"Enhancement of Cancer Vaccine Therapy by Systemic Delivery of a Tumor Targeting Salmonella-based STAT3 shRNA Suppresses the Growth of Established Melanoma Tumors".

Edwin R Manuel ^{1, *}, Céline A Blache ^{1, *}, Rebecca Paquette ¹, Teodora I Kaltcheva ¹, Hidenobu Ishizaki ², Joshua D.I. Ellenhorn ³, Michael Hensel ⁴, Leonid Metelitsa ⁵, and Don J. Diamond ^1,

¹ Division of Translational Vaccine Research, Laboratory of vaccine research, City of Hope, Duarte, USA
² Department of General and Oncologic Surgery and Division of Translational Vaccine Research, City of Hope National Medical Center
³ General & Oncologic Surgery, City of Hope National Medical Center
⁴ Universitäts-klinikum Erlangen, Mikrobiologisches Institut, Erlangen; Germany
⁵ Pediatrics, Immunology, Baylor College of Medicine. Houston, Texas, USA

* Equal contribution by both authors.

Corresponding Author:
Don J. Diamond, Division of Translational Vaccine Research, Laboratory of vaccine research, City of Hope, Duarte, USA, 1500 East Duarte Road, Duarte, CA, 91010, USA
Phone: (626)-256-4673    Fax: (626)-301-8981    Email:  ddiamond@coh.org

Received December 30, 2010.  Revision received April 4, 2011.  Accepted April 21, 2011.

Cancer vaccine therapies have only achieved limited success when focusing on effector immunity with the goal of eliciting robust tumor-specific T cell responses. More recently, there is an emerging understanding that effective immunity can only be achieved by coordinate disruption of tumor-derived immune suppression. Towards that goal, we have developed a potent Salmonella-based vaccine expressing codon-optimized survivin (CO-SVN) referred to as 3342Max. When used alone as a therapeutic vaccine, 3342Max can attenuate growth of aggressive murine melanomas overexpressing SVN. However, under more immunosuppressive conditions, such as those associated with larger tumor volumes, we found that the vaccine was ineffective. Vaccine efficacy could be rescued if tumor-bearing mice were treated initially with Salmonella encoding a shRNA targeting the tolerogenic molecule STAT3 (YS1646-shSTAT3). In vaccinated mice, silencing STAT3 increased the proliferation and granzyme B levels of intratumoral CD4+ and CD8+ T cells. The combined strategy also increased apoptosis in tumors of treated mice, enhancing tumor-specific killing of tumor targets. Interestingly, mice treated with YS1646-shSTAT3 or 3342Max alone were similarly unsuccessful in rejecting established tumors, while the combined regimen was highly potent. Our findings establish that a combined strategy of silencing immunosuppressive molecules followed by vaccination can act synergistically to attenuate tumor growth, and they offer a novel translational direction to improve tumor immunotherapy.

http://cancerres.aacrjournals.org/content/early/2011/04/27/0008-5472.CAN-10-4676/suppl/DC1

Survivin (SVN) is a member of the inhibitor of apoptosis protein (IAP) family whose function is
involved in prolonging cell survival and cell cycle control (1, 2). SVN is an ideal tumor-associated
antigen (TAA) for therapeutic vaccination because it is overexpressed by essentially all solid tumors
and is poorly expressed in normal adult tissues (3). Increased expression of SVN is also observed in
endothelial cells during angiogenesis, thereby serving as an additional target for therapy (4). In
animal tumor models, downregulation or inactivation of SVN has already been shown to inhibit
tumor growth (5-7). Therefore, strategies to boost tumor-specific responses, such as using adjuvants
or immunogenic vectors, will be critical to the success of therapeutic vaccination (8, 9).

Even when favorable vaccination conditions are discovered that promote robust tumor-specific
immunity, these responses can eventually be compromised by expanding numbers of intratumoral
regulatory T cells and myeloid-derived suppressor cells (10-13). Signal transducer and activator of
transcription 3 (STAT3) has been recognized as an oncogenic transcription factor in myeloid and
tumor cells that, when activated, inhibits production of immunostimulatory molecules and promotes
expression of immunosuppressive molecules (14-16). A promising approach to inactivating STAT3
is the use of siRNA or shRNA, usually administered intratumorally, alone or conjugated to
molecules that target specific cell populations (17, 18).

