Vaccination with Messenger RNA (mRNA)

Steve Pascolo
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
2 Messenger RNA: Structure, Production, and Specic Optimizations for Vaccination
Purposes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
2.1 Messenger RNA for Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
2.2 Messenger RNA for Clinical Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
2.3 Optimization of mRNA for Vaccination. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
3 Methods of mRNA-Based Vaccination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
3.1 Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
3.2 Adjuvants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
4 Results of Clinical Trials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
4.1 Monocyte-Derived Dendritic Cells Transfected with mRNA In Vitro . . . . . . . . . . . 231
4.2 Direct Injection of mRNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
Abstract Both DNA and mRNA can be used as vehicles for gene therapy. Because
the immune system is naturally activated by foreign nucleic acids thanks to the pres-
ence of Toll-like Receptors (TLR) in endosomes (TLR3, 7, and 8 detect exogenous
RNA, while TLR9 can detect exogenous DNA), the delivery of foreign nucleic acids
usually induces an immune response directed against the encoded protein. Many
preclinical and clinical studies were performed using DNA-based experimental vac-
cines. However, no such products are yet approved for the human population. Mean-
while, the naturally transient and cytosolically active mRNA molecules are seen as
a possibly safer and more potent alternative to DNA for gene vaccination. Opti-
mized mRNA (improved for codon usage, stability, antigen-processing characteris-
tics of the encoded protein, etc.) were demonstrated to be potent gene vaccination
vehicles when delivered naked, in liposomes, coated on particles or transfected in
dendritic cells in vitro. Human clinical trials indicate that the delivery of mRNA
Steve Pascolo
Institut for Cell Biology, Department of Immunology, University of Tuebingen, Auf der Morgen-
stelle 15, 72076 Tuebingen, Germany
steve.pascolo@uni-tuebingen.de
S. Bauer, G. Hartmann (eds.), Toll-Like Receptors (TLRs) and Innate Immunity. 221
Handbook of Experimental Pharmacology 183.
c Springer-Verlag Berlin Heidelberg 2008
222 S. Pascolo
naked or transfected in dendritic cells induces the expected antigen-specic im-
mune response. Follow-up efcacy studies are on the way. Meanwhile, mRNA can
be produced in large amounts and GMP quality, allowing the further development of
mRNA-based therapies. This chapter describes the structure of mRNA, its possible
optimizations for immunization purposes, the different methods of delivery used in
preclinical studies, and nally the results of clinical trial where mRNA is the active
pharmaceutical ingredient of new innovative vaccines.
1 Introduction
The seminal article of Wolf et al. shows that naked minimal nucleic acid vectors in
the form of plasmid DNA (pDNA) or messenger RNA (mRNA) that code for a pro-
tein in an eukaryotic cell are spontaneously taken up and expressed in mouse mus-
cles (Wolff et al., 1990). Thus, in vivo injected foreign nucleic acids can somehow
penetrate in the cytosole (mRNA) and the nucleus (pDNA) of somatic cells before
being degraded by ubiquitous extracellular nucleases. This phenomenon is more
surprising for mRNA than for pDNA since the former is degraded within seconds in
contact to the abundant extracellular RNases (Probst et al., 2006). The uptake mech-
anismis saturable and can be competed away for both pDNAreview by (Wolff and
Budker, 2005)and mRNA (Probst et al., Gene Ther. 2007 Aug 14(15): 117580).
It involves the movement of vesicles and probably specic receptors. Other sites
than the skeletal muscle can be used for in vivo gene delivery using naked nucleic
acids: skin, liver, and heart muscle, for example; review by Nishikawa (Nishikawa
and Hashida, 2002). Following Wolff and associates results, the utilization of min-
imal nucleic acid vectors for local expression of an antigen to be recognized by
the immune system was undertaken as shown with pDNA rst (Ulmer et al., 1993)
and mRNA thereafter (Conry et al., 1995). The capacity of the immune system to
recognize specically the protein expressed from the foreign nucleic acid is prob-
ably linked to the capacity of immune cells such as dendritic cells to sense these
exogenous genetic molecules through TLR9 for bacterial DNA or TLR3 and TLR7
or TLR8 for double stranded RNA (dsRNA) and single stranded RNA (ssRNA),
respectively.
The utilization of mRNA has several superlative advantages compared to pDNA:
(i) At the peak of expression, the amount of protein produced through injection of
naked mRNA is higher than the amount of protein produced by the injection
of the same amount of naked pDNA (Probst, Gene Ther. 2007 Aug 14(15):
117580)
(ii) Due to its transient nature, mRNA is expressed during a controlled period of
few days while pDNA-expression is uncontrolled: Its expression in mice can
be transient or last for months depending on the random fate of these stable
molecules (the integration of pDNA in the genome could result in the long-
term expression of the transgene). This guarantees that long-term expression
Vaccination with Messenger RNA (mRNA) 223
and its possible consecutive tolerization of the immune response will not occur
with mRNA-based vaccines.
(iii) Because mRNAcannot modify the genetic information of the host, it is not con-
sidered a gene therapy approach by the authorities. Thus, for example, tedious
genotoxicity evaluation in animals can be avoided.
For these reasons, several methods based on mRNA for vaccination were tested
in mice and further evaluated in humans. This chapter summarizes the features of
the mRNA that are needed for vaccine formulation, the different methods that were
developed and validated in mice, and nally the result of phase I/II clinical trials.
2 Messenger RNA: Structure, Production, and Specic
Optimizations for Vaccination Purposes
2.1 Messenger RNA for Research
A mature eukaryotic mRNA has three characteristic structural elements: the 5

