-701416 – E-mail: fs@soton.ac.uk

attack, and we need to tailor our vaccines with thisin mind.

We have investigated idiotypic determinants, whichare expressed by the immunoglobulin of neoplastic Blymphocytes as defined clonal markers, with the pri-vate idiotypic determinants being tumour-specific.They arise from the processes of genetic recombina-tion and somatic mutation which occur during nor-mal B-cell maturation, and are largely preserved fol-lowing neoplastic transformation. Idiotypic Ig can bea surface protein, as in most lymphomas, or a secret-ed protein, as in multiple myeloma. Development ofvaccines which can specifically suppress both thesecategories of B-cell tumour would have relevance forother tumour antigens found in these cellular sites.Idiotypic protein vaccines have been found to induceprotective immunity against B-cell lymphoma in sev-eral mouse models, largely mediated by anti-idiotyp-ic antibody,1 and a small clinical trial in patients withlow grade lymphoma is showing promising results.2However, idiotypic protein is difficult to prepare onan individual patient basis, and antibody is not like-ly to be useful if there is a significant level of secretedprotein. Both these problems could be solved byturning to DNA vaccines which are simple to con-struct, and which are known to activate T cellresponses.

DNA vaccinesDNA plasmid vaccines consist of a backbone of

bacterial DNA containing a cDNA sequence encod-ing the potential antigen. Transcription is usually dri-ven by the powerful CMV promoter, and stimulationof the immune system occurs due to unmethylatedCpG dinucleotide repeats within the backbonesequence.3 Vaccination with DNA containing genesfrom pathogenic organisms has been shown to beeffective in inducing all arms of the immune response,and in protecting against infection.4 Injection is usu-ally in a muscle or skin site, and a gene gun has beenused to deliver DNA to intradermal sites. There is evi-dence for direct transfection of antigen-presentingcells (APCs) in skin, and possibly in muscle. Howev-er, delivery of antigen from a muscle site is probablymainly by secretion and uptake by APCs, or by cross-priming. Clinical trials of DNA vaccines against infec-tious diseases are in progress, with immune respons-es being induced.5

Vaccination as a treatment for cancer is an attrac-tive option, particularly in the setting of a lowlevel of residual disease. However, the task ofactivating a defeated immune system to recognise anddestroy persistent tumour cells is formidable. Haema-tological tumours are a major challenge since thosecells have survived the full power of the immune sys-tem, possibly by inducing tolerance. Three develop-ments, all arising from molecular biological technol-ogy, may allow us to overcome the anticipated obsta-cles. First, there has been a large expansion in ourknowledge of candidate tumour antigens; second, wehave a greater understanding of immune mecha-nisms; third, vaccine vehicles are being developed fordelivery of tumour antigens via pathways able to acti-vate the most effective immune attack.

Our focus is on tumours of B lymphocytes, whichinclude a broad range of diseases, with various cur-rent treatment options. Low grade tumours areresponsive to chemotherapy, but are likely to relapseand are usually incurable. Even tumours such as dif-fuse large cell lymphoma, which may be cured bystandard chemotherapy, prove lethal in 60% of cas-es, and plasma cell tumours such as multiple myelo-ma have a poor outlook with current treatment.However, patients with B-cell tumours can oftenachieve remission, in some cases with transplant sup-port, offering an opportunity to intervene with animmunotherapeutic approach. For vaccines it is ofcourse necessary that remission allows recovery ofimmune capacity.

A widening range of potential tumour antigens isbeing identified on B-cell tumours, including viralantigens, mutated proto-oncogene products, onco-foetal antigens, sequences arising from chromoson-al translocation events, mucins and idiotypes. Can-didate antigens can be placed into various categoriesdepending on how they are expressed by the tumourcell. One group of antigens is expressed as glycopro-teins at the cell surface; a second group is presentedas peptides bound to MHC class I or II molecules;and a third group consists of secreted antigens. Eachcategory of tumour antigen will require an appropri-ate immune effector mechanism for successful

Haematologica 1999; 84:(EHA-4 educational book):11-13Educational Session 2Chairman: F.K Stevenson