Advances in the generation of attenuated enteric bacterial vectors, such as Salmonella typhimurium,
facilitates the highly translational tumor-specific delivery of antigens or plasmids (19-21). The
vector itself acts as an adjuvant to elicit innate immunity and aid in generation of adaptive immunity
against recombinant antigen. The most common Salmonella vaccines employ Salmonella
pathogenicity 1 (SPI1) type 3 secretion systems (T3SS), which only produce recombinant antigen in
a defined timeframe as the pathogen penetrates the host cell (22). More advanced vaccine designs
utilize SPI2 T3SS, which switches on recombinant antigen production when the Salmonella have
entered the host cell, allowing for extended antigen production (23). Numerous studies have
documented strains that colonize hypoxic regions of solid tumors weeks following intravenous
injection, with no detectable bacteria in peripheral organs, making it an ideal delivery vehicle for
targeting shRNA therapeutics into solid tumors (24-26).

In this report, we demonstrate a novel strategy utilizing two therapeutic agents delivered
systemically that are inadequate to control tumor growth as single agents but succeed as a combined
therapy. Specifically, attenuated Salmonella typhimurium carrying either a STAT3-specific shRNA
plasmid (YS1646-shSTAT3) or an SVN expression plasmid (3342Max) were administered
consecutively and observed to function synergistically leading to effective tumor rejection. The
combined approach improves the prospects for successful vaccination against cancer by altering the
tumor microenvironment to be less antagonistic to tumor infiltrating T cells such as those stimulated
by vaccine-encoded TAAs.

Materials and Methods:

Animals, tumor lines, and bacterial strains.

C57BL/6 mice (Jackson, 6-8 weeks) were obtained
from breeding colonies housed at the City of Hope (COH) Animal Research Center (Duarte, CA).
The B16F10 murine melanoma line was a kind gift from Drs. Hua Yu and Marcin Kortylewski
(COH, Duarte, CA). Cells were maintained in DMEM containing 10% FBS. S. typhimurium strains
MVP728 (purD-/htrA-) and YS1646 (ATCC#202165) were cultured by shaking at 37°C in LB or
LB-O media.

Salmonella SPI2 expression vectors, shRNA plasmids, and generation of recombinant

Salmonella. pWSK29 constructs containing the SPI2 expression cassettes for LisA (2810) or SVN
(3342) are described elsewhere (23). For construction of pWSK29 encoding Salmonella codon
optimized survivin (CO-SVN), 2810 was digested with XbaI/EcoRV and the gel purified pWSK29
backbone was used to clone the CO-SVN gene (Genscript, Piscataway, NJ) engineered with
XbaI/EcoRV sites for in frame fusion with the sseF gene. shRNA constructs against STAT3
(Origene, Rockville, MD) were tested for silencing by stable transfection of B16F10 cells followed
by western blot (WB) analysis using polyclonal rabbit antibody against STAT3 (Santa Cruz
Biotech, Santa Cruz, CA). The pGFP-V-RS vector containing the 29-mer shRNA sequence
ACCTGAAGACCAAGTTCATCTGTGTGACA (ID#GI556360) exhibited >70% STAT3
knockdown and was selected for generation of recombinant YS1646. SPI2 expression vectors and
shRNA plasmids were electroporated into MVP728 or YS1646, respectively, with a BTX600
electroporator (BTX, San Diego, CA).

Western blot analysis.

WB for Salmonella expression of SVN was carried out as described
previously (23). Briefly, 3342 and 3342Max were grown overnight in MOPS based media (Sigma)
at 37ºC containing either low phosphate (113 uM) to induce SPI2 expression or high phosphate (25
mM). Bacterial pellets were boiled in SDS loading buffer and equal amounts of lysate were loaded.
Blots were probed using a monoclonal rabbit antibody (ab76424) against SVN (Abcam, Cambridge,
MA).

Tumor challenge, vaccination, and shRNA therapy.

For tumor challenge, 10⁵ B16F10 cells were
injected subcutaneously into C57BL/6 mice. Tumor growth was monitored daily or every other day
using a caliper. For testing vaccination alone, MVP728 carrying 2810, 3342 or 3342Max were
administrated by gavage twice, 4 days apart, when tumors reached 3.5-4 mm in diameter at 10⁸ cfu.
For combined therapy, PBS, YS1646-STAT3 or -scrambled was first injected at 10⁷ cfu in C57BL/6
mice when tumor volumes were   ~50 mm³ (7-8 mm in diameter) followed by gavage with PBS or
10⁷ cfu MVP728-2810 or -3342Max.

