gene therapy for hematological disorders

The observed lack of clinical utility results from several major hurdles, including inefficient gene transfer to desired target cells, espe-cially stem cells, poor in vivo expression of

CHAPTER 6

Gene Therapy for Hematological

Disorders

CYNTHIA E DUNBAR, M.D and TONG WU, M.D.

INTRODUCTION

Hematopoietic cells are an attractive target for gene therapy for two main reasons First, it is possible to easily collect and then manipulate hematopoietic cells in vitro Second, many congenital and acquired diseases are potentially curable by genetic correction of hematopoietic cells, especially hematopoietic stem cells (HSCs, see Fig 6.1) For hematological disorders, the target cell(s) in which gene expression

is required are red blood cells (RBC), lymphocytes, granulocytes, or other mature blood elements Ideally, the transgene is integrated into the chromatin of pluripo-tent HSCs, ensuring the continuous production of genetically modified blood cells of the desired lineage for the lifetime of the patient Other potential cellular targets with potential utility in the treatment of hematologic diseases include dendritic cells, tumor cells, and endothelial cells Hepatocytes, myocytes, and keratinocytes can be considered as “factories” for soluble factors with clinical utility

in hematologic diseases such as hemophilia (see Chapter 7) Relevant targets and applications for gene therapy of hematopoietic or immune system disorders are summarized in Table 6.1

Many important advances in our understanding of hematopoiesis, stem cell engraftment, and other basic principles have resulted from animal models, in vitro studies, and early clinical trials of gene marking or gene therapy For example, studies using retrovirally marked murine stem cells show tracking and a quantita-tive analysis of murine stem cell behavior Experiments overexpressing oncogenes

or cytokines in hematopoietic cells have elucidated the in vivo role of these pro-teins Early clinical gene marking trials demonstrated the long-term engrafting capa-bility of peripheral blood stem cells The observed lack of clinical utility results from several major hurdles, including inefficient gene transfer to desired target cells, espe-cially stem cells, poor in vivo expression of introduced genes, and immune responses against gene products recognized as foreign Further basic research investigations

133

Copyright © 2001 by Wiley-Liss, Inc ISBNs: 0-471-39188-3 (Hardback); 0-471-22387-5 (Electronic)

FIGURE 6.1 Hierachal model of lymphohemopoiesis A primitive lymphohemopoietic cell

is capable of producing lymphoid stem cells for lymphopoiesis or myeloid stem cells for hemopoiesis These stem cells give rise to progressively more differentiated progenitor cells that eventually give rise to lineage-specific terminally differentiated effector cells.

into new or modified vector systems and target cell biology are necessary to move the field forward into real clinical utility

REQUIREMENTS FOR GENE TRANSFER INTO HEMATOPOIETIC CELLS

Ex Vivo Versus in Vivo Gene Transfer

Specific aspects of gene transfer techniques are advantageous for gene therapy approaches when applied to hematological diseases Aspects of ex vivo gene trans-fer as well as certain gene transtrans-fer vector systems are particularly useful in the experimental therapy of hematological diseases Hematopoietic cells such as stem cells or lymphocytes are generally transduced ex vivo because these cells can be easily collected, cultured, and transduced in vitro (see Chapter 1) Subsequently, they can be reinfused intravenously Ex vivo transduction allows for a controlled exposure of only the desired targets to vector particles It is less likely to produce

an immune response or be impeded by complement-induced vector inactivation However, limited data indicate that direct in vivo injection of vector into the marrow space can transduce primitive cells But, there is no evidence that this in vivo method currently has any advantages over the more fully characterized ex vivo transduc-tion approaches In vivo gene transfer is most appropriate for target cells that cannot

be easily harvested or manipulated ex vivo, such as airway epithelium, vascular endothelium, and differentiated muscle cells