Biotherapy strategies inhaematological malignancies

F. K. Stevenson et al.12

DNA vaccines against B-cell tumoursTo construct a DNA vector for idiotypic vaccina-

tion, it is necessary to identify the VH and VL genesused to encode tumour Ig, and this can be achievedby PCR/cloning methods.6 We have chosen to assem-ble genes as single chain Fv (scFv), but it is also pos-sible to construct vectors to express whole Ig, con-taining either human or mouse constant regionsequences.7 It soon became clear that DNA vaccinescontaining scFv alone, or as homologous Ig, couldnot activate a significant anti-idiotypic immuneresponse.8 The presence of human constant regionsequences improved the antibody response toattached mouse V-region sequences,7 and inclusion ofcytokine sequences or peptides could also increaseresponses. However, we obtained a dramaticimprovement in anti-idiotypic responses by fusing agene encoding a fragment derived from tetanus toxinto the 3’ end of the scFv. Attachment of this non-tox-

ic fragment (fragment C[FrC]) activated anti-idiotyp-ic responses against a range of human scFv sequencesin mice, where scFv alone had failed.9 Importantly,the scFv appeared to fold effectively in the scFv-FrCfusion protein.9 Promotion of immunity probablyresults from both increased recognition by APCs, andby a massive increase in T-cell help. This help is pro-vided by T cells specific for FrC and, since gene fusionis an absolute requirement,10 this acts via a classicalhapten-carrier mechanism.

To assess the generality of the approach for othertumour antigens, carcinoembryonic antigen (CEA),we replaced the scFv with a different tumour-associ-ated antigen, also expressed at the cell surface. A sim-ilar promotional effect of fusion was seen with thisantigen, with CEA sequence alone inducing poor anti-body responses, whereas the CEA-FrC fusion geneinduced very high levels of anti-CEA antibodies. Thissuggests that the same fusion format may be usefulfor many surface-expressed tumour antigens.

Figure 1. Induction of protec-tive immunity against B-celltumours by DNAscFv-FrC vac-cines. Mice were vaccinatedwith DNAscFv-FrC or DNAscFvderived from either: A, amouse lymphoma (A31); or B,a mouse myeloma (5T33), atdays 0, 21 and 42. Anti-idio-typic antibody levels againstthe tumour Ig were measured,and mice were then chal-lenged with tumour. In eachcase, anti-idiotypic antibodiesand protection were inducedby the DNAscFv-FrC fusionvaccine but not significantlyby the DNAscFv alone.

Session #2 – Biotherapy strategies in haematological malignancies 13

DNA vaccines induce protective immunity against tumour

A DNA vaccine containing scFv-FrC fusion genesderived from a mouse lymphoma, A31, was injectedinto syngeneic mice and was found to induce anti-idiotypic antibody. Importantly, this vaccine protect-ed mice against challenge with a malignant lym-phoma10 (Figure 1). It appeared, therefore, that thisdesign was effective as a vaccine against a cell surfacetumour antigen. We then tested the same vaccinedesign against a mouse myeloma, 5T33. This myelo-ma is one of a series which show characteristics sim-ilar to human myeloma, including osteolytic lesions.The neoplastic plasma cells are completely surface Ig-negative, but secrete monoclonal Ig. Interestingly, thescFv-FrC fusion gene also induced protective immu-nity against this tumour10 (Figure 1). Although thevaccine-induced high levels of anti-idiotypic antibody,it was unlikely that antibody could mediate protec-tion. This was confirmed by the fact that vaccinationwith idiotypic protein/CFA, which induced compara-ble levels of anti-idiotypic antibody, completely failedto protect against tumour.10 We were unable to definea motif in the V-gene sequences suitable for bindingto MHC class I, and it is therefore likely, but not yetproven, that CD4+ T cells are involved in protection.

Clinical trials of scFv-FrC DNA vaccineA phase 1 clinical trial of idiotypic DNA vaccines

containing only scFv has been carried out in 7 patientswith end-stage low grade follicular centre lymphoma(FCL) largely to assess any toxicity. No ill effects wereobserved, and we have now been allowed to proceedwith a second trial using DNA scFv-FrC fusion genes.The patients have FCL in first remission. We shallinvestigate escalating doses of DNA, and will mea-sure antibodies against both tumour-derived scFv,and against FrC. This will provide information on theability of patients to respond to both a tumour anti-gen and an exogenous pathogen-derived antigen,both delivered by DNA.

SummaryDNA vaccines against cancer have to activate an

inadequate or damaged immune system in order toattack residual cancer cells. Although the potentialproblem of tolerance may be overcome by transplan-tation, provision of high levels of T-cell help is likely tobe an important factor in stimulating effectiveimmune pathways. The fusion gene approachappears to provide the required help, and offers a

rational design for raising both antibody and T-cellmediated attack against lymphoma and myeloma,which express idiotypic antigen at the cell surface oras a secreted protein respectively. Intriguingly, pre-liminary data indicate that the fusion gene approachpromotes antibody responses against a different cellsurface tumour antigen, CEA. Strategies for usingDNA vaccines to induce attack on processed peptidesbound to MHC class I molecules are also being devel-oped. We hope and anticipate that all categories oftumour antigen may be susceptible to this powerfulnew technology. The critical clinical requirement,however, will be to treat the presenting tumour withmaintenance or restoration of immune capacity. Weawait results of the preliminary clinical trials with greatinterest.

