Iuly 2002
`
`7
`
`'
`
`7“
`
`a
`
`the earliest known hominid
`
`‘ g‘lnternK‘tlionalWeeklyjour fiafsgence. www.nature.com/nature
`anti”:
`
`‘VNJM‘AM.
`a“ or)
`Jr"
`
`,aI
`
`' v.‘
`i
`7)“;
`
`
`
`00/001l0132
`
`5-Dé“:
`ORIALLIBRARY
`
`418/02
`06-1201
`is?
`
`éu
`AD
`
`I.|.I.nl|.l...lllu..Aim;“5319EN
`
`RNA
`-
`Nature Insught
`....................................................................................................................................
`
`‘
`
`,w ,,
`
`....................................................................................................................................
`
`...................................................................................................................................
`
`Catalytic converters The road to longer life
`
`naturejobs
`‘r
`:1 N; 1th n m medicime
`
`
`
`1
`
`MTX1039
`
`

`

`insight review articles
`
`
`rm: material may be protected by Copyright law [Title 17 us. cm)
`
`
`
`
`Emerging clinical applications of RNA
`
`Bruce A. Sullenger & Eli Gilboa
`
`Department ofSurgery, Duke University Medical Center, Durham, North Carolina 27710 USA (e—mail' 11 5ullenger@cgct dukeedu)
`
`RNA is a versatile biological macromolecule that is crucial in mobilizing and interpreting our genetic
`information. It is not surprising then that researchers have sought to exploit the inherent properties of RNAs
`so as to interfere with or repair dysfunctional nucleic acids or proteins and to stimulate the production of
`therapeutic gene products in a variety of pathological situations. The first generation of the resulting RNA
`therapeutics are now being evaluated in clinical trials, raising significant interest in this emerging area of
`medical research.
`
`
`
`
`as
`he concept of using RNA molecules
`therapeutic agents is relatively new, but has
`received increasing attention during the past
`decade. Much of this interest stems from a
`
`that
`variety of basic scientific discoveries
`underscore the seminal role of RNA molecules in the
`
`utilization of genetic instructions in all living systems and
`the versatility of these molecules in nature. RNA molecules
`can adopt a wide variety of conformations and perform a
`range of cellular functions. Certain RNAs fold to form
`catalytic centres, whereas others have structures that allow
`them to make specific RNA—RNA, RNA—DNA or
`RNA—protein interactions.
`Such realizations have led translational researchers to
`
`attempt to exploit various facets of RNA biology and
`chemistry to combat human disease. Significant progress
`has been made towards this goal and the first RNA—based
`therapeutics are now being evaluated in clinical trials for the
`treatment of disorders ranging from cancer to infectious
`diseases. The therapeutic RNAs that have so far received the
`most attention can be grouped into four categories: gene
`inhibitors,
`gene
`amenders, protein inhibitors
`and
`immunostimulatory RNAs. Here we review the develop—
`ment of these various RNA-based approaches to therapy
`and provide an update on the progress towards moving this
`emerging class of molecular therapeutics into and through
`clinical evaluation. Finally, we will discuss the developmen-
`tal difficulties that various RNA therapeutics still face and
`consider potential solutions to those problems.
`
`RNA~mediated inhibition of gene expression
`Regulation of gene expression by an RNA that is comple-
`mentary to a target messenger RNA (mRNA) was first
`recognized as a naturally occurring process in prokaryotes‘.
`These complementary RNAs,
`termed antisense RNAs,
`specifically recognize their target transcripts by forming
`base pairs with them in a sequence-dependent manner. The
`formation of this RNA duplex is believed to lead to the
`degradation of the target RNA or the inhibition of its trans-
`lation. The ability to inhibit specific genes after gene transfer
`of antisense expression cassettes was first demonstrated
`almost two decades ago in bacteria by Pestka etal.2 and Cole-
`man et al.3 and in eukaryotic cells by Izant and Wientraub".
`Following these initial studies, numerous reports appeared
`that described the potential utility of antisense RNA for the
`inhibition of a wide array of genes in mammalian cellss.