Quantitative PCR for detection of STAT3 levels:

Mice bearing B16F10 tumors (  ~50mm³) were
i.v. injected with 10⁷ cfu of YS1646-scrambled, -shSTAT3, or PBS twice, 4 days apart. At days 3,
7, and 10, mice (n=3) were sacrificed and RNA was extracted from tumor homogenates for
generation of single stranded cDNA (Fermentas, Glen Burnie, MD). To quantitate STAT3 levels,
SYBR®-Green qPCR analysis (BD Biosciences, Franklin Lakes, NJ) using primers specific for
STAT3 (Forward: 5’-CATGGGCTATAAGATCATGGATGCGAC-3’, Reverse:
5’-AGGGCTCAGCACCTTCACCGTTATTTC-3’) was carried out using GAPDH (Forward:
5’-CAAGGTCATCCATGACAACTTTG-3’, Reverse: 5’-GTCCACCACCCTGTTGCTGTAG-3’) for
normalization.

Immunofluorescence staining.

For detection of intracellular SVN expression from recombinant
Salmonella, RAW264.7 macrophages seeded on coverslips were infected for 30 minutes at an MOI
of 10 with wildtype MVP728, 3342, or 3342Max. Cells were incubated overnight in DMEM-10
containing 10 µg/mL gentamicin. Cells were fixed/permeabilized with 1:1 acetone:methanol and
stained with conjugated antibodies FITC-LPS (Santa Cruz Biotech, Santa Cruz, CA) and PE-HA
(Covance, Princeton, NJ) overnight at 4ºC followed by DAPI. Cells were imaged on an Axiovert
200 using live imaging software (Axiovision, Skokie, IL). Images shown are representative of cells
observed within multiple fields.

Flow cytometry.

Conjugated mAbs directed to PECy7-CD8, PerCP-CD45, and PE-phospho-STAT3 were
purchased from BD Pharmingen (San Diego, CA) and mAb to APC-Cy7-CD4, APCF4/
80, FITC-Ki-67, PE-Granzyme B, and FITC-Annexin V were purchased from eBioscience (San
Diego, CA). Intracellular phosphor-STAT3, Granzyme B, Ki-67 and Annexin V staining were
performed following the manufacturer's protocol (eBioscience). Samples were run on a FACSCanto
(Becton Dickinson, La Jolla, CA) and analyzed using FlowJo™ software (TreeStar, Ashland,
OR).

Cytotoxicity assay.

Cytotoxicity against B16F10 melanoma cells in treated mice was determined using
a standard ⁵¹Cr release assay (27). Briefly, effectors were derived from spleens of B16F10-
bearing C57BL6 mice (n=4) i.v. injected with either PBS, 10⁷ cfu of YS1646-shSTAT3 or -
scrambled followed by gavage with PBS or 10⁷ cfu of 3342Max or 2810 4 days later. Mice were
sacrificed ~1 week post-gavage and splenocytes were co-incubated with RMA-S cells loaded with
human SVN library (27). Effectors were then co-incubated for 4 hours with 5,000 ⁵¹Cr-loaded
B16F10 targets in 96-well plates at ratios of 100:1, 20:1, and 4:1 (in triplicate). Radioactivity
released into the supernatant was measured using a Cobra Quantum gamma counter (PerkinElmer).
Percent specific lysis: (experimental release - spontaneous release)/(maximum release -
spontaneous release) x 100.

Statistical analysis.

Statistical significance for comparisons between two or more groups was
calculated with the Graphpad Prism Software v4.03 using the Student’s t test or one-way ANOVA,
respectively. A p value <0.05 was considered significant. All experiments were typically performed
at least in duplicate, and all data are presented as mean ± SEM. *p<0.05, **p<0.01, and
***p<0.001.

Results:

Construction and evaluation of SVN expression vectors

Previous work using the MVP728 bacterial vector transformed with plasmid 3342, which expresses
SVN, demonstrated partial success in rejecting murine models of colon carcinoma and glioblastoma
(23). We found that SVN expression from 3342 was suboptimal when compared to LisA expression
from 2810 (data not shown). We hypothesized that codon optimization (CO) of SVN to Salmonella
preferred codons would allow for increased stability and protein expression leading to greater antitumor
effects (28, 29). To test this hypothesis, a Salmonella typhimurium CO-SVN sequence was
generated using an online algorithm (30) and then synthesized (Genscript). As shown in Fig. 1A,
the low copy plasmid pWSK29 was engineered to encode the SPI2 chaperone protein sscB (31) and
sseF protein fused to either LisA (2810), SVN (3342), or CO-SVN (3342Max) for expression and
secretion by MVP728. Ultimately, expression of these genes would be regulated by the SPI2
specific promoter for sseA.