Cap,
the coding sequence starting with usually an ATG codon in a Kozak surrounding
and ending at a stop codon, and nally a poly-A tail of several hundred residues.
For vaccination purposes, mRNA can be puried from cells such as tumor cells
and formulated in an immunogenic solution in order to trigger an immune response
against a broad range of antigens. However, for mRNA-based vaccination, in vitro
transcribed mRNA is usually used. It is produced through molecular biology meth-
ods. First, the gene of interest is cloned in a plasmid vector that contains (i) an
upstream promoter exclusively recognized by processive RNA polymerases avail-
able as recombinant proteins such as the T7, S6, or T3 RNA polymerases from
bacteriophages; (ii) a downstream poly A sequence of a minimum of 30 bases; and
(iii) a unique restriction site downstream the poly A-tail. A bacterial clone contain-
ing this plasmid is cultured at 37

C in an adequate bouillon (usually LB medium

that contains the antibiotic for which the plasmid carries specic resistance). Bac-
teria are then collected by centrifugation and lysed (usually using sodium hydrox-
ide and SDS). Proteins and bacteria genomic DNA are precipitated by potassium
acetate. The cleared lysate contains the plasmids that can be isolated using, for ex-
ample, anion exchange columns (see, for example, www.qiagen.com). The eluted
pDNA is highly pure. It can be precipitated by salts (ammonium acetate or sodium
chloride, for example) plus isopropanol or ethanol. The nucleic acid pellet is resus-
pended in water or tris-EDTA buffer, quantied by spectrophotometry (absorbance
at 260 nm) and stored in the cold (4

C for short term and 20

C or 80

C for long
term). Messenger RNA is produced from pDNA in an in vitro transcription reac-
tion. To this end the pDNA is linearized thanks to the unique restriction site that is
downstream of the poly-A tail. After digestion, the proteins are extracted by phenol-
chloroform, the pDNA is recovered by ethanol precipitation, and, after a wash
in ethanol 75%, resuspended in water. The run-off transcription of this linearized
224 S. Pascolo
plasmid is performed by the addition of an adequate buffer, the RNA polymerase
specic for the upstream promoter (T7, SP6, or T3 RNA polymerase), the four nu-
cleotides in their triphosphate form (ATP, UTP, GTP, and CTP) and a four-fold
excess of Cap analogue (the dinucleotide methyl-7-Guanin(5

) PPP(5

) Guanin,
in short m7G(5

) ppp(5

)G) compared to GTP. This will guarantee that approxi-

mately 80% of the RNA molecules will start with a Cap instead of a G-residue
(the rst base of the RNA is dictated by the sequence of the promoter and set as
a G in T7, SP6, and T3 promoters). After two hours incubation at 37

C, a DNase
is added that will destroy the plasmid. Thereafter, long mRNA molecules are se-
lectively recovered by precipitation with lithium chloride. After a wash with 75%
ethanol, the mRNA pellet is resuspended in water and quantied by OD260. For
a detailed transcription protocol and overview of RNA recovery methods, refer to
the manual of commercially available transcription kits such as those from Ambion
(mMessagemMachine at www.ambion.com), for example. Messenger RNA resus-
pended in water can be stored at 4

C for days and at 20

C or 80

C for long-term
storage.
2.2 Messenger RNA for Clinical Applications
For good manufacturing practice (GMP) production, the antibiotic used in fermenta-
tion of the pDNA should preferably not be of the ampicillin family. This avoids po-
tential clinical problems due to penicillin allergies. The production process is similar
for research (as described above) and pharmaceutical grade products. However, for
mRNA in GMP quality, the nucleic acids can be puried using a chromatographic
method that allows elimination of traces of contaminants from the transcription
reaction (proteins, DNA fragments, endotoxins). Eventually the chromatography
method can also allow the recovery of the mRNA according to its size. The ad-
vantage of this method is that it eliminates abortive (shorter) or aberrant (eventually
longer) transcripts produced during the enzymatic reaction (see www.curevac.com
for more information). Should the chromatography method use other ions than
sodium, a precipitation of the mRNA with NaCl and ethanol will guarantee that
the counterion in the RNA batch is sodium.
The nal batch should appear as a transparent colorless solution and fulll the
following specications:
(i) Identity: Sequencing of the plasmid used for in vitro transcription should show
100% identity with the expected sequence. At best, reverse transcription and
sequencing of the nal mRNA could be performed. Agarose gel electrophoresis
is used to document the size of the mRNA and prove that only one species of
mRNA (one size) is present. Susceptibility to RNase can additionally be used
as a proof of molecular identity.
(ii) Content: Quantication should be performed by light absorbance at 260 nm.
Osmolarity and pH should also be measured. All these values should be in-
between prespecied limits.
Vaccination with Messenger RNA (mRNA) 225
(iii) Purity: Residual proteins, chromosomal bacteria DNA, bacteria RNA, and
endotoxin must be below specied limits. Sterility must be controlled by
standard microbiological assays. Residual pDNA and aberrant mRNA tran-
scripts (smaller or larger byproducts of the transcription) should be checked.
The former can be done by using quantitative PCR with primers specic for
the plasmid that was used for transcription. The latter can be done by agarose
gel electrophoresis. The amount of eventual contaminants should remain below
specied limits.
Moreover, counterions (which should be NaCl because of the nal precipitation
of the nucleic acids with alcohol plus NaCl), residual solvents (if they are used
during the production), and potency (functional assay using, for example, transfec-
tion of cells and verication of the expression of the protein of interest or testing
of the immune response after adequate application in an animal model) can ideally
be tested.
Pharmaceutical grade mRNA is offered by two companies: Asuragen in the USA
(www.asuragen.com) and CureVac in Europe (www.curevac.com).
2.3 Optimization of mRNA for Vaccination
It can be assumed, although it is not rmly proven, that a high and long-term expres-
sion of the protein encoded by the mRNA would favor the efcacy of the mRNA-
based vaccine. In order to enhance the transcription rate (high expression) and
stability (long-term expression) of the mRNA, several features of the molecule can
be optimized.
2.3.1 Optimizing the 5