`However, these studies also indicated that the efficacy of
`antisense—mediated gene inhibition was usually dependent
`on the presence ofa considerable excess ofantisense RNA to
`
`252
`
`2
`
`target RNA in the cells. Therefore methods were sought to
`express large quantities of antisense RNAs in transduced
`cells6 or to generate antisense RNAs that could destroy
`multiple target RNAs and thus reduce the need for an excess
`ofthe inhibitory RNA.
`The discovery that certain RNAs can perform catalysis”8
`has led to the development of a class of therapeutic RNAs
`called trans-cleaving ribozymes. Such ribozymes bind
`substrate RNAs through base—pairing interactions, cleave
`the bound target RNA, release the cleavage products and are
`recycled so that they can repeat this process multiple times in
`vitro (Fig. 1a). The observation that such ribozymes can be
`repeatedly targeted to cleave Virtually any pathogenic
`transcript in vitrog‘lo led to much speculation about their
`potential therapeutic value in viva”’”. Much progress has
`been made towards assessing the potential utility of trans-
`cleaving ribozymes, with the hammerhead and hairpin
`ribozyme13 being the main focus of this translational effort
`(see review in this issue by Doudna and Cech, pages
`222—228).
`Several phase I and II clinical trials have been initiated
`using trans—cleaving ribozymes in a small number of
`patients with infectious diseases or cancer. In these studies
`the ribozymes have been delivered to the patients either by
`gene therapy methods or by direct injection of a synthetic
`ribozyme. The gene therapy—based trials have focused upon
`developing ribozyme-based treatments for
`individuals
`infected with the human immunodeficiency virus (HIV).
`Three separate groups have used retroviral vectors to intro-
`duce expression cassettes for anti-HIV ribozymes into CD4+
`lymphocytes or CD34+ haematopoietic precursors ex vivo
`that have been taken from the infected patient or from an
`identical twin”‘” (Fig. 1b). The transduced cells are then
`infused into the patient and the engraftment and survival of
`the ribozyme-containing cells are monitored.
`Initial results from these studies suggest that transfer of
`ribozyme-encoding genes to HIV-infected individuals is
`well
`tolerated and transduced cells can persist
`in the
`patient”. Moreover, preliminary reports suggest
`that
`anti-HIV ribozyme—containing cells may possess a transient
`survival advantage in the patient compared with cells trans-
`duced with a control vector”. Unfortunately, such studies
`also indicate that gene transfer into long-term progenitors
`has not been accomplished because transduced cells
`decrease to below detection by one year after infusion (I. I.
`Rossi, personal communication). Larger clinical trials must
`now be performed to evaluate the efficacy of anti-HIV
`ribozymes. Critical factors that will influence the success of
`such trials will be the development of gene-transfer systems
`NATURE | VOL 418 | 11 JULY 2002 [ www.nature.com/nature
`
`
`
`2
`
`

`

`
`
`
`
`
`
`
`
`
`
`a
`
`Pathogenic mRNA—
`Binding of target FINN
`Trans-cleaving ribozyme
`
`Cleavage of
`target RNAv
`Product release
`
`5
`
`Infuse
`transduced
`cells
`
`
`Patienw
`
`
`follow up
`
`a
`
`Gene m
`transfer
`Ribozyme or
`decoy vector
`
`Control vector
`
`
`
`Isolate CD4+
`or CD34‘ cells
`
`
`
`
`
`
`:mwwfifim’geneifimer
`,
`
`mammmmemare mmoduwdlnmihehatlm ‘
`
`
`
`
`
`intravenous or subcutaneous delivery of this compound is well
`tolerated and that plasma levels could be maintained for prolonged
`periods after subcutaneous delivery”. Currently its therapeutic
`efficacy is being evaluated in phase II trials for breast and colorectal
`cancer. The other two synthetic ribozymes to enter clinical trials tar-
`get mRNA ofhuman epidermal growth factor receptor type 2, which
`is overexpressed in many breast cancers, and hepatitis C virus (HCV)
`RNA, which is associated with liver cirrhosis and hepatocellular car-
`cinoma. Results from these initial efficacy studies will provide the
`first significant insight into the long—term utility of trans-cleaving
`ribozymes as therapeutic agents.