Vector Systems and Nonviral Vectors

The choice of an appropriate vector system depends on the biology of the desired target cell and the need for transient versus prolonged gene expresssion (see Chapter 4) Both viral and nonviral vectors have been utilized to transduce hematopoie-tic target cells If prolonged correction or modification of hematopoiehematopoie-tic cells is required, then vectors such as retroviruses that efficiently integrate into target cell chromosomes are necessary, otherwise new genetic material will be lost as HSCs or other targets such as lymphocytes proliferate On the other hand, if transient expres-sion is required, for instance, in the production of leukemic cell tumor vaccines, then nonintegrating but efficiently expressing vectors such as adenoviruses may be pre-ferred The vast majority of preclinical and clinical investigations of hematopoietic cell gene transfer utilize viral vectors, taking advantage of the characteristics of the virus that have evolved over time to efficiently infect target cells The viral genes and replication machinery are replaced with nonviral transgene sequences of interest For murine retroviruses, the Moloney murine leukemia virus (MuLV) vectors are the vectors of choice since they have not been supplanted by any other vector system for most hematologic applications Thus, MuLV vectors have been employed

in almost every clinical study to date The main advantages of MuLV vectors are their ability to integrate a stable proviral form into the target cell genome, the avail-ability of stable producer cell lines, the lack of toxicity to target cells, and almost 10 years of experience in using them safely in clinical trials Over the past several years,

TABLE 6.1 Relevant Targets and Applications for Gene Therapy of Hematopoietic or Immune System Disorders

Cancer (TIL) AIDS (intracellular immunization)

Antisense to oncogenes Tumor vaccines Suicide genes

Growth factors

Keratinocytes

a number of modifications in the genetic sequences included in packaging cell lines has greatly decreased the risk of recombination events, and sensitive methods for detecting replication-competent virus have been established and are strictly utilized

in all clinical trials There have been no documented adverse events related to inser-tional mutagenesis in early human clinical studies or in preclinical animal studies using replication-defective viral vectors

There appear to be two major limitations to the use of MuLV vectors for hematopoietic stem cell transduction First, cells must pass through the mitotic phase of the cell cycle in order for the vector to gain access to the chromatin and integrate (Fig 6.2) Most stem cells reside in the G0 phase of the cell cycle, and manipulations that stimulate these cells to cycle ex vivo may result in irreversible lineage commitment or apoptosis Second, the receptor for MuLV retroviral vectors (amphotropic vectors) on human and primate cells has been identified and appears

to be broadly expressed in most human tissues However, the low levels of this receptor on primitive HSCs may be limiting To redirect receptor specificity, pseudo-typing of vectors has been employed by replacement of MuLV envelope proteins with gibbon ape leukemia virus (GALV) envelope proteins This technique improves transduction efficiency of mature lymphocytes and possibly hematopoi-etic stem cells The vesicular stomatitis virus (VSV) envelope protein allows direct membrane fusion, circumventing the need for a specific cell surface receptor, but toxicity of the envelope protein to both producer cell lines and target cells hinders development of this approach

vectors based on lentiviruses such as the human immunodeficiency virus (HIV)-1

or 2 Certain characteristics of HIV may overcome some of the limitations of the MuLV vectors Pseudotyping of HIV-based vectors with VSV or amphotropic enve-lope proteins would allow transduction of hematopoietic progenitor and stem cells Use of the HIV envelope gene would allow specific transduction of CD4+targets HIV and other lentiviruses transduce target cells without the need for cell division The mechanism for this property is not fully understood But, the dissection of the HIV genome and incorporation of the nuclear transport mechanism(s) into other-wise standard MuLV vectors for gene therapy has not been successful Beyond these

FIGURE 6.2 Importance of cellular activation by growth factors or cytokines to induce mitosis for transduction by Moloney murine leukemia virus (MuLV) Cells must pass through the mitotic phase of the cell cycle (M, middle frame) in order for the vector to gain access

to the chromatin and integrate into the genome (right frame).

efforts, there are obviously major safety concerns that preclude clinical applications

of HIV Absolutely convincing preclinical data regarding efficacy and lack of replication-competent virus must be obtained prior to human use Non-HIV-1 lentiviral vectors are also of great interest and are very early in development, as are vectors based on the human foamy virus (HFV), another retrovirus that appears to have little pathogenicity