We determined whether CO actually increased SVN expression by growing the recombinant
Salmonella in SPI2-inducing conditions (32) (Fig. 1B). Under non-inducing conditions (PCN+P),
we found no significant expression of SVN. Surprisingly, under inducing conditions (PCN-P), we
observed much greater SVN expression from 3342Max compared to the non-optimized 3342. To
further evaluate SVN expression and secretion by 3342Max, we infected RAW264.7 murine
macrophages with either 3342 or 3342Max to determine intracellular expression of SVN by
immunofluorescence. As shown in Fig. 1C (HA-SVN panel), we observed greater SVN expression
(characterized by more foci) compared to 3342. As expected, mAb staining for both the LPS (LPSSt
panel) and HA (HA-survivin panel) localized to the cytoplasm and overlapped in the Merge
panel, independent of the nuclear DAPI staining. These data suggest that optimization of SVN
tailored to preferred Salmonella codons greatly improves recombinant antigen expression.

CO-SVN enhances suppression of tumor growth

We next evaluated whether enhanced expression of SVN by 3342Max translated into a more
efficacious vaccine using a B16F10 melanoma, which naturally overexpresses SVN (inset of Fig.
2A). Subcutaneously injected tumor cells were allowed to grow until a palpable tumor was present,
generally 3.5-4 mm in diameter (<10mm³). Mice were then gavaged twice (4 days apart) with either
PBS or MVP728 harboring 2810, 3342, or 3342Max constructs. As shown in Fig. 2A, 3342Max
vaccination was superior to all other experimental treatments in attenuating tumor growth (p<0.01).
We then determined the lymphocyte subsets that were most responsible for the attenuation by
carrying out in vivo antibody depletions of CD8+, CD4+, or NK populations in tumor bearing mice
vaccinated with 3342Max as in Fig. 2A. We observed that depletion of CD8+ T lymphocytes
resulted in significant loss of tumor growth control with an intermediate effect of NK depletion,
which has been described previously (17). These data suggest that vaccination with 3342Max elicits
superior CD8+ T cell responses that limit tumor growth, likely a result of enhanced SVN expression.

Attenuation of STAT3 mRNA levels using shRNA

When subcutaneous B16F10 tumors were grown to larger volumes before treatment (7-8 mm in
diameter, ~50mm³), we discovered that 3342Max vaccination had no efficacy to attenuate growth
(data not shown). Since we demonstrated that 3342Max works efficiently in less demanding
circumstances (Fig. 2), we presumed that failure under more stringent conditions was likely the
result of greater levels of tumor-derived immunosuppression (33). To determine if we could rescue
the efficacy of the vaccine, we sought to manipulate the tumor microenvironment by silencing the
tolerogenic molecule STAT3 (15, 34, 35). We chose to inactivate STAT3 mRNA expression using
an shRNA expression plasmid carried by the tumor-targeting Salmonella strain YS1646 (36). We
first tested several commercially available shRNA plasmids (Origene) to silence the expression of
STAT3 in stably transfected B16F10 tumor lines. As shown in Fig. 3A, shSTAT3#60 showed
dramatic silencing (>70%) of endogenous STAT3 when compared to scrambled shRNA control
plasmid. Other shSTAT3 plasmids had intermediate to no effect on endogenous STAT3 expression.
Targeted silencing of STAT3 combined with 3342Max results in significant suppression of
tumor growth in a more advanced melanoma tumor model