`The critical factors that will most likely determine the success of
`these synthetic ribozyme efficacy trials are the ability to deliver
`ribozymes efficiently into the appropriate cells in vivo, and the level
`and duration oftarget-gene inhibition that is required to alter disease
`pathophysiology and slow disease progression. For ribozymes to
`benefit individuals with chronic disorders such as cancer and HCV
`(or HIV)
`infection,
`long-term, high-level
`inhibition of target
`transcripts will probablybe required. This maybe difficult to achieve
`in practice, especially when targeting highly expressed viral RNAs.
`For example, we observed that a hammerhead ribozyme was able to
`inhibit the replication of a murine retrovirus by up to 90% by
`co-localizing the ribozyme with its Viral target inside cells, but we
`found that higher levels of inhibition were difficult to achieve even
`when a large excess of the ribozyme was present in the cells”. Thus,
`the utility of trans—cleaving ribozymes may be limited to conditions
`were a modest reduction in pathogenic gene expression will result in
`therapeutic efficacy or to combination therapies were the ribozyme
`can act in concert with other therapeutics (for example, antiviral or
`chemotherapeutic agents) that largely control the pathogen.
`Finally, two additional RNA—based strategies for gene inhibition
`in mammalian cells have recently been described. First, in many
`eukaryotic cells, expression of an mRNA can be inhibited by a
`double-stranded RNA that corresponds to the sequence ofthe target-
`ed transcript. This process, known as RNA interference (RNAi; see
`review in this issue by Hannon, pages 244—251), has been shown to
`function in mammalian cells after introduction of short (~21
`nucleotides), synthetic duplex RNAs23 or expression of similar—24— 7
`length transcripts by RNA polymerase 111 (Pol III) promoters
`.
`Using this Pol III expression approach, Lee and colleagues“ have
`shown that RNAi can inhibit HIV gene expression by up_to 4 logs in
`co—transfection experiments. These and other results have raised
`much interest about the potential utility of the RNAi approach for
`therapeutic applications.
`Second, in contrast to trans—cleaving ribozymes and RNAi, which
`target pathogenic RNA for destruction, mobile group II introns can
`be targeted to insert into and inactivate pathogenic DNA. The
`obvious advantage of targeting DNA rather than RNA is that such
`disruption would be a one-time event that would effectively knock
`out 100% of the RNAs issuing from the disrupted gene. Lambowitz
`and colleagues have shown2a that a mobile group II intron can be
`retargeted to insert specifically into HIV pro—viral DNA or the HIV
`co—receptor gene CCR5. Gene disruption was efficient in bacteria
`where the HIV pol gene could be disrupted in over 60% of the cells.
`Group II mobilization into extra-chromosomal copies ofthe HIVpol
`and CCR genes in mammalian cells has been shown to be possible",
`but mobilization into genomic DNA has yet
`to be reported.
`Translational researchers must now begin to determine whether
`these two gene-inhibition strategies can be utilized for therapeutic
`applications. Information gleaned from the ongoing clinical trials
`evaluating trans-cleaving ribozymes will undoubtedly facilitate the
`clinical development of this next generation of RNA—based gene
`inhibitors.
`
`insight review articles#
`
`
`
`that can efficiently transduce pluripotent haematopoietic stem cells
`and the generation of improved ribozyme expression cassettes that
`can increase the survival advantage oftransduced cells. In this regard,
`encouraging preclinical studies suggest
`that co—localization of
`ribozymes with their viral target RNAs inside cells may enhance
`ribozyme activity in vivo’7‘18, and use of combinations of inhibitory
`ribozymes and decoy RNAs may yield more potent inhibitors ofHIV
`against which it will be difficult for the virus to develop resistance.
`Three different nuclease-resistant synthetic ribozymes are being
`evaluated in clinical trials”. Each of these trials uses a hammerhead
`ribozyme derivative that contains chemical modifications that great-
`ly increase the ribozyme’s stability in biological fluids“). Moreover,
`methods have been developed that enable large-scale synthesis ofthis
`new class of therapeutic agent under good manufacturing practice
`protocols”). All three ofthese synthetic ribozymes target RNAs whose
`expression is associated with the induction or progression of cancer
`and all three have shown promising results in preclinical cell and
`animal experimentszm.