For adenoassociated virus (AAV), utility in hematopoietic stem cell gene transfer

is unlikely However, applications requiring only transient expression in lymphocytes

or dendritic cells are attractive Most recently, promising data has been obtained using AAV to transduce muscle cells in vivo, allowing prolonged production of soluble factors important in hematologic diseases such as factor IX for hemophilia

or erythropoietin for anemia of chronic renal failure AAV vectors package 5.2 kb of new genetic material precluding the transfer of large genes such as factor VIII Adenovirus (Ad) vectors have been explored primarily for in vivo gene delivery for the transfection of both dividing and nondividing cells The immune response induced by Ad vectors, although a major disadvantage, is also being considered as

a possible advantage for transduction of tumor cells with cytokines, co-stimulatory molecules, or other immune modulators in cancer vaccine protocols (see Chapter 13) These applications, thoroughly investigated in solid tumor animal models, are also being applied to hematologic malignancies such as leukemias and lymphomas Normal primitive hematopoietic cells can be transduced by Ad, but only with very highly concentrated vector preparations that also result in significant toxicity Transient expression in primitive cells may be of interest in manipulating homing after transplantation

The simplest approach to gene transfer is to use naked plasmid deoxyribonucleic acid (DNA), with necessary control sequences and the transgene, as the vector The advantages of nonviral vectors include the lack of any risk of generation of replication-competent infectious particles, independence from target cell cycling during transduction, and elimination of antivector immune response induced by viral proteins There are few size constraints However, transduction efficiency of primary cells is very low, and physical methods such as electroporation or chemical shock used to increase gene transfer efficiency of plasmids into cell lines are either inefficient or toxic Encapsulation by lipsomes has been useful for some primary cell types, as has conjugation to molecular conjugates including polyamines and inacti-vated adenovirus However, none of these nonviral methods has shown any promise

in the transduction of hematopoietic stem or progenitor cells Limited success has been reported transducing primary human lymphocytes with a device called the

“gene gun,” introducing plasmid DNA into cells using colloid gold particles None

of these vectors integrate, and expression levels are generally lower than reported with viral vectors

HEMATOPOIETIC STEM AND PROGENITOR CELLS AS TARGETS FOR GENE THERAPY

The concept of genetic correction or modification of HSCs has been an ongoing primary focus of gene therapy research The properties of both self-renewal and dif-ferentiation of HSC can provide for the continuous maintenance of the transgene

in cells of hematopoetic origin, including red blood cells, platelets, neutrophils, and

lymphocytes Less obvious are the application to tissue macrophages, dendritic cells, and central nervous system microglial cells (Chapter 9) Lineage-specific control ele-ments need to be included to allow for differential expression in the appropriate mature cell type; for example, the use of hemoglobin gene enhancers to target expression to red blood cells The genetic correction of these cells offer a potential curative, one-time therapy for a wide variety of congenital disorders such as hemoglobinopathies, immunodeficiencies, or metabolic storage diseases Gene therapy also allows consideration of novel approaches to malignancies and HIV infection such as differential chemoprotection and intracellular immunization (see Chapter 11)

The feasibility of harvesting transplantable stem cells from the bone marrow (BM) and the maintenance in short-term ex vivo cell culture were a crucial advan-tages in early animal studies The discovery and isolation of hematopoietic cytokines

in the mid-1980s allowed successful ex vivo culture and transduction, resulting in the first successful demonstration of efficient gene transfer into murine repopulating stem cells More recently, the discovery of alternative sources of stem cells such as mobilized PB and umbilical cord blood (UCB) broadens the potential for HSC gene therapy to neonates or conditions requiring very high dose stem cell reinfusion However, several obstacles have limited progress toward efficient gene transfer into HSCs Some are methodologic No in vitro assays exist to identify and quanti-tate true human stem cells Further, gene transfer strategies efficient in transduction

of in vitro surrogates, such as day 14 colony forming units (CFU) or the primitive multipotential long-term culture initiating cells (LTCIC), have not resulted in similar high levels of transduction of actual repopulating cells in early clinical trials or large animal models Thus, optimization of protocols and testing of new approaches has been hampered An additional obstacle is the observation that the most primitive pluripotent hematopoietic cells appear to be predominantly in the quiescent G0phase of the cell cycle These cells are thus resistant to transduction with MuLV retroviral vectors (Fig 6.2) Attempts to increase cycling of primitive cells during transduction by prolonged culture in the presence of various combina-tions of hematopoietic cytokines has resulted in decreased engrafting ability This is due to either loss of self-renewal properties, induction of apoptosis, or alteration in homing ability Additionally, a characteristic of primative hematopoietic stem and progenitor cells that inhibits efficient gene transfer is the low level of expression

of receptors for a number of vectors including retroviruses and adenoassociated viruses Lastly, many clinical applications are in nonmalignant disease where the use

of high-dose ablative conditioning therapy prior to reinfusion of genetically cor-rected autologous stem cells is unacceptably toxic Only with the use of high doses

of stem cells can significant levels of engraftment occur without the use of high-dose conditioning chemotherapy or total body irradiation