Cap Structure
The Cap dinucleotide analogue usually used in the transcription reaction can be in-
corporated in two directions: either the 3

OH of the pentose carrying the methylated

guanine or the 3

OH of the pentose carrying the non-methylated guanine is used

for the phosphodiester bridge to the second residue of the mRNA. Only, the second
case will lead to a functional mRNA because the initiation elongation factor 4 (eIF4)
protein recognizes the methylated base at the 5

end position only. Thus, using the

standard Cap analogue, only half of the 80% mRNA molecules carrying the Cap
are functional. As a remedy, a modied Cap can be used. The most common one is
called ARCA (anti-reverse Cap) and consists of a Cap analogue with a modication
on the 3

OH of the ribose carrying the methylated guanine (Stepinski et al., 2001).

Consequently, this side cannot be used for the incorporation of the Cap. Thus, only
correctly capped molecules are generated. Moreover, the modication on the sugar
may in some unknown ways improve the mRNA since ARCA Cap mRNA mole-
cules are not two-fold but ve-fold more efciently translated than mRNA made
with the standard Cap.
226 S. Pascolo
Another method to get only correctly capped molecules is to do a transcription
reaction without Cap analogue and to perform on the synthesized mRNA a cap-
ping reaction. This reaction can be done using the vaccinia virus-encoded capping
complex. It consists of two subunitsD1 and D12and is also known as guanylyl-
transferase. This enzyme is commercially available as a recombinant molecule
(www.ambion.com). In the presence of GTP and S-adenosyl methionine (SAM),
the guanylyltransferase can add a natural Cap structure (7-methylguanosine) to the
5

triphosphate of a RNA molecule. However, because the efcacy of this reac-

tion cannot be controlled (no simple test allows one to check the level of capping
on mRNA molecules), the enzymatic capping is not used routinely. Excess of Cap
anologue or, at best, ARCA compared to GTP in the transcription reaction are the
standard, reliable methods to produce capped mRNA.
2.3.2 Optimizing the Untranslated Regions
The half-life of mRNA molecules in the cytosole is highly regulated (Ross, 1995). It
is in the range of minutes, for example, for messenger-coding cell cycle-controlling
proteins or cytokines; up to weeks, for example, for messenger-coding globins in red
blood cells. Destabilization and stabilization sequences are usually located at the un-
translated (UTR) 3

end of the mRNA (between the stop codon and the poly-A tail).
The most common destabilization mode is signaled by AU rich sequences called
AUREs in the 3

UTR. Thus, if they are present in the cDNA of interest, those AU-
REs sequences should be deleted from the pDNA construct that is used to produce
the mRNA. On the opposite side, the most common stabilization mode is signaled
by pyrimidine-rich sequences located in the 3

UTRs. They are recognized by the

ubiquitous -complex (Holcik and Liebhaber, 1997). The most characterized se-
quences recognized by the alpha complex are the approximately 180-base or ap-
proximately 80-base UTRs in the - or -globin genes, respectively. They allow the
long-term stability of those mRNA in terminally differentiated red blood cells de-
void of nuclei. - or -globin UTRs can be added to the end of any cDNA of interest
in the pDNA construct that is used to produce the mRNA. Messenger RNA used for
therapy usually possess such globin stabilization UTRs (Hoerr et al., 2000; Klencke
et al., 2002; Conry et al., 1995; Malone et al., 1989). Together with a poly-A tail of
a minimum of 30 residues but of preferably more than 100 residues (Mockey et al.,
2006; Holtkamp et al., 2006), the globin UTRs provide stability to the mRNA.
2.3.3 Optimizing the Translated Region
Optimizing the Translation Process
To optimize the production of the desired protein frommRNAmolecules, the correct
start codon should be the rst ATGtriplet met by the ribosome after it starts scanning
the mRNA from the 5

end towards the 3

end. This start codon should be in Kozak

Vaccination with Messenger RNA (mRNA) 227
surrounding which is: A/GNNATGG (Kozak, 1978). On this sequence the ribosome
will incorporate the N-terminal methionine and move toward the 3