References

1. George AJT, Tutt AL, Stevenson FK. Anti-idiotypicmechanisms involved in suppression of a mouse B celllymphoma, BCL. J Immunol 1987; 138:628-34.

2. Hsu F, Caspar CB, Czerwinski D. Tumor-specific idio-type vaccines in the treatment of patients withrelapsed low-grade non-Hodgkin’s lymphoma. Blood1997; 89:3129-35.

3. Sato Y, Roman M, Tighe H, et al. Immunostimulato-ry DNA sequences necessary for effective intradermalgene immunization. Science 1996; 273:352-4.

4. Ulmer JB, Donnelly JJ, Parker SE, et al. Heterologousprotection against influenza by injection of DNAencoding a viral protein. Science 1993; 259:1745-9.

5. Wang R, Doolan DL, Le TP, et al. Induction of anti-gen-specific cytoxic T lymphocytes in humans by amalaria DNA vaccine. Science 1998; 282:476-80.

6. Hawkins RE, Zhu D, Ovecka M, et al. Idiotypic vacci-nation against human B-cell lymphoma. Rescue ofvariable region gene sequences from biopsy materialfor assembly as single-chain Fv personal vaccines.Blood 1994; 83:3279-88.

7. Syrenglas AD, Chen TT, Levy R. DNA immunizationinduces protective immunity against B cell lymphoma.Nat Med 1996; 2:1038-41.

8. Stevenson FK, Zhu D, King CA, Ashworth LJ, Kumar S,Hawkins RE. Idiotypic DNA vaccines against B-celllymphoma. Immunol Rev 1995; 145:211-28.

9. Spellerberg MB, Zhu D, Thompsett A, King CA, Ham-blin TJ, Stevenson FK. DNA vaccines against lym-phoma. Promotion of anti-idiotypic antibody respons-es induced by single-chain Fv genes by fusion totetanus toxin fragment C. J Immunol 1997; 159:1885-92.

10. King CA, Spellerberg MB, Zhu D, et al. DNA vaccineswith single chain Fv fused to fragment C of tetanustoxin induce protective immunity against lymphomaand myeloma. Nat Med 1998; 4:1281-6.

Monoclonal antibodies in the treatment of non-Hodgkin’s lymphoma patientsBERTRAND COIFFIERHospices Civils de Lyon and Université Claude Bernard, Lyon, France

Correspondence: Prof. B. Coiffier, Service d’Hématologie, CH Lyon-Sud,69495 Pierre-Bénite Cedex, France. Tel. international +33 478861194 – Fax: international +33 478 866566 – e.mail: coiffier@hema-tologie.univ-lyon1.fr

tope attached to the antibody as well as the antigen.Despite the simplicity of the concept, the choice ofthe antigen and the antibody is difficult. The selectionof a suitable antigen was the first step for thesetreatments. The antigen must not be shared by criti-cal tissues such as haematopoietic stem cells, mustnot be associated with much toxicity if all target cellsare eliminated, and only be present on lymphomacells. Unfortunately, specific antigens for B or T lym-phoma cells are unknown and all antigens are sharedwith the normal B or T cell counterparts. The targetantigen must be present either on all lymphoma cellsor on the self-renewing clonogenic population of lym-phoma cells. This target antigen must preferentiallyhave a critical role in the homeostasis of these lym-phoma cells and be necessary for the cell survival.Lymphoma cells should not be able to escape theantibody effect through the development of antigenvariants, antigen-negative clones, the shedding of theantigen in the extracellular compartment, or themodulation of the antigen on the cell surface. IfMoAb-toxin conjugates need to be internalised forthe toxin to access the critical cellular processes,unconjugated MoAbs must remain on the cell surfaceto allow the Fc portion of the antibody to activateimmunologic mechanisms. Antigen density andMoAb binding affinity may influence the cytotoxicefficacy for unconjugated antibodies but radioim-munoconjugates emit particles with enough energy tokill adjacent cells, cells with a low antigen density ornon-antigen-bearing cells. However, they may kill vitalnormal cells and increase the toxicity of the treat-ment.

The MoAb must reach tumour cells in every parts ofthe organism and at all sites of the disease. This maybe a problem in large tumours that are poorly vascu-larised. The presence of circulating antigens may be aproblem leading to rapid clearance of the MoAb. TheMoAb must not be eliminated through immunologicmechanisms because of its own difference from thehost. When xenophobic Ab are used, rapid appear-ance of human anti-mouse antibodies (HAMA) mayalter the pharmacokinetics of the MoAb, particularlyduring the re-treatment phases, and then furtherdecrease their activity. Genetic engineering hasallowed the humanisation of antibodies and the cre-ation of chimaeric proteins with a small antigen-bind-ing mouse part and a large human constant Fc region.