Preclinical Studies

Initial retroviral gene transfer into murine hematopoietic repopulating cells was achieved in 1984 The discovery, availability, and application of various hematopoi-etic growth factors improved the efficiency of ex vivo retroviral transduction of murine hematopoietic cells Several different combinations of growth factors have been successfully used for supporting gene transfer into murine stem cells These

include the combination of interleukin 3 (IL-3), interleukin 6 (IL-6), and stem cell factor (SCF) Inclusion of recently discovered early acting growth factors such as

flt-3 ligand and megakaryocyte growth and development factor

(MGDF)/throm-bopoietin (TPO) have augmented the level of genetically modified cells These cytokines and growth factors maintain primitive cell physiology ex vivo and poten-tially stimulate primitive cells to cycle without differentiation They may also up-regulate retroviral cell surface receptors Other manipulations that have been found beneficial in the murine system include (1) treatment of animals with 5-fluorouracil before marrow harvest to stimulate cycling of primitive cells, (2) the co-culture of target cells directly on a layer of retroviral producer cells or other stromal support, (3) the use of high titer (greater than 105viral particles per ml) vector and (4) co-localization of vector and target cells using fibronectin-coated dishes

Under these enhanced conditions, retroviral gene transfer into murine BM hematopoietic cells is now achieved in vivo with long-term marking at 10 to 100%

in all cell lineages The persistence of vector sequences in short-lived granulocytes and in multiple-lineage hematopoietic cells from serially transplanted mice indicates that murine repopulating stem cells can be successfully modified with retroviral vectors Other supportive data include retroviral integration site analysis docu-menting the common transduced clones from different lineages The repopulation

of murine stem cells in nonablative or partially ablative conditioning transplant models has been increased by pretreatment of recipient mice with G-CSF/SCF These results in the murine model have raised concerns about long-term expres-sion of transgenes from integrated vectors Studies have shown poor or decreasing

in vivo expression of the transgene or transgenes, especially with serial transplants, despite persistence of vector sequences A hypothesis for this down-regulation in expression is the methylation of specific sequences in the vector promoter and enhancer regions To counter this down-regulation in gene expression, many modi-fications have been made in basic MuLV vectors These include the exchange of control sequences in the long terminal repeats (LTRs) with sequences from other retroviruses with lineage specificity of expression and the mutagenesis of putative negative regulatory sequences Data suggest that modified vectors show improved long-term in vivo expression, although, equivalent long-term expression from stan-dard MuLV vectors has been acheived under certain circumstances

Evaluation of ex vivo gene transfer protocols using human cells mainly relies on

in vitro progenitor cell assays, including CFU (representing committed progenitors), and long-term culture initiating cell (LTCIC), a putative in vitro stem cell surrogate Using similar optimized conditions to the murine model, 50% or more progenitor colonies were transduced by retroviral vectors Equally high LTCIC transduction has also been observed Although BM has been the traditional source for HSCs, optimized gene transfer into CFU or LTCIC indicates that mobilized PB and UCB can be sources for HSCs

Purification for primitive cells by panning—the exposure of whole BM or mobi-lized PB to antibodies directed against cell surface antigens found only on primi-tive cells, such as CD34—followed by flow cytometric sorting or immunoabsorp-tion results in the isolaimmunoabsorp-tion of approximately 1 to 5% of total cells These enriched progenitor cells have reconstituting properties in clinical transplantation protocols Selection for CD34+/CD38-or HLA-DR populations can further purify stem cells Recent studies show that CD34- cell populations also possess repopulating activity,

potentially arguing against the use of CD34-enriched cells for gene transfer and other applications Use of purified target cells permits practical culture volumes and higher vector particle to target cell ratios (MOI) during transduction, thereby increasing gene transfer efficiency