end to the next

codon. No alternative cryptic start codon should be upstream of the correct start
codon since they would induce the production of unwanted upstream protein and
reduce the production of the expected antigen.
Most amino acids are encoded by several codons. Some of them are more fa-
vorable since they correspond to more abundant tRNA. This feature allows the ri-
bosome to perform translation quickly. Thus, mRNA highly translated, such as, for
example, mRNA coding for structure proteins (actin, etc.) contain mostly favorable
codons. Unfavorable codons slow down the ribosome and may also induce abortion
of the translation process. Because the relative abundance of tRNA is not conserved
between species, the codon preference or codon usage is also species-specic. For
example alanine, which is encoded by four codons, would be optimally coded by
GCG in Escherichia coli, by GCT in Saccharomyces cerevisiae, by GCA in Bacil-
lus subtilis, and by GCC in Homo sapiens. As a consequence, genes from bacteria
or fungi are not well translated by the mammalian cell machinery. As a standard
method for optimizing gene therapies, even when the therapeutic gene is of mam-
malian origin, a codon optimization should be performed to enhance the efcacy of
transgene expression in humans. To this end the protein sequence of interest is con-
verted into an optimal mRNA sequence using algorithms that take into account the
precise codon usage of human cells. This virtual gene is then created using assem-
bly of chemically synthesized DNA oligonucleotides. Many companies offer this
service such as Geneart (www.geneart.com) or Entelechon (www.entelechon.com)
or Blueheron (www.blueheronbio.com), for example. The synthetic gene can be
cloned in an mRNA production vector as described above. Not only will this allow
the generation of mRNA that is optimal for the translation machinery but it will also
facilitate the description and full documentation of the genetic starting material for
authorities in charge of clinical trials.
Optimizing the Antigen Presentation
Although epitopes presented by MHC class I molecule may mostly come from
quickly degraded natural defective translation products (DRiPs), it is possible
to specically enhance this antigen presentation pathway thanks to the addition
of ubiquitin, calreticulin (Anthony et al., 1999), or herpes simplex virus VP22
(Chhabra et al., 2004) moieties. These additional protein sequences can increase
the utilization of the attached antigen for generation of epitopes to be loaded on
MHC I molecules. Although these methods were not used for mRNA-based vac-
cines, they were proven useful to enhance the immunogenicity of antigens deliv-
ered by pDNA-vaccination. Thus, mRNA-based vaccinations could be expected to
be similarly optimized thanks to the introduction of sequence tags coding for such
MHC I optimizing moieties.
The presentation of MCH class II epitopes depends on the availability of the anti-
gen in endosomes of antigen presenting cells (APCs). MHC II antigen processing
228 S. Pascolo
and loading takes place in these vesicles. For antigens that are normally located in
the cytosole, the addition of sequences addressing to the endosomes is a validated
method to enhance mRNA-based vaccines at least in vitro using transfected APCs.
In this context, invariant chain (Momburg et al., 1993) or LAMP (Ruff et al., 1997)
moieties that naturally target the attached protein to the endosomes were shown
to optimize MHC II presentation of the mRNA-encoded antigen. Another generic
method to enhance the presentation of an antigen to the immune system is to address
it to the secretory pathway resulting from a signal sequence introduced in front of
the antigen (a leader sequence), and at the same time, target it to APCs by adding the
sequence of an APC-ligand such as CD40L, Flt-3L, or CTLA4 (Boyle et al., 1997;
Hung et al., 2001). These combined modications allow the secreted antigen to be
taken up by APCs and thereby enhance the efcacy of cross-presentation (natural
uptake by an APC of an antigen expressed by a somatic cell).
3 Methods of mRNA-Based Vaccination
3.1 Delivery
There are four different formulations of mRNA, which were described in mice
to trigger the development of an immune response against the coded antigen. In
chronological order of publication, those are:
1. Encapsulation of the mRNA into cationic liposomes (Martinon et al., 1993;
Hoerr et al., 2000; Hess et al., 2006). Mice injected with as little as one mi-
crogram of mRNA encapsulated in liposomes developed an immune response
against the encoded foreign antigen. Intravenous delivery was the most efcient
route of vaccination, although intradermal and eventually subcutaneous injec-
tions could also induce immunity. Instead of liposomes, the mRNA can be en-
trapped in cationic polymers such as protamine, with which it spontaneously
forms particles (Hoerr et al., 2000; Scheel et al., 2005). Intradermal injections
allow these complexes to trigger an immune response specic for the encoded
protein (Hoerr et al., 2000).
2. Direct injection of naked mRNA into the dermis (Conry et al., 1995 and Hoerr
et al., 2000). Surprisingly, in spite of the presence of ubiquitous RNases on the
skin (Probst et al., 2006), naked mRNA when injected in the dermis of mice is
spontaneously taken up and translated. The expressed protein is recognized as
antigen by the immune system. Formulation of the mRNA in a buffer that con-
tains calcium such as the Ringer solution promotes the spontaneous active uptake
of the injected mRNA (Probst et al., Gene Ther. 2007 Aug 14(15):117580).
3. Needleless delivery of gold particles coated by mRNA using a gene gun (Qiu
et al., 1996; Steitz et al., 2006). The mRNA is precipitated on micrometric gold
particles. They are dried inside a tube, which is cut in small pieces used as
cartouche in the helium-powered gene gun. By shooting, helium goes through
Vaccination with Messenger RNA (mRNA) 229
the cartridge and propels the gold beads, which, thanks to their kinetic energy,
penetrate through the stratum corneum and directly transfect with the loaded
mRNA all type of cells present in the dermis. This mechanical transfection al-
lows the expression of the mRNA-encoded protein. It is followed by the trigger-
ing of an immune response against foreign and also against self mRNA-encoded
protein (Steitz et al., 2006).
4. Transfection of in vitro generated autologous APCs that are re-administered to
patients (Boczkowski et al., 1996). APCs, generally monocyte-derived dendritic
cells (MoDC), are prepared from the patients peripheral blood. This requires
one week of in vitro culture. The cells are then transfected with the mRNA using
usually electroporation. After this transfection they are matured by culture one or
two days in a medium containing maturation signals and then re-injected into the
patients. Mature transfected APCs can efciently stimulate in vitro and in vivo
nave or memory CD4- and CD8-positive T cells specic for epitopes from the
foreign mRNA-encoded protein.