We next generated YS1646 carrying the shSTAT3#60 plasmid (YS1646-shSTAT3) to test whether
systemic delivery of Salmonella by i.v. route could silence STAT3 expression in situ in the tumor.
Mice bearing subcutaneous B16F10 tumors (~50 mm³) were injected twice i.v. with 10⁷ cfu of
YS1646-shSTAT3, -scrambled, or PBS 4 days apart. Post-treatment, no significant attenuation of
tumor growth was observed for mice treated with YS1646-shSTAT3 alone compared to control
groups (data not shown). These same results were also observed in less stringent conditions where
initial tumor volumes were <10mm³ (data not shown). Nonetheless, tumors were isolated,
homogenized, and total RNA extracted for quantitative PCR. Surprisingly, there was significant
silencing of STAT3 three days after YS1646-shSTAT3 administration compared to mice
administered YS1646-scrambled or PBS (Fig. 3B). STAT3 silencing continued to increase on day 7
in the shSTAT3 group, which is consistent with the continued effectiveness of the therapeutic
strategy (see below). On day 10, STAT3 silencing moderated, but was still lower than the control
groups. Confirmation that YS1646-shSTAT3 succeeded in specifically silencing STAT3 mRNA
but failed to reject tumors as a single agent motivated us to combine delivery of shSTAT3 and
3342Max vaccination in mice with significantly larger B16F10 tumors. Therefore, mice bearing
B16F10 tumors   ~50 mm³ were i.v. injected with 10⁷ cfu of YS1646-shSTAT3, -scrambled, or PBS.

Four days later, mice were gavaged with 10⁷ cfu of 3342Max, 2810, or PBS. As shown in Fig. 3C,
the combination of shSTAT3+3342Max rescues the activity of the vaccination to attenuate tumor
growth significantly better than control groups. These results suggest that combining shSTAT3
therapy and SVN vaccination is a powerful synergistic approach to attenuate tumor growth.
Decreased phospho-STAT3 levels are observed in tumor macrophages following shSTAT3
and 3342Max treatment

We hypothesized that the success of the combined treatment was in part due to suppression of
phospho-STAT3 levels in specific immune populations. Therefore, we used flow cytometry to
determine the levels of activated STAT3 in specific immune subsets present in the tumor following
treatment. We found no significant changes in phospho-STAT3 levels for CD4+, CD8+, CD11c+, or
Gr1+CD11b+ cells in all treatment groups (Supplemental Fig. 1). However, we did observe
significantly decreased phospho-STAT3 levels (p<0.05) in F4/80+ macrophages for the
shSTAT3+3342Max treated group (Fig. 4A). Surprisingly, no significant decreases of phospo-
STAT3 were observed for the shSTAT3+2810 group. These results suggest that only the
shSTAT3+3342Max treatment is able to prevent activation of STAT3 in the F4/80+ subset, likely a
result of early STAT3 silencing followed by tumor growth control, whereas shSTAT3+2810 is
unable to do so regardless of early STAT3 silencing due to uncontrolled tumor growth.

Combined shSTAT3 and 3342Max administration enhances infiltration of T lymphocytes

Since it was known that ablation of STAT3 increases intratumoral immune function (14, 17, 35), we
first examined the frequency and functional status of intratumoral CD4+ and CD8+ T cells in
vaccinated mice. The percentage of B16F10 intratumoral CD4+ and CD8+ T cells was statistically
greater in mice treated with shSTAT3+3342Max than in the scrambled+3342Max or
shSTAT3+2810 treatment groups (Fig. 4B). We next evaluated the proliferative index of these
intratumoral CD4+ and CD8+ T cells by determining Ki67+ expression levels. Both CD4+ and CD8+
populations expressed higher levels of Ki67+ in the shSTAT3+3342Max group compared to control
groups (Fig. 4C and D). The markedly higher proliferation potential suggests that the combined
shRNA and vaccination treatments allow for a proliferative expansion of intratumoral T cells, and
the increased frequency may therefore not solely be explained by a redistribution of existing T cells
from other sites.

YS1646-shSTAT3 enhances tumor-specific cytotoxic responses and tumor cell apoptosis

We addressed tumor cell death by evaluating the extent of apoptosis using Annexin V staining of
gated CD45- cells, mainly tumor cells (37), from all of the treatment groups. The CD45- cells
revealed significantly higher apoptotic frequencies in mice treated with shSTAT3+3342Max than
the control groups (Fig. 5A). The increased apoptosis of tumor cells could either be explained by
the cytotoxic activity of immune cells or possibly by a shSTAT3-based mechanism to enhance
apoptotic signal transduction. To address immune-based mechanisms, we investigated function of
the CD8+ T cell subset by evaluating granzyme B levels in B16F10 tumor-bearing mice ( ~50mm³)
treated with shSTAT3+3342Max versus groups treated with scrambled+3342Max or
shSTAT3+2810 (Fig. 5B). The proportion of CD8+ T cells expressing granzyme B in the mouse
group treated with shSTAT3+3342Max was dramatically higher than both control groups. These
results suggested a potential cytotoxic mechanism for tumor growth control, which we further
assessed using a direct in vitro cytotoxicity assay.