As data emerge suggesting that the use of in vitro surrogate assays do not predict levels of gene transfer seen in vivo in early human clinical trials, attention has refo-cused on studying in vivo repopulating cells One approach is the use of large animal models since the stem cell dynamics, cytokine responsiveness, and retroviral re-ceptor properties appear to be similar between humans and nonhuman primates However, very few research centers have the facilities and resources to carry out such transplant studies, and thus current studies are feasible as small proof of principle experiments, with little ability to study the impact of changing multiple variables Rhesus or cynamologous monkeys and baboons are currently used most extensively The persistence of vector sequences was first observed in a rhesus monkey transplantation model in 1989 In this seminal study, the CD34-enriched marrow cells were transduced with a high titer vector producer cell line (greater than 108–10 viral particles per ml) secreting both human IL-6 and gibbon IL-3 However, this high titer producer cell line also produced significant titers of replication-competent helper virus due to recombination between vector and helper sequences in the producer cell line Thus, in vivo marking in these animals could not

be interpreted Moreover, high-grade T-cell lymphomas were found in some re-cipients several months posttransplantation because of insertional mutagenesis by the replication-competent contaminating virus This complication resulted in wide agreement that it is absolutely necessary to use helper-free producer cell lines and vector stocks in any clinical application As well, it is necessary to assess safety in large animals before human clinical use

Subsequent studies have documented long-term genetic modification of multiple hematopoietic lineages in primates using a number of different helper-free retrovi-ral vectors These successful transductions have been performed in the presence of growth factors, using unpurified or CD34-enriched BM or mobilized PB cells Lower levels of gene-modified circulating cells were reported when compared to the mouse model (generally less than 0.01 to 1%), although similar optimized transduction con-ditions were used in both systems Improved marking levels of up to 1 to 4% have been reported by transducing growth factor-stimulated PB or BM hematopoietic cells in the presence of a cell line engineered to express a transmembrane form of

human SCF Recently, studies report further encouraging data when flt-3 ligand

is added to the transduction cytokine combination, either in the presence of a fibronectin support surface or autologous stroma Marking levels of 10 to 20% in vivo for at least 20 weeks were confirmed by Southern blotting

Some important results of retroviral transduction were obtained from the canine autologous transplantation model For instance, effective transduction of G-CSF-mobilized peripheral blood repopulating cells was first observed in the dog Par-tially or fully ablative conditioning was necessary to obtain detectable engraftment with transduced HSCs Using this model, high levels (up to 10%) of transduced marrow CFU after transplantation have been reported using a 3-week long-term marrow culture for transduction and reinfusion without conditioning

The expense and difficulty of transplanting large animals have resulted in the transplantation of gene-modified human hematopoietic cells in immunodeficient mice as an alternative model The major obstacle of this method is the low-level

engraftment with human cells Improved results have been obtained by inclusion of co-transplantation of stromal cells secreting human IL-3, the use of more immun-odeficient strains such as NOD/SCID, and transplantation into immunimmun-odeficient transgenic mice expressing human cytokines Identical retroviral integration sites were documented in human myeloid and T-cell clones obtained from a mouse posttransplantation, suggesting that pluripotent human HSCs were transduced Cord blood cells engraft with greater efficiency than adult BM or mobilized PB Thus studies have employ CB to a greater extent and extrapolate the data to other cell sources for gene therapy The predictive value of data derived from xenograft models remains to be proven through the direct comparison with results from human clinical studies, thereby tracking the same gene-modified cell population in both patients and immunodeficient mice

Clinical Genetic Marking Studies

Genetic marking of cells with an integrating vector is a unique method for tracking autologous transplanted cells and their progeny in vivo Early human clinical gene transfer trials used retroviral vectors carrying nontherapeutic marker genes to duce a fraction of an autologous graft in patients undergoing autologous trans-plantation for an underlying malignancy These studies provided proof of principle and safety data

Several studies used retroviral marking to track whether reinfused tumor cells contribute to relapse after autologous transplantation In two pediatric gene-marking studies, unpurged autologous marrow from children with acute myeloid leukemia or neuroblastoma was briefly exposed to a retroviral vector carrying the