`In 1998, the first ofthese ribozymes entered a phase I trial target-
`ing flt-1 mRNA, which encodes the high-affinity receptor for the
`angiogenic protein vascular endothelial growth factor (VEGF).
`Results from this and two subsequent phase I trials show that daily
`
`NATURE l VOL 418 l 11 IULY 2002 l www.nature.com/nature
`
`RNA-mediated repair of genetic instructions
`Genetic instructions are usually revised as they are converted from
`DNA to RNA to protein in human cells. Most ofthis revision occurs at
`253
`
`3
`
`3
`
`

`

`insight review articles
`
`the RNA level when splicing removes intron sequences from
`precursor transcripts and ligates together flanking exon sequences to
`generate mature RNAs. Interestingly, manyofthe molecules involved
`in revising RNA messages are RNAs themselves. Several recent stud-
`ies exploit this facet of RNA biology and describe the development of
`an intriguing new class of therapeutic RNAs that can perform trans—
`splicing to repair clinically relevant mutant transcripts. The concept
`of RNA repair has received much attention as a novel approach to
`gene therapy. Compared with more traditional strategies, RNA
`repair should engender the regulated expression of the corrected
`gene product and simultaneously reduce the expression of the
`mutant gene product. This property makes RNA repair an attractive
`strategy to attempt in the treatment of genetic disorders associated
`with mutant genes that are highly regulated or that encode deleteri-
`ous mutant proteins.
`The initial studies focused on RNA repair used a trans-splicing
`version of a group I ribozyme to repair mutant lacZ transcripts in
`bacteria29 and mammalian cells”. These studies showed that the
`
`ribozyme was able to repair the mutant RNA by recognizing the tar-
`get transcript by base pairing with it, cleaving off mutant sequences
`and ligating a wild-type sequence onto the cleavage product (Fig. 2a).
`More recently, trans-splicing ribozymes have been generated that can
`amend mutant transcripts associated with myotonic dystrophy31 and
`many cancers (mutant p53 tumour-suppressor transcripts)32 in
`mammalian cell lines, and with sickle cell anaemia in erythrocyte
`precursors isolated from patients with sickle cell disease”.
`A second, related approach to RNA repair uses spliceosomes to
`revise mutant transcripts by trans-splicing“. In this case the spliceo-
`some performs the trans—splicing reaction upon a pre- trans—splicing
`RNA molecule (termed a PTM) and a target RNA (Fig. 2b). An
`expression cassette for the PTM is delivered to the cell, whereas the
`spliceosome and target RNA are supplied by the cell. Puttaraju et al.
`first demonstrated that spliceosome—mediated RNA trans-splicing
`(SMaRT)
`could be used to reprogram human chorionic
`gonadotropin B—polypeptide mRNAs and to repair mutant lacZ
`transcripts in cell culture”. Subsequently, encouraging results have
`shown that the SMaRT approach can repair a clinically relevant
`fraction of a mutant cystic fibrosis transmembrane conductance
`regulator (CFTR) transcript (CFTRAF508) in human cystic fibrosis
`airway-epithelial cells grown in culture or in animal xenografts“.
`Such repair resulted in partial restoration of Cl" transport in the
`CFTR—deficient cells to 12—15% of the level observed in wild—type
`CFTR—containing cells.
`A critical question requiring further study for both ribozyme—
`and spliceosome-mediated repair of mutant RNAs is the specificity
`ofRNA revision. The specificity ofspliceosomal trans—splicing seems
`to be low, at least in certain instances, with the spliceosome trans-
`splicing the PTM sequence onto many unintended target RNAs in
`mammalian cells”. Initial studies suggested that the specificity of
`ribozyme-mediated repair may also be low”; but more recent
`investigations have described modifications to the ribozyme to
`address the issue oftarget specificity directly32‘38“°. Definitive studies
`are now warranted to address this question directly and to facilitate
`the development of more specific trans—splicing agents if needed.