Although all four technologies were validated in mice, only methods (2), naked
mRNA, and (4), transfected MoDC were evaluated in humans through phase I/II tri-
als as presented below in paragraph 4. Probably because of formulation/stability and
toxicity issues, the liposome formulation of mRNA was not yet tested in humans.
Since the gene gun delivery method was shown to be very efcient and absolutely
safe in humans with pDNA vaccines, it was not tested using mRNA.
3.2 Adjuvants
3.2.1 Vaccination by Direct Application of mRNA
When the mRNA is delivered naked, encapsulated, or through a gene gun, mostly
somatic cells are transfected. Thus, as opposed to the mRNA-transfection in APCs
(see next paragraph), using these three methods, cross-presentation of the antigen
may be the primary mechanism for the priming of the immune response. To this
end professional APCs need to be present at the site of mRNA delivery. These
pick the antigen produced by somatic cells and, being mature, carry it to the drain-
ing secondary lymphoid tissues, where stimulation of specic lymphocytes takes
place. Thus, enrichment of the mRNA injection site with APCs could be a method
to improve the efcacy of vaccination. One well-characterized reagent available at
pharmaceutical grade and capable of attracting or inducing APCs is GM-CSF. In
mice, several peptide or pDNA-based vaccine formulations were found to be en-
hanced thanks to concomitant GM-CSF injections. In humans, peptide vaccines are
often delivered with recombinant GM-CSF. Accordingly, GM-CSF was tested as
an adjuvant for mRNA vaccination in mice using both the naked and liposome for-
mulations. Using naked mRNA injections (Carralot et al., 2004), the application of
recombinant GM-CSF either at the site of mRNA injection (dermis: ear pinna) or
at a distant site (subcutaneous: ank) was efcient in enhancing the Th1-type of
230 S. Pascolo
immunity against the mRNA-encoded antigen. However, the recombinant GM-CSF
had to be administered one day after the mRNA. Injecting recombinant GM-CSF
before or at the same time as the mRNA did not result in any improvement of the
immune response. Using liposome-encapsulated mRNA, the co-administration of
mRNA coding for a foreign antigen (Ovalbumin) and mRNA coding for GM-CSF
could enhance the immunogenicity of the vaccine as well as the development of
the primed T cells towards memory lymphocytes (Hess et al., 2006). In this set-up,
GM-CSF is actually present some time after the injection since uptake of the li-
posomes and translation of the mRNA require several hours. Thus, in both reports
(Carralot et al., 2004; Hess et al., 2006), the potential of GM-CSF to enhance the ef-
cacy of mRNA-vaccines is evidenced when this chemokine is present after mRNA
injection and not, as could be hypothesized from the known capacity of GM-CSF to
attract/differentiate DCs, before mRNA application.
3.2.2 Vaccination Using mRNA-Transfected APCs
The optimal status of vaccinating MoDCs is still a mater of debate: Although in vitro
matured DCs are more efcient than immature DCs for priming T cells, they are
eventually too exhausted for migrating to the lymph nodes in vivo. This would sug-
gest that DCs on the way to maturation rather than fully matured DCs should be
injected. Consequently, the optimal maturation signal to use in vitro should induce a
slow maturation and induce the capacity to migrate to lymph nodes (through CCR7
expression, for example). Once in the lymph node, the DCs should express adequate
cytokines such as IL-12 and co-stimulation signals such as CD80 or CD86. To this
end, a cocktail of four molecules is frequently used in order to induce maturation of
vaccine DCs: TNF-alpha, IL-6, IL-1 beta, and prostaglandin E2 (Thurner et al.,
1999). Two general improvements for the DC-based vaccines have been tested:
(i) the pretreatment of the DC-injection site with inammatory cytokines (Martin-
Fontecha et al., 2003), and (ii) the direct intralymphatic injection (Grover et al.,
2006). The rst method is based on induction of the CCR7 ligand CCL21 in lym-
phatic endothelial cells by inammatory cytokines. This creates a lymphatic avenue
from the skin to the lymph node. DCs subsequently injected at the same site use this
pre-made avenue and migrate efciently to the draining lymph node. This approach
was not yet reported in human clinical trials. The second method consists in the
mechanical delivery of the DCs at the site where they prime T cells, which is the
lymph node. Such a method can be performed in humans thanks to the monitoring
of the injection by ultrasound. Injection is performed while the tip of the needle
is in a targeted subcutaneous lymph node such as an inguinal lymph node. Intra-
lymphatic injection of peptide pulsed dendritic cells was reported to enhance the
efcacy of vaccination. Up to now however, as detailed below, mRNA-transfected
dendritic cells were not more potent in inducing an immune response when injected
via intralymph node compared to intradermal (Kyte and Gaudernack, 2006).
Vaccination with Messenger RNA (mRNA) 231
4 Results of Clinical Trials
Messenger RNA-based vaccines in humans were up to now only tested as
immunotherapies to induce or boost anti-tumor immunity in cancer patients.
4.1 Monocyte-Derived Dendritic Cells Transfected with mRNA
In Vitro
As mentioned above, the method described by the team of E. Gilboa in 1996 (Gilboa
et al., 2003; Boczkowski et al., 1996) was the rst to be used as an mRNA-based
vaccination regimen in humans. In one of the rst trials metastatic prostate cancer
patients received several injections of autologous MoDCs transfected with prostate-
specic antigen (PSA) coding mRNA (Heiser et al., 2002). After this treatment, spe-
cic T cells against PSA could be evidenced in the peripheral blood of all patients
(9/9). No severe adverse event was recorded. Thus, mRNA-tranfected MoDCs can
safely trigger the desired T cell immunity in human patients. This trial also indicates
that the newly induced T cell response may have had a clinical impact (decrease in
the log slope of PSA and transient clearance of circulating tumor cells) in some pa-
tients. In another trial made by the same team, MoDCs were transfected with the
total mRNA from autologous renal tumors (Gilboa et al., 2003; Su et al., 2003).
Most treated patients (6/7) responded to the vaccine as shown by monitoring antitu-
mor T cells. Superior results were obtained in a subsequent trial where on top of the
mRNA vaccine, patients received ONTAK, which is a IL-2-toxin fusion that kills
CD25-expressing T cells such as the CD4+/CD25+Tregs (Dannull et al., 2005). In
this case the ablation of Treg could allow a better proliferation and activity of antitu-
mor (antivaccine) T cells. Several other groups have also evaluated this technology.
In particular, the team of G. Gaudernack published the results of two trials, in which
either allogeneic total tumor mRNA (prepared from three prostate tumor cell lines)
or autologous total tumor mRNA (prepared from a surgically removed melanoma
metastasis) was transfected by electroporation in autologous MoDCs and injected