During the last ten years, significant progresshas been made in the treatment of non-Hodgkin’s lymphoma (NHL) patients. Therefinement of the classification of the lymphomasand the incorporation of prognostic indices in deci-sion making1 have allowed identification of groupsof patients who may be cured and others that needfurther research to find the correct therapeuticoptions. Therapeutic improvements came from theintroduction of G-CSF allowing higher dose of cura-tive drugs with less risk of severe infections, the devel-opment of high-dose therapy with autotransplanta-tion, and the addition of interferon to multidrugchemotherapy regimens. Through these options,around half of the patients with de novo NHL mayexpect to be alive ten years later. However, the diffi-culty of overcoming tumour cell resistance to stan-dard drugs, the toxicity of the newly developed regi-mens, and the ageing population of patients chal-lenge us to develop less toxic but more effective treat-ments.

The idea of developing monoclonal antibodies(MoAb) against cancer cells, particularly lymphomacells, appeared more than 20 years ago with thedescription of the different antigens found on cellmembranes. Since the first attempt reported in1980,2 the exponential increase in the progress inmolecular biology and protein engineering hasrecently culminated in the approval, by drug agen-cies, of rituximab and, in a near future, of radio-labelled antibodies, for in vivo treatment of lymphomapatients. Although monoclonal antibodies were usedfor a long time to purge haematopoietic stem cells exvivo before autotransplant or for T-cell depletion ofallogeneic cells and they are currently being devel-oped against non-lymphoma cancer cells, this reviewwill only focus on the future roles for MoAb in thetreatment of lymphoma patients.

Three main approaches have been used in thedevelopment of MoAb therapy. Unconjugated anti-bodies mediate cell death through different mecha-nisms related to the antigen and the antibodies. Con-jugated antibodies act through a toxin or a radioiso-

Haematologica 1999; 84:(EHA-4 educational book):14-18Educational Session 2Chairman: F.K Stevenson

Biotherapy strategies inhaematological malignancies

These chimaeric MoAbs have substantially lowerimmunogenicity and, thus, prolonged half-life. Theyalso have an improved ability to mediate complement-dependent cytotoxicity (CDC) and antibody-depen-dent cell-mediated cytotoxicity (ADCC), whichincreases their potency.

The MoAb may kill the lymphoma cells through avariety of different mechanisms. Radioimmunocon-jugates or immunotoxins kill cells through the emis-sion of particles or the internalisation of the toxin.Unconjugated MoAb may trigger CDC or ADCC.They may have direct cytotoxic effects on tumourcells,3 either on blocking the binding of an endoge-nous ligand, which deprives the cell of a critical sur-vival signal, or by mimicking it, which triggers growtharrest. These functions may potentiate the effect ofchemotherapy.

A large variety of antigens may potentially be cho-sen as targets. While the early trials focused on Ig idio-type, the CD20 antigen is probably the ideal target forB-cell lymphomas. It is not expressed on stem cells orprecursor B-cells but is found on normal mature B-cells and malignant B-cells, with the exception of plas-ma cells and myeloma cells. It is usually present on allcells of the tumour clone. It is expressed in high den-sity on all B-cell lymphoma but not chronic lympho-cytic leukaemia cells. This antigen is stable in themembrane of B-cell, does not have any known vari-ant, is not shed, and does not modulate or internalisein response to antibody binding. While its biologicfunction is not fully known, it appears to function asa calcium channel and it either forms a membranepore or controls a pore that is involved in calciumtransfer during the cell cycle. The majority of the mol-ecule is within the membrane or inside the cell andthere is a small loop of 40 amino acids outside thecell. All anti-CD20 appear to bind to the same sectionof this external loop except L26, which binds to anintracellular epitope of the molecule.

Unconjugated MoAbsUnconjugated MoAbs constitute the simplest appli-

cation of targeted MoAb treatment. Table 1 lists thedifferent antigens that have been used. The first trialsused MoAbs directed against the idiotype of the sur-face Ig of lymphoma cells which certainly represents aunique tumour-specific antigen. Most of these trialswere conducted by Levy at Stanford.4 In different trials,anti-idiotype antibodies produced responses in 50% to70% of the patients. Although the median duration ofthese responses was only 6 months, some patients withcomplete response had long remissions. However,patients relapsed with idiotype-negative cells. The pres-ence of circulating shed idiotype and the formation ofHAMA further limited the efficacy of unconjugatedMoAbs. This, associated with the constraint of makinganti-idiotype specific for each patient, precluded fur-ther development of this therapy.