`Even though some issues concerning RNA repair remain unresolved,
`the preclinical studies performed so far on therapeutic applications
`of trans-splicing are encouraging. They demonstrate that trans-
`splicing can amend mutant transcripts associated with a variety of
`human diseases and repair target RNAs with efficiencies that would
`be expected to be clinically beneficial, at least for treating many
`recessive disorders.
`
`
`
`RNA as a protein antagonist
`Many small RNAs can fold into three-dimensional structures that
`allow them to bind target proteins with high affinity and specificity.
`Several RNA viruses such as HIV use this property of RNA to
`recruit viral and host proteins to perform essential functions in Viral
`
`254
`
`4
`
`a
`
`Mutant transcript
`
`Ribozyme with
`corrective 3’ exon
`.
`
`—:_N'N’N’N'N’u zX_
`NNNNNG 3’exon
`Ribozyme
`
`
`Substrate
`binding
`
`
`1 target transcript
`
`
`Ribozyme—catalysed
`cleavage of
`
`Ligation of corrective
`RNA sequence onto
`cleaved transcript
`n N’N'N’N'N’u ‘XF
`Corrected RNA
`
`
`cis-
`splicing
`
`
`
`trans-
`splicing
`
`
`
`
`Corrected RNA
`
`Figure 2 Tmns-splicing-mediated repair of mutanttranscripts. a, Ribozyme-mediated
`repair Trans—splicing ribozymes recognize mutantRNAs upstream of a mutation site
`(Xm). The mutant RNA is cleaved and an exon with a wild—type sequence (th) Is
`ligated onto the cleavage productto generate a corrected transcript. h, Spliceosome-
`mediated RNA trans-splicing repairof a mutantiranscript. A target RNAcontaining a
`mutation In itsthird exon Gilmnormaiiy undergoes’cisispiicingto generatean mRNA
`matemwesademcfiwgempmduct.AWm$ngNnmdawb (PTM,red)
`with award-type exon 3 (th) can impedethe assessing processand engender
`trans-splicing to yield an mRNAwith acorrectsequenCe.
`’
`‘
`~
`
`replication. For example, HIV uses small-structured RNA elements
`termed the trans—activation response region (TAR) and Rev response
`element (RRE) to recruit the viral regulatory proteins Tat and Rev to
`control viral gene expression. The use of small—structured RNAs to
`directly bind and inhibit the activity ofa pathogenic protein was first
`explored using the HIV TAR sequence (Fig. 3a). Expression of TAR
`‘decoy’ RNAs in CD4+ T cells was shown to competitively inhibit Tat
`binding to the viral TAR RNA and render cells highly resistant to HIV
`replication (Fig. 3b)“. Subsequent studies demonstrated that RRE
`also could act as a decoy RNA to block Rev activity and inhibit
`HIV replication”.
`These observations suggested that short RNA decoys might be
`useful therapeutic agents for inhibiting HIV replication in vivo and
`have led to a phase I gene therapy-based clinical trial designed to test
`safety and feasibility. In this trial, retroviral vectors were used to
`NATURE I VOL 418 I l 1 JULY 2002 I www.nature.com/nature
`
`
`
`4
`
`

`

`
`
`Viral replication
`vTAR RNA
`
`b
`
`Control
`
`insight review articlesf
`
`Day 6
`
`Day 10
`
`Day 17
`
`Day 21
`
`
`
`Tat protein
`
`Viral transcription
`
`
`No Replication
`
`TAFidecoy @
`
`No trans-activation
`
`
`
`TAR I.-
`
`
`
`incubi/ Targetprotein
`
`RNA library
`
`
`Partition-
`
`bound RNA
`
`Repeat proces/v
`
`- I
`I - Amplifyselected RNA I]
`
`I]
`
`spread through the cultures was monitored by immunofluoresCent staining of cells
`(yellow cells) at various days iollo‘wing infection. Adaptedwith permission from ref. 41 .
`c. isolation of RNA aptamers that bind target proteins uslngthe SELEX (systematic
`evolution of ligands by exponential enrichment) process. After several rounds of
`selection. the RNAs remaining in the selected pool are cloned and sequenced to
`identity the high-affinity RNA aptamers for the target protein of interest.