four times at weekly intervals. Those investigators could record the induction of
a T cell response against the tumor in approximately half the vaccinated patients
(Kyte et al., 2006) (Mu et al., 2005): 12/19 in the prostate cancer trial; 10/19 in
the melanoma trial. In the prostate carcinoma trial, 11 patients had stable disease
as judged by PSA levels in the blood. Ten of those eleven patients had a detectable
vaccine-induced T cell immune response while only two out of the eight patients
with progressive disease (increasing PSA levels) had a detectable vaccine-induced
T cell response. Thus, this allogeneic vaccination strategy may have had a posi-
tive impact on the tumor disease. In both of these trials, the intradermal injection
of transfected MoDCs was compared to the intralymph node injection. However,
surprisingly, the vaccine induced T cell response rate was higher in the group of
patients receiving intradermal injections (8/9 in the prostate trial) compared to the
group of patients receiving intralymph node injections (4/10 in the prostate trial).
232 S. Pascolo
Thus, using the MoDC preparation protocol proposed by Gaudernacks team, intra-
dermal injections are sufcient to trigger a potent and clinically relevant antitumor
immune response. Follow-up trials designed to further improve this strategy are on-
going. They include the concomitant systemic injection of IL-2 or the application
of TLR agonists on the skin at the injection site (Kyte and Gaudernack, 2006).
To conclude, mRNA transfected MoDCs is a safe and efcacious vaccine formu-
lation that can trigger antigen-specic T cells. The forthcoming results of phase II
and phase III trials will, it is hoped, rmly demonstrate that this method can induce
clinically relevant T cell immune response that can be benecial for patients health
and survival (for an update, see www.clinicaltrials.gov).
4.2 Direct Injection of mRNA
The direct intradermal injection of mRNA was only recently evaluated in humans
although it was described in mice as a vaccination method earlier than the mRNA-
transfected MoDC technology. In the rst trial, progressive stage III and stage IV
melanoma patients were recruited. Total mRNA corresponding to the whole tran-
scriptome of a surgically removed autologous metastasis was produced, formu-
lated, and injected in the dermis. One day after the mRNA injection, recombinant
GM-CSF was applied subcutaneously in the same area. Fifteen patients were re-
cruited and altogether 114 mRNA injections were performed. No WHO grade III or
grade IV adverse event was recorded. Three of thirteen evaluable patients showed
an increased antitumor antibody response during the treatment period (Weide et al.,
Journal of Immunu therapy, in press). The antivaccine T cell response was eventu-
ally induced in 5 of 13 evaluable patients. Although no objective clinical response
was observed in this trial, some patients showed an unexpected favorable course
of disease. This technology is being further evaluated in melanoma patients and in
renal cell carcinoma patients. However, in these subsequent trials dened mRNA
species are used (cocktails of 4 to 6 different mRNA coding for dened tumor
antigens). Thus, immunomonitoring will be facilitated compared to the autologous
whole mRNA approach used in the rst trial. These studies should indicate whether
the direct injection of mRNA is a method that can efciently prime an immune
response in human.
5 Conclusion
Four methods of mRNA-based vaccinations were validated in animal models and
two of them (mRNA transfection in vitro in MoDCs and direct injection in the der-
mis of naked mRNA) were evaluated in clinical trials. The very encouraging results
obtained by different academic groups using mRNA-transfected MoDCs may lead
to the validation in pivotal phase II or phase III trials of this method. As an alterna-
tive, the simpler direct injection of mRNA needs to be further evaluated in phase I/II
clinical trials. Meanwhile, our understanding of the immune systemand in particular
Vaccination with Messenger RNA (mRNA) 233
of the suppression mechanisms as well as of the adjuvant effects through TLR sig-
naling allow us to design more efcacious vaccination strategies that will certainly
help in turning mRNA-based vaccine strategies into potent therapeutic and prophy-
lactic modalities.
References
Anonymous (2003) Anti-vascular endothelial growth factor therapy for subfoveal choroidal neo-
vascularization secondary to age-related macular degeneration: Phase II study results. Ophthal-
mol 110: 979986
Anthony LS, Wu H, Sweet H, Turnnir C, Boux LJ, Mizzen LA (1999) Priming of CD8+ CTL
effector cells in mice by immunization with a stress protein-inuenza virus nucleoprotein fusion
molecule. Vaccine 17: 373383
Boczkowski D, Nair SK, Snyder D, Gilboa E (1996) Dendritic cells pulsed with RNA are potent
antigen-presenting cells in vitro and in vivo. J Exp. Med. 184: 465472
Boyle JS, Koniaras C, Lew AM (1997) Inuence of cellular location of expressed antigen on the
efcacy of DNA vaccination: Cytotoxic T lymphocyte and antibody responses are suboptimal
when antigen is cytoplasmic after intramuscular DNA immunization. Int Immunol 9: 1897
1906
Carralot JP, Probst J, Hoerr I, Scheel B, Teufel R, Jung G, Rammensee HG, Pascolo S (2004) Polar-
ization of immunity induced by direct injection of naked sequence-stabilized mRNA vaccines.
Cell Mol Life Sci 61: 24182424
Chhabra A, Mehrotra S, Chakraborty NG, Mukherji B, Dorsky DI (2004) Cross-presentation of a
human tumor antigen delivered to dendritic cells by HSV VP22-mediated protein translocation.
Eur J Immunol 34: 28242833
Conry RM, LoBuglio AF, Wright M, Sumerel L, Pike MJ, Johanning F, Benjamin R, Lu D, Curiel
DT (1995) Characterization of a messenger RNA polynucleotide vaccine vector. Cancer Res
55: 13971400
Dannull J, Su Z, Rizzieri D, Yang BK, Coleman D, Yancey D, Zhang A, Dahm P, Chao N, Gilboa
E, Vieweg J (2005) Enhancement of vaccine-mediated antitumor immunity in cancer patients
after depletion of regulatory T cells. J Clin Invest 115: 36233633
Grover A, Kim GJ, Lizee G, Tschoi M, Wang G, Wunderlich JR, Rosenberg SA, Hwang ST, Hwu
P (2006) Intralymphatic dendritic cell vaccination induces tumor antigen-specic, skin-homing
T lymphocytes. Clin Cancer Res 12: 58015808
Heiser A, Coleman D, Dannull J, Yancey D, Maurice MA, Lallas CD, Dahm P, Niedzwiecki D,
Gilboa E, Vieweg J (2002) Autologous dendritic cells transfected with prostate-specic antigen
RNA stimulate CTL responses against metastatic prostate tumors. J Clin Invest 109: 409417
Hess PR, Boczkowski D, Nair SK, Snyder D, Gilboa E (2006) Vaccination with mRNAs encoding
tumor-associated antigens and granulocyte-macrophage colony-stimulating factor efciently
primes CTL responses, but is insufcient to overcome tolerance to a model tumor/self antigen.
Cancer Immunol Immunother 55: 672683
Hoerr I, Obst R, Rammensee HG, Jung G (2000) In vivo application of RNA leads to induction of
specic cytotoxic T lymphocytes and antibodies. Eur J Immunol 30: 17
Holcik M, Liebhaber SA (1997) Four highly stable eukaryotic mRNAs assemble 3