Subsequently, investigators used pan-B antigens,

such as CD20, and humanised antibodies. Rituximab(MabThera®) is the most extensively studied uncon-jugated MoAb to date. This chimeric antibody con-sists of the murine variable regions from the 2B8MoAb grafted onto a human IgG1 constant region. Ina phase I trial, the dose limiting toxicity was notreached which attests to a low side-effect profile ofthe drug.5,6 Most phase II trials used the dose of 375mg/m2 once a week for 4 weeks.7,8 These trials accruedpredominantly patients with indolent follicular lym-phoma (FL), refractory to standard chemotherapyregimens. In more than 200 patients, the responserate was around 50%, with 6% complete responses,and responses were observed in different subgroupswith adverse prognostic features, such as previousautotransplant or bulky tumour. Many patients hadno detectable residual tumour cells in blood or bonemarrow, even as detected by PCR analysis for thet(14;18) translocation (molecular remission). Themedian time to response was about 2 months, withmany patients showing progressive responses for sev-eral months. This may be correlated with the longhalf-life of the antibody, some patients having resid-ual levels detectable 6 months after the last infusion.Median time to progression in responding patientswas longer than 12 months. Interestingly, patientswho progressed after a first response could be re-treated and 50% of them responded. Several succes-sive responses were observed in some patients.Because of these results drug agencies approved theindication of rituximab for relapsing FL patients.

Most adverse events were minor and associated withthe first infusion. They consisted primarily of fever,chills, mild nausea, mild fatigue, or malaise. Rarely,patients developed more serious reactions, includinghypotension, bronchospasm, or sensation of throatswelling. These symptoms were usually managed bytemporarily slowing or stopping the antibody infusion.The most serious reactions were observed in patientswith peripheral blood involvement or large tumours.These reactions could be prevented by stopping the

Session #2 – Biotherapy strategies in haematological malignancies 15

Table 1. Different antigens chosen for unconjugated MoAbtherapy in lymphoproliferative diseases.

Antigens Monoclonal antibodies Humanised

CD4 cMT412 ChimaericCD5 T101 NoCD10 J5 NoCD19 CLB-CD19 NoCD20 1F5 NoCD20 IDEC-C2B8 (rituximab) ChimaericCD21 OKB7 NoCD25 Anti-TAC NoCD52 Campath-1M No (rat)CD52 Campath-1H ChimaericHLA-DR LYM-1 NoIg idiotype anti-idiotype No

infusion once mild adverse reactions occur. Myelo-suppression was rare. As expected, the normal B-lym-phocytes rapidly declined and recovered over 6 to 9months. However, there was no increase in infectionrate and no occurrence of opportunistic infections,probably because Ig levels and T-cells remained nor-mal. HAMA was observed in less than 1% of thepatients and HACA (anti-chimaeric antibody) was notobserved.

Subsequently, rituximab was used in patients withmore aggressive B-cell lymphoma, mantle cell lym-phoma (MCL) and diffuse large B-cell lymphoma(DLCL). In a phase II trial, patients with relapsing dis-ease showed a 32% response rate.9 In another phaseII trial, mantle cell lymphoma patients, in first line ther-apy or relapsing, showed a 40% and 30% responserate, respectively. Most of the responses in these stud-ies were incomplete, with less than 10% being com-plete responses, and median time-to-progression wasless than 12 months. These results indicate that ritux-imab has an anti-tumour activity in nearly all B-celllymphomas. Currently, it is being tested in chroniclymphocytic leukaemia, post-transplant lymphopro-liferative diseases, and multiple myeloma. Althoughplasma cells do not express the CD20 antigen, thereis some indication that the clonogenic precursors may.

Rituximab has been used in conjunction withchemotherapy in FL and DLCL patients, mostly withthe CHOP regimen.10 In a small study of 40 patientsa response rate of 95% was reached with 55% com-plete responses. At time of publication, 74% of theresponding patients had not progressed with a medi-an follow-up of 29 months. Seven of the 8 patientswith bcl-2-rearrangement converted to PCR negativi-ty after completion of the treatment. The adverseevents were not different from those expected fromthe CHOP regimen or rituximab treatment. No spe-cific toxicity was observed with the combination ofthese drugs. Similar results were described in DLCLpatients. Currently, large co-operative group ran-domised trials are addressing the use of rituximab inconjunction with multidrug chemotherapy as first linetreatment or as maintenance after chemotherapy.Other trials are testing in vivo purging before the har-vest of peripheral stem cells for autotransplant. Untilthese results are complete, the definitive place of rit-uximab cannot be established.