`
`ir Eh g i
`
`,
`'5
`
`J
`
`
`
`
`
`Figure 3 RNA ligand-mediated inhibition of protein function. a, Trans-activation
`response region (TAR) decoy‘mediated inhibition at HN. During HIV replication, the viral
`Tat protein (blue) binds the viral TAR RNA (vTAR; green) and trans-activates viral gene
`emession and replication (top). TAR decoy RNAs (red) can compete for Tat binding and
`oornpetitively inhibit the Tat—VTAR Interaction and stop viral trans-activation and
`replication (bottom). I). Expression of TAR decoys can render cells resistant to HIV. TAR
`decoy- and control vector-containing CD4‘ T cells were challenged with HIV-1 and viral
`
`introduce expression cassettes for RRE decoys into haematopoietic
`progenitors that had been isolated from, and re-infused into, HIV-
`infected paediatric patients“3 (Fig. 1b). This initial study showed that
`gene transfer and expression of RRE decoys was well tolerated, but as
`with other HIV gene-therapy trials, the results emphasize the need
`for improved gene-transfer techniques to transduce pluripotent
`haematopoietic progenitors43.
`The observation that TAR and RRE decoy RNAs could be used to
`competitively inhibit viral protein function and replication suggest—
`ed that other small-structured RNA molecules might be able to bind
`pathogenic target proteins and inhibit their activity. Concurrentwith
`this observation, work from two groups suggested that iterative in
`vitro selection methods could be used to isolate high-affinity RNA
`ligands from large pools of randomized RNA sequences (vast RNA
`shape libraries) that could bind to proteins and small molecules“‘”.
`The resulting RNA ligands were termed aptamers by Ellington and
`Szostak45 and the selection process was named SELEX (systematic
`evolution ofligands byexponential enrichment) byTuerk and Gold“
`(Fig. 3c). SELEX has now been used to identify RNA aptamers that
`can bind to and inhibit the activity of a wide variety of proteins (for
`reviews, see refs 46, 47). The affinities ofaptamers for their targets are
`similar to the affinities achieved by monoclonal antibodies for their
`antigens (Kd values typically in the low nanomolar to high picomolar
`range)“. In contrast
`to antibodies, however, aptamers can be
`NATURE I VOL 418 | 11 JULY 2002 l www.maturecom/nature
`
`chemically synthesized to produce large quantities of these com-
`pounds for in viva experimentation and clinical trials. Moreover, as
`with trans-cleaving ribozymes, synthetic aptamers can be modified
`to have greatly enhanced plasma stability and circulating half-lives
`and seem to exhibit low toxicity and immunogenicity in ViV047—49.
`So far only a few aptamers have been evaluated in animal models
`of disease (reviewed in refs 47, 50) and two of these have received
`the most attention. A DNA aptamer to thrombin functions as a
`potent anticoagulant that is able to maintain the patency ofan extra-
`corporeal circuit in sheep and replace heparin in a canine cardiopul—
`monary bypass mode151'53. A 2’-flouro-modif1ed RNA aptamer to
`VEGF-165 is able to inhibit neovascularization in a rat corneal-
`pocket angiogenesis assay and block glomerular endothelial cell
`proliferation and apoptosrs in rats with mesangioproliferative
`nephritis‘g‘“. This VEGF aptamer is also the first therapeutic aptamer
`to be administered to humans and is currently being evaluated in a
`phase II/III clinical trial in patients with age—related macular degen—
`eration. In this instance, the synthetic aptamer is being injected
`directly into the vitreous humour ofthe eye to assess treatment safety
`and efficacy. Factors that are likely to influence the trial success
`include the retention time ofthe aptamer in the vitreous humour and
`the relative importance of VEGF—165 on the progression of age-
`related macular degeneration in these patients. The observation that
`the VEGF aptamer remains active and can be recovered from the
`
`5
`
`

`

`insight review articles
`
`~~mem4Tmlmentofcaneerpafientswm1
`
`NW
`NW
`WAmplified
`
`v..'Amplified
`\flex
`MonocytesI
`
`Tumour RNA
`loading
`
`tumour RNA
`
`DC maturation l
`
`W subsequénnntuslons
`lntethepatlem
`
`I
`
`
`
`
`
`
`
`cDNA
`
`
`
`
`
`
`C
`
`
`T
`
`
`
`u...~m,wmflu-n9—";st
`
`Mononuclear
`cells
`
`/ Leukapheresis
`
`vitreous humour of rhesus monkeys 28 days after injection55 is
`encouraging with regard to prospects for long-term ocular retention
`ofthe compound.