untranslated
region RNA-protein complexes sharing cis and trans components. Proc Natl Acad Sci USA 94:
24102414
Holtkamp S, Kreiter S, Selmi A, Simon P, Koslowski M, Huber C, Tureci O, Sahin U (2006)
Modication of antigen-encoding RNA increases stability, translational efcacy, and T cell
stimulatory capacity of dendritic cells. Blood 108: 40094017
234 S. Pascolo
Hung CF, Hsu KF, Cheng WF, Chai CY, He L, Ling M, Wu TC (2001) Enhancement of DNA
vaccine potency by linkage of antigen gene to a gene encoding the extracellular domain of
Fms-like tyrosine kinase 3-ligand. Cancer Res 61: 10801088
Klencke B, Matijevic M, Urban RG, Lathey JL, Hedley ML, Berry M, Thatcher J, Weinberg V,
Wilson J, Darragh T, Jay N, Da Costa M, Palefsky JM (2002) Encapsulated plasmid DNA
treatment for human papillomavirus 16-associated anal dysplasia: A Phase I study of ZYC101.
Clin Cancer Res 8: 10281037
Kozak M (1978) How do eukaryotic ribosomes select initiation regions in messenger RNA? Cell
15: 11091123
Kyte JA, Gaudernack G (2006) Immuno-gene therapy of cancer with tumour-mRNA transfected
dendritic cells. Cancer Immunol Immunother 55: 14321442
Kyte JA, Mu L, Aamdal S, Kvalheim G, Dueland S, Hauser M, Gullestad HP, Ryder T, Lislerud K,
Hammerstad H, Gaudernack G (2006) Phase I/II trial of melanoma therapy with dendritic cells
transfected with autologous tumor-mRNA Cancer Gene Ther 13: 905918
Malone RW, Felgner PL, Verma IM (1989) Cationic liposome-mediated RNA transfection. Proc
Natl Acad Sci USA 86: 60776081
Martin-Fontecha A, Sebastiani S, Hopken UE, Uguccioni M, Lipp M, Lanzavecchia A, Sallusto F
(2003) Regulation of dendritic cell migration to the draining lymph node: Impact on T lympho-
cyte trafc and priming. J Exp Med 198: 615621
Martinon F, Krishnan S, Lenzen G, Magne R, Gomard E, Guillet JG, Levy JP, Meulien P (1993)
Induction of virus-specic cytotoxic T lymphocytes in vivo by liposome-entrapped mRNA. Eur
J Immunol 23: 17191722
Mockey M, Goncalves C, Dupuy FP, Lemoine FM, Pichon C, Midoux P (2006) mRNAtransfection
of dendritic cells: Synergistic effect of ARCA mRNA capping with Poly(A) chains in cis and
in trans for a high protein expression level. Biochem Biophys Res Commun 340: 10621068
Momburg F, Fuchs S, Drexler J, Busch R, Post M, Hammerling GJ, Adorini L (1993) Epitope-
specic enhancement of antigen presentation by invariant chain. J Exp Med. 178: 14531458
Mu LJ, Kyte JA, Kvalheim G, Aamdal S, Dueland S, Hauser M, Hammerstad H, Waehre H, Raabe
N, Gaudernack G (2005) Immunotherapy with allotumour mRNA-transfected dendritic cells in
androgen-resistant prostate cancer patients. Br J Cancer 93: 749756
Nishikawa M, Hashida M (2002) Nonviral approaches satisfying various requirements for effective
in vivo gene therapy. Biol Pharm Bull 25: 275283
Probst J, Brechtel S, Scheel B, Hoerr I, Jung G, Rammensee HG, Pascolo S (2006) Characterization
of the ribonuclease activity on the skin surface. Genet Vaccines Ther 4: 4
Probst J, Weide B, Scheel B, Pichler BJ, Hoerr I, Rammensee HG, Pascolo S (2007) Spontaneous
cellular uptake of exogenous messenger RNA in vivo is nucleic acid-specic, saturable and ion
dependent. Gene Ther. 2007 Aug; 14(15):117580. Epub 2007 May 3
Qiu P, Ziegelhoffer P, Sun J, Yang NS (1996) Gene gun delivery of mRNA in situ results in efcient
transgene expression and genetic immunization. Gene Ther 3: 262268
Ross J (1995) mRNA stability in mammalian cells. Microbiol Rev 59: 423450
Ruff AL, Guarnieri FG, Staveley-OCarroll K, Siliciano RF, August JT (1997) The enhanced im-
mune response to the HIV gp160/LAMP chimeric gene product targeted to the lysosome mem-
brane protein trafcking pathway. J Biol Chem 272: 86718678
Steitz J, Britten CM, Wolfel T, Tuting T (2006) Effective induction of anti-melanoma immunity
following genetic vaccination with synthetic mRNA coding for the fusion protein EGFPTRP2.
Cancer Immunol Immunother 55: 246253
Stepinski J, Waddell C, Stolarski R, Darzynkiewicz E, Rhoads RE (2001) Synthesis and properties
of mRNAs containing the novel anti-reverse Cap analogs 7-methyl(3