Another humanised MoAb, CAMPATH-1H, hasundergone evaluation in different indolent B-cell pro-liferations. CAMPATH-1 MoAb is directed againstCD52, an antigen expressed by B- and T-lymphocytes,granulocytes, and monocytes but not by stem cells. Aseries of trials documented responses in CLL and T-cellprolymphocytic leukaemia.11 Adverse events consistedof infusion-related reactions, grade 4 neutropenia,thrombocytopenia, or lymphopenia, which led to pro-found immunosuppression and multiple opportunis-tic infections. Complementary studies are in progressto define a safer schedule of administration.

Studies are only beginning or have been short livedwith other MoAb and no or few further developmentsare expected with them.

ImmunotoxinsAn alternative approach to increase the activity of

MoAb is the development of an immunotoxin, a con-struct conjugating the antibody to cytotoxic plant orbacterial toxic proteins. The commonly used toxins,ricin and diphtheria toxin, are highly potent naturalproducts that disrupt protein synthesis. Unlike uncon-jugated MoAb, immunotoxins must be internalisedafter antigen binding to allow the toxin access to thecytosol. Although the conjugation to MoAb conferssome target specificity, the toxin continues to medi-ate non-specific toxicity to normal tissues. Deglyco-sylated ricin A-chain has been used to eliminate suchnon-specific toxicity.

The vast majority of the immunotoxin trials havebeen phase I studies designed to determine the max-imum tolerated dose (Table 2). These trials haveshown that therapeutic serum levels may be achievedwith tolerable toxicity. A relatively uniform toxicity hasbeen observed with vascular leak syndrome, hepato-toxicity, and myalgia. A strong immunologic responseagainst the construct or the toxin was observed andre-treatment was not feasible in most patients. Thedifferent trials have shown a low response rate of 10%to 25% partial responses without durable efficacy. Thefuture of this therapy will depend on decreasing toxi-city, decreasing immune response against the con-struct, and on increasing the anti-tumour activity.

Radiolabelled antibodiesThese consist of a radionucleide, usually 131iodine

(131I) or 90yttrium (90Y) emitting –particles, coupled toa MoAb (Table 3). These compounds may selective-ly deliver ionizing radiation to tumour cells. Theseagents seem to possess several advantages over oth-er antibody constructs: they do not rely on recruit-ment of patients’ immune effector mechanisms andthe –particles are capable of killing cells from a dis-tance of several cell diameters permitting the killing ofantigen-negative tumour cells. Most studies of radio-

B. Coiffier16

Table 2. Antigens, antibodies, and toxins used for immuno-toxin therapy.

Antigens Antibodies Toxins

CD5 anti-H65-RTA blocked ricinCD19 IgG-HD37-dgA deglycosylated ricin A-chainCD19 anti-B4-bR blocked ricinCD22 RFB4-dgA deglycosylated ricin A-chainCD25 DAB389 IL-2 truncated diphtheria toxinCD25 DAB486 IL-2 truncated diphtheria toxinCD30 BerH2-saporin saporin

labelled MoAb require careful dosimetry measure-ments before the administration of the therapeuticdoses to ensure that the radiation doses delivered totumour sites exceed the doses to normal tissues. Alarge tumour burden, particularly in the spleen, caninterfere with the distribution of radiation. To min-imise the doses distributed to normal tissues, a firstinfusion of non-radiolabelled (cold) MoAb is oftenadministered to the patients. Although 131I and 90Yboth emit b-particles, 90Y emits higher energy parti-cles, which have a deeper tissue penetration, and 131Ialso emits –radiation and has a longer half-life. Theselast characteristics may present safety concerns limit-ing its use to large centres with strict radiation isola-tion. Both compounds require an onsite radiophar-macy and dosimetry calculations that may also limittheir application.

Radiolabelled MoAb therapy is associated withmyelosuppression, toxicity not found with coldMoAb, although a considerable inter-patient vari-ability has been observed. With low dose radiationtherapy neutropenia and thrombocytopenia areobserved 3 to 4 weeks after the infusion and may per-sist for 16 weeks but with higher doses these haema-tological effects are more rapid, profound, and pro-longed. Extensive bone marrow involvement can leadto a greater binding and a larger radiation dose deliv-ered to the normal haematopoietic cells. In additionthe same adverse effects as those which are observedwith cold MoAb treatment exist and the use of iodineconjugates can cause hypothyroidism. Murine MoAbtherapy is associated with the appearance of HAMAthat may limit the possibility of re-treatment.