Neo gene Genetically marked tumor cells were detected in several patients at

relapse This observation suggested that the reinfused marrow had contributed to progression and that purging was necessary One adult marking study did not detect marked tumor cells in patients with acute leukemia at relapse, but overall trans-duction efficiencies in this study were lower Marked relapses were demonstrated

in chronic myelogenous leukemia: bcr/abl+ marrow CFU-C were shown to contain

the marker gene No marked relapses have been detected in adult patients with mul-tiple myeloma and breast cancer transplanted with genetically marked bone marrow and peripheral blood cells However, the marrow and blood cells were CD34-enriched before transduction, thus purging the starting population by at least 2 logs

of tumor cells

Another outcome of these marking studies was to assess in vivo gene transfer efficiency In the pediatric study, a fraction of the bone marrow graft was briefly exposed to retroviral supernatant without growth factors or autologous stroma As many as 5 to 20% of marrow CFU were shown to be neomycin-resistant between

6 and 18 months posttransplantation, suggesting effective transduction and ongoing transgene expression This surprisingly high level of stable marked marrow prog-enitors may be explained in part by active cell cycle kinetics of the primitive HSCs from these children likely due to their young age Additionally, the primitive HSCs may have been undergoing hematopoietic recovery from high-dose chemotherapy just before BM collection However, only 0.1 to 1% of circulating mature cells were marked

Treated adults have undergone autologous bone marrow and mobilized periph-eral blood stem cell transplantation for multiple myeloma and breast cancer Bone

marrow and peripheral blood CD34-enriched cells were transduced with different

retroviral vectors containing the Neo gene in order to assess the relative

contribu-tion to marking and engraftment of marrow and peripheral blood populacontribu-tions Transduction was performed for 3 days in the presence of the cytokines IL-3, IL-6, and SCF Circulating marked cells were detected after engraftment in all patients Marked cells were also detected in three of nine recipients for over 18 months Although granulocytes, B cells, and T cells were positive for the transgene, the gene transfer efficiency was lower than in the pediatric studies Less than 0.1% of circu-lating cells were marked long term, and no high-level marking of marrow CFU-C was detected Because both the bone marrow and peripheral blood grafts con-tributed to long-term marking, this study documented that mobilized peripheral blood grafts can produce multilineage engraftment This study was also important evidence that allogeneic transplantation could be performed safely with this cell source These investigators also tested the brief single transduction protocol that was effective in the pediatric study, but no persistent marking was detected in adult patients

Clinical Studies Using Therapeutic Genes

A main objective of gene therapy is the replacement of defective or missing genes

in congenital diseases A number of single-gene disorders such as the hemoglo-binopathies, Fanconi anemia, chronic granulomatous disease, and Gaucher disease have been the focus of clinical trials The hematological deficiencies in these disor-ders can be successfully treated by allogeneic BMT, implying that normal stem cells can reverse the pathophysiology of the disorders Despite the low level of gene transfer into long-term repopulating stem cells achieved in large animal models and early human marking studies, several clinical trials exploring potentially therapeu-tic genes have been reported or are ongoing (Table 6.2) Important information has been obtained on safety and feasibility of stem cell engraftment without ablation, and there are glimmers of hope regarding clinical benefit

Severe combined immunodeficiency due to adenosine deaminase (ADA) muta-tions was the first disease involving gene therapy of hematopoietic cells for several reasons The human ADA gene was cloned in the early 1980s and the small 1.5-kb (cDNA) could easily fit into a retroviral vector along a selectable marker gene such

as Neo Even a low level of gene transfer efficiency might be efficacious because

ADA normal cells should have an in vivo survival and proliferative advantage Thus, the correction of only 1 to 5% of target cells may have clinical benefit Hematopoi-etic stem cells could be better gene correction targets than T cells in this and other immunodeficiency disorders because of the potential for permanent and complete reconstitution of the T-cell repertoire However, it has been difficult to achieve stable long-term efficient transduction of HSCs, thus T cells were the initial targets chosen To directly address this issue, two ADA-deficient children in Italy received both autologous bone marrow and mature T lymphocytes transduced with

distin-guishable retroviral vectors carrying both the ADA and Neo genes The patients

were then repeatedly reinfused with both cell products without conditioning In the first year, vector-containing T cells originated from the transduced mature T cells; but, with time, there was a shift to vector-containing T cells originating from transduced bone marrow cells A normalization of the immune repertoire and