`Because most existing drugs are protein antagonists, the ultimate
`success of aptamers as therapeutic agents will probably depend on
`how well they compete with other classes of therapeutic compounds.
`In particular, the use ofmonoclonal antibodies has garnered increas-
`ing interest from both physicians and the pharmaceutical industry.
`Monoclonal antibodies are being generated and tested against many
`of the same proteins and for similar indications as most of the
`aptamers that are in pre-clinical development. Moreover, several
`antibodies have already progressed to market and are being used for
`the treatment ofa variety ofdisorders. Therefore, in this competitive
`landscape, properties of aptamers that distinguish them from
`monoclonal antibodies (and other protein inhibitors) will have to
`be exploited for aptamers to penetrate the therapeutic market and
`ultimately fulfil their potential and become broadly useful pharma-
`ceutical agents.
`
`lmmunolherapy using mRNA-transfected dendritic cells
`Specific active immunotherapy ofcancer —— stimulating the patient’s
`immune system to recognize and eliminate tumour cells — is emerg-
`ing as a promising modality for treating cancer recurrence and
`low—volume metastatic disease. There is considerable evidence that
`
`the cytotoxic T lymphocyte (CTL) arm of the immune response is
`crucial in controlling tumour growth. CTLs recognize short (8—10
`amino acids) antigenic peptides in association with major histocom-
`patibility complex (MHC) molecules displayed on the cell surface.
`The peptides are generated in the cytoplasm via the proteolytic action
`of the multi—unit proteolytic complex, the proteosome. They are
`shuttled from the cytoplasm to the endoplasmic reticulum, where
`they associate with nascent MHC molecules, and are then transport—
`ed to the cell surface for recognition by CTLS“.
`Naive CTLs generated in the thymus undergo an activation
`process to acquire the ability to kill their targets or secrete so-called
`‘effector’ cytokines such as interferon-y. Bone marrow-derived den—
`dritic cells (DCs) displaying the appropriate MHC—peptide complex
`on the cell surface are the primary cell type capable ofactivating naive
`CTL557. Once activated, the CTL can recognize and kill any somatic
`cell presenting the MHC—peptide complex. Thus, to stimulate a CTL
`response against the tumour, the tumour antigens have to reach the
`
`256
`
`6
`
`cytoplasm of a DC. In the cancer patient the capture of tumour
`antigens by DCs and stimulation of tumour—specific CTLs is pre-
`sumed to be inefficient and a limiting factor in stimulating protective
`immunity.
`One approach to stimulate effective CTL responses in cancer
`patients would be to reconstruct this process in vitro, that is, to isolate
`DCs from the patient, load them with tumour antigens in such a
`manner that the antigens reach the cytoplasm, and inject the antigen-
`loaded cells back into the patient. This approach has been
`accomplished by incubating DCs with peptides and proteins, or by
`transfecting the cells with DNA constructs. Transfecting DCs with
`mRNA-encoding (tumour) antigens is yet another way to load DCs
`with antigens“. mRNA can be isolated directly from tumour cells or
`synthesized in vitro from complementary DNA templates. DCs
`transfected with mRNA-encoding specific antigens or total tumour-
`derived RNA elicited potent CTL responses and tumour immunity in
`micess'“, and DCs generated from healthy volunteers or from cancer
`patients transfected with tumour RNA stimulated CTL responses in
`culture“'“‘".