The trials with non-myeloablative doses have used131I targeted to HLA-DR, CD20, CD21, or CD22 anti-gens and have achieved responses in 5% to 80% oftreated patients. Preliminary studies with the 131I-labelled Lym1 antibody in refractory patients yieldeda 50% response rate.12 The 131I-labelled anti-CD20antibody (Bexxar®) produced durable CR in patientswith recurrent FL.13 Patients received an unlabelled

dose of the antibody followed by a trace-imaging doselabelled with 131I. One or two weeks later, the patientreceived an additional dose of unlabelled antibodyfollowed by a therapeutic dose of 131I calculated toproduce 75 cGy to the whole body. A 60% responserate was observed, with 27% CR, and a median dura-tion of CR longer than 1 year. In de novo patients, ahigher response rate was reported. Longer follow-upin more patients is necessary to determine whetherthis approach will result in long-term disease control.In phase I-II trials, 90Y-labelled 2B8 antibody admin-istered after a rituximab infusion yielded a 82%response rate. Larger numbers of patients are neededto determine this antibody’s place.

Even greater response rates have been reported instudies that used myeloablative doses of 131I-labelledanti-CD20 antibody. Patients received a therapeuticinfusion of the antibody, were isolated, and 10-12days later were given their cryopreserved stem cells.14Toxicity of this approach included severe infectionsand cardiomyopathy. However, 70% to 80% of theCR patients remained disease-free 18 months later.Longer follow-up and randomised study will deter-mine whether this approach confers survival advan-tage over conventional high-dose regimens.

ConclusionsIn conclusion, impressive responses have been doc-

umented with MoAb therapy, even if their role in thetreatment of lymphoma remains to be determined.Although the unconjugated MoAb, and particularlyrituximab, have showed activity as single agents, theymay have a greater role in combination with chemo-therapy or as maintenance in responding patients.Their ability to purge blood and bone marrow to anundetectable level of lymphoma cells may allow re-infusion of minimally contaminated haematopoieticstem cell harvests after high dose therapy. The com-bination of these MoAb with a radionucleide mayresult in greater efficacy but a greater toxicity, partic-ularly myelosuppression. The definitive use of theseMoAb cannot be recommended before the results ofcurrent prospective studies are available.

References

1. Coiffier B. Non-Hodgkin’s lymphomas. In: Cavalli F,Hansen HH, Kaye SB, eds. Textbook of medical oncol-ogy. London: Martin Dunitz Ltd; 1997. p. 265-87.

2. Nadler LM, Stashenko P, Hardy R, et al. Serotherapyof a patient with a monoclonal antibody directedagainst a human lymphoma-associated antigen. Can-cer Res 1980; 40:3147-54.

3. Shan D, Ledbetter JA, Press OW. Apoptosis of malig-nant human B cells by ligation of CD20 with mono-clonal antibodies. Blood 1998; 91:1644-52.

4. Brown SL, Miller RA, Levy R. Antiidiotype antibodytherapy of B-cell lymphoma. Semin Oncol 1989; 16:199-10.

5. Maloney DG, Liles TM, Czerwinski DK, et al. Phase I

Session #2 – Biotherapy strategies in haematological malignancies 17

Table 3. Radioimmunoconjugates used in the treatment oflymphomas.

Antigens Antibodies Isotopes Dose in mCi

Ig anti-idiotype 90Y 10 - 55HLA-DR LYM-1 131I 25 - 1050CD5 T101 131I or 90Y 25 - 150CD20 B1 131I 30 - 850CD20 1F5 131I 600CD20 C2B8 90Y 10 - 60CD21 OKB7 131I 90 - 200CD22 LL2 131I 15 - 350CD25 Anti-TAC 90Y 5 - 70CD37 MB1 131I 25 - 650Ferritin Anti-ferritin 131I or 90Y 20 - 100

B. Coiffier18

clinical trial using escalating single-dose infusion ofchimeric anti-CD20 monoclonal antibody (IDEC-C2B8) in patients with recurrent B-cell lymphoma.Blood 1994; 84:2457-66.

6. Maloney D, Grillo-López A, Bodkin D, et al. IDEC-C2B8: Results of a phase I multiple-dose trial inpatients with relapsed non-Hodgkin’s lymphoma. JClin Oncol 1997; 15:3266-72.

7. Maloney D, Grillo-López A, White C, et al. IDEC-C2B8(rituximab) anti-CD20 monoclonal antibody therapyin patients with relapsed low-grade non-Hodgkin’slymphoma. Blood 1997; 90:2188-95.

8. McLaughlin P, Grillo-Lopez AJ, Link BK, et al. Ritux-imab chimeric anti-CD20 monoclonal antibody ther-apy for relapsed indolent lymphoma: half of patientsrespond to a four-dose treatment program. J ClinOncol 1998; 16:2825-33.