-O-methyl)GpppG and
7-methyl (3

-deoxy)GpppG RNA 7: 14861495

Su Z, Dannull J, Heiser A, Yancey D, Pruitt S, Madden J, Coleman D, Niedzwiecki D, Gilboa
E, Vieweg J (2003) Immunological and clinical responses in metastatic renal cancer patients
vaccinated with tumor RNA-transfected dendritic cells. Cancer Res 63: 21272133
Vaccination with Messenger RNA (mRNA) 235
Thurner B, Roder C, Dieckmann D, Heuer M, Kruse M, Glaser A, Keikavoussi P, Kampgen E,
Bender A, Schuler G (1999) Generation of large numbers of fully mature and stable dendritic
cells from leukapheresis products for clinical application. J Immunol Meth 223: 115
Ulmer JB, Donnelly JJ, Parker SE, Rhodes GH, Felgner PL, Dwarki VJ, Gromkowski SH, Deck
RR, DeWitt CM, Friedman Aand (1993) Heterologous protection against inuenza by injection
of DNA encoding a viral protein. Science 259: 17451749
Wolff JA, Budker V (2005) The mechanism of naked DNA uptake and expression. Adv Genet
54: 320
Wolff JA, Malone RW, Williams P, Chong W, Acsadi G, Jani A, Felgner PL (1990) Direct gene
transfer into mouse muscle in vivo. Science 247: 14651468