`
`Initial studies have used cationic lipids to facilitate the uptake of
`RNA by DCsss'“. Remarkably, incubation of DCs with RNA alone
`was also sufficient to sensitize them to stimulate CTLs70—72'7‘. This is
`despite the low efficiency of RNA transfer (measured by gene expres-
`sion), and no doubt reflects the sensitivity of the immune system to
`recognize minute amounts ofantigens not detectable by convention-
`al means. Recently an improved method for DC transfection with
`RNA was developed using electroporation, which rivals the best
`transfection methods for nucleic acids“.69
`Just how efficient is the mRNA transfection protocol? Using
`functional end points such as CTL priming or induction of tumour
`immunity in mice, studies comparing several loading techniques —
`including loading DCs with peptides and proteins, transfection with
`cDNA plasmids or transduction with vaccinia vectors — found that
`mRNA loading was invariably superiorés‘fifi‘és‘ég. Additionally, it is
`often not fully appreciated that the generation of mRNA-encoding
`specific antigens is a simple process. Given the sequence, a cDNA
`template can be generated from the cell by reverse transcription
`followed by amplification using the polymerase chain reaction and
`transcribed into RNA in a matter of a few hours. This can be
`
`compared to the task of generating the corresponding protein or
`identifying the antigenic peptides.
`NATURE | VOL 418 I 11 JULY 2002 | www.nature.com/nature
`
`6
`
`

`

`Use of mRNA-encoded antigens for cancer vaccination offers
`another potentially useful benefit. Only a handful oftumour antigens
`
`have been discovered so far, mostly from melanoma patients, and
`there is growing evidence that many of these antigens are not well
`
`suited for vaccination75’76. The alternative option is to vaccinate with
`tumour-derived antigenic mixtures, an approach which animal
`
`studies suggest is surprisingly effective. In reality, however, it is not
`possible to obtain sufficient tumour tissue from most cancer patients
`yto generate the amount ofantigens thought necessary for an effective
`
`vaccination protocol. Consequently, a significant proportion
`(perhaps most) of cancer patients would not benefit from current
`
`vaccination strategies, regardless of how effective they might be. Use
`. omeNA as source ofantigen offers a solution that has been shown to
`
`workéz’“. Biologically active RNA can be amplified from microscopic
`amounts of tumour tissue (for instance, from frozen sections or
`from needle biopsies) to provide a virtually inexhaustible amount of
`
`antigen from practically every patient.
`Perhaps no less important is that use of mRNA—encoded antigens
`
`can address the concern associated with the emergence oftreatment—
`resistant tumour variants which lost
`the antigens targeted by
`
`vaccination, a common occurrence in cancer therapy. Using amplifi-
`cation protocols, sufficient RNA can be readily generated from a
`
`microscopic amount of a newly emerging antigen—loss tumour
`variant, before it is too late.
`
`In summary, the preclinical experience suggests that cancer
`
`vaccination with tumour RNA-transfected DCs may constitute a
`highly effective and broadly applicable treatment for patients with
`
`recurring cancer. Whether such observations can be translated to the
`
`clinic remains to be seen, but hints from initial clinical trials are not
`
`discouraging.
`The primary drawback of DC—based vaccination is that it is a cus-
`
`tomized form of cell therapy; DCs (and tumour RNA) are generated
`from each patient and require in vitro manipulation ofthe cells before
`re-injection into the patient (Fig. 4). This adds cost and complexity to
`
`the treatment and remains a challenge owing to the yet unproven
`nature of this new paradigm in human therapy. The expectation,
`which we believe is well founded, is that improved outcome to other—
`wise intractable diseases will offset the added cost and complexity
`
`associated with this form oftherapy. A phase I clinical trial in patients
`with metastatic prostate cancer vaccinated with prostate-specific
`antigen (PSA) RNA-transfected DCs (PSA is a common tumour-
`associated antigen in patients with prostate cancer) was recently
`concluded”. The DC therapy seems to be safe and, despite the
`advanced state of the disease, all patients have responded immuno-
`logically to the vaccine, that is, they have all exhibited PSA—specific
`T—cell responses. Surprisingly, six out of seven available patients
`exhibited a (very) modest clinically related response, a small but
`statistically significant impact on the blood PSA levels. This modest
`effect is unlikely to translate to clinical benefit to the pa