9. Coiffier B, Haioun C, Ketterer N, et al. Rituximab(anti-CD20 monoclonal antibody) for the treatmentof patients with relapsing or refractory aggressive lym-phoma. A multicenter phase II study. Blood 1998; 92:1927-32.

10. Czuczman MS, Grillo-Lopez AJ, White CA, et al. Treat-ment of patients with low-grade B-cell lymphoma withthe combination of chimeric anti-CD20 monoclonalantibody and CHOP chemotherapy. J Clin Oncol1999; 17:268-76.

11. Osterborg A, Dyer MJS, Bunjes D, et al. Phase II mul-ticenter study of human CD52 antibody in previouslytreated chronic lymphocytic leukemia. J Clin Oncol1997; 15:1567-74.

12. DeNardo GL, DeNardo SJ, Goldstein DS, et al. Maxi-mum-tolerated dose, toxicity, and efficacy of I-131-Lym-1 antibody for fractionated radioimmunothera-py of non-Hodgkin’s lymphoma. J Clin Oncol 1998;16:3246-56.

13. Kaminski MS, Zasadny KR, Francis IR, et al. Iodine-131. Anti-B1 radioimmunotherapy for B-cell lym-phoma. J Clin Oncol 1996; 14:1974-81.

14. Press OW, Eary JF, Appelbaum FR, et al. Phase II trialof I-131-B1 (anti-CD20) antibody therapy with autol-ogous stem cell transplantation for relapsed B cell lym-phomas. Lancet 1995; 346: 336-40.

Antisense oligonucleotides for haematological malignanciesFINBARR E. COTTERLRF Molecular Haematology Unit, Institute of Child Health, London UK

Correspondence: Finbarr E. Cotter, LRF Molecular Haematology Unit,Institute of Child Health, 30 Guilford Street, London WC1N 1EH, UK.Tel. international +44-171-8138191 – Fax. international +44-171-8138100 – E-mail. f.cotter@ich.ucl.ac.uk

to improved ASO molecules and the prospect ofeffective molecular therapy.

Antisense oligonucleotide design

Optimising choice of targetThe ability of naturally occurring RNA to anneal is

a crucial process for living cells. However, the abilityto design artificial antisense oligonucleotides has beenhampered by the frequent inability of the ASO toanneal successfully to the mRNA. Efficient RNA-ASOannealing involves the interaction between highlystructured RNA elements. Empirical trial of a largeselection of oligonucleotides for an mRNA sequencehas tended to be the method for effective ASO selec-tion, with many ineffective molecules being discarded.A lack of understanding of the structure of RNA andthe rules governing the annealing properties underliethis failure. Two approaches have been effective inimproving ASO design. The first is an empiricalapproach, essentially generating a large number ofsynthetic oligonucleotides to cover all the possibleselections and hybridising these against the targetmRNA. This may be performed by generating griddedarrays of ASO logically covering all computations fora mRNA onto a glass plate and then hybridising thelabelled message against the grid. The ASO that canaccess the RNA structure and anneal gives a positivesignal. Identification of a good ASO molecule is read-ily provided but the technique is limited to examininga maximum of 400 bases of RNA for each grid. Inmost cases this is quite adequate. A greater length ofgene specific RNA can be examined by hybridising thelabelled RNA against random or semirandom libraries(pools of single stranded ASO) of oligonucleotides,the latter having been shown to be more effective. Thelibrary approach is more prone to some false negativeresults but does not require the polymer technologyrequired to attach the oligonucleotides to the glass forthe gridded arrays. Both are extremely effective atidentifying accessible RNA sites for ASO. The secondASO selection approach is based on the computersupported structural design of RNA. The secondarystructure of the target RNA may be predicted by theprogramme mfold 2.0 and the structural parametersrecorded. Favourable ASO structures are searched forby examining for a maximal number of external basesand components and a minimised loop degree. Sim-plistically this means an ASO that complements anmRNA region with a weak secondary structure will

The growth, differentiation, appearance andfunction of all living cells is dictated by pro-teins. The code for each protein is stored in theform of double helix DNA within the nucleus of everycell. Expression of such a gene is performed by tran-scribing the base sequence of the DNA into singlestranded messenger RNA (mRNA) which is able tomove from the nucleus to the cytoplasmic compart-ment of the cell were it is translated into a protein.Many disease processes are characterised by inap-propriate or excessive expression of normal or chi-maeric proteins as a consequence of altered DNAtranscription. Current drugs predominantly functionby altering protein expression within the target cell,however, the process is far from efficient and oftenlacks specificity. It is an attractive proposition to usenovel approaches to block transcription or transla-tion of individual genes in order to lower disease-causing proteins. An emerging and powerfulapproach is the use of antisense ol