Tumor-specific cytotoxicity contributes to control of established subcutaneous B16F10 tumors

We determined if T cells obtained from B16F10 tumor-bearing mice treated with
shSTAT3+3342Max possessed functional capacity to kill survivin-expressing tumor cells in vitro
by conducting a chromium release assay (Fig. 5C). Splenocytes harvested from B16F10 tumor-
bearing mice (n=3) treated as in Fig. 5A were in vitro stimulated with a human SVN peptide library,
then evaluated for in vitro cytotoxic recognition and killing of chromium-loaded B16F10 tumor
targets. Mice treated with either scrambled+3342Max or shSTAT3+2810 could not effectively kill
B16F10 tumor cells. In contrast, splenocytes from all mice receiving shSTAT3+3342Max treatment
were very effective at killing B16F10 tumor targets (0.001<p<0.01) at all effector ratios (Fig. 5D).
These results further suggest that a potential mechanism of tumor growth attenuation is by tumorspecific
T cells, stimulated through SVN vaccination or SVN peptide stimulation, directly killing
tumor cells, though only when mice are pre-treated with shSTAT3 and then vaccinated with
3342Max.

Discussion:

The goal of these studies was to discover a translational approach that would provide durable
control of solid tumor growth. Our initial hypothesis was that SVN as a ubiquitously expressed
TAA would provide the widest versatility for vaccination. In contrast to Salmonella-based SVN
vaccines used in previous studies, which have been relatively ineffective when used alone, the goal
was to find a vaccine strategy that would not require additional cytokine or chemokine components
for effectiveness (8, 9, 23). A simplified one component regimen based on Salmonella delivery of
SVN was the initial goal of this study. Our use of oral systemic administration of Salmonella
transformed with SVN expression plasmids was similar to other reports describing Salmonella
routes of administration. The advantage to this approach is that Salmonella are efficiently
recognized by antigen processing macrophages in the gut or other mucosal sites (38, 39).

Initial trials using 3342 were only partially successful against small tumors that had just a few days
to develop vascularization (Fig. 2). By investigating the expression levels of SVN from 3342 and
bacterial LisA from 2810, we discovered significantly lower levels of SVN by WB analysis
compared to the bacterial LisA protein under identical conditions (data not shown). Since the
bacterial LisA protein was so heavily expressed in Salmonella, we theorized that changing the
sequence of SVN from human to Salmonella typimurium preferred codons might achieve the same
goal. In preliminary experiments, we found a gradation of effectiveness against growth of
established subcutaneous tumors dependent on expression levels (data not shown). The
predictability of the increasing effectiveness to reject established tumors made it unnecessary to
continue to simultaneously evaluate all forms, and the most effective form (3342Max) was
exclusively used in all further comparisons. Ultimately, we found that control of B16F10 tumor
growth using 3342Max vaccination only worked shortly after tumor challenge, when tumors
became palpable. Upon treatment of mice with larger B16F10 tumors of volumes ~50 mm³, the
vaccine was unable to attenuate tumor growth. This was not surprising as immunosuppressive
mechanisms likely became more established within the growing tumor, and the single modality
vaccination had no means to overcome them (11, 40).

Despite advances in therapeutic vaccines, there now exists numerous studies that support the idea of
tumor-derived immunosuppression contributing to tumor evasion (41). These mechanisms include
the secretion of TGF-b or IL-10 leading to Th2 polarization (42-44) or production of IDO by
myeloid cells to induce the generation of Tregs and T cell anergy (45, 46). The novel mechanisms
by which STAT3 causes immunosuppression are just beginning to be unraveled (14, 15, 47), and
with its other roles in tumor progression, has become a multi-faceted target that could potentially
attenuate tumor growth on its own or enhance the anti-tumor effects of any immunotherapy.
Disrupting tumor-induced immunosuppression has generally been studied in the context of its
effects that are independent of antigen-specific vaccination. Only rarely has the combination of
disrupting immunosuppressive mechanisms within tumors been combined with vaccination to limit
tumor growth. A more recent study has examined tumor-associated stromal cells expressing
fibroblast activation protein-a (FAP) as a source of immunosuppression in a model of pancreatic
ductal adenocarcinoma (48). Administration of a therapeutic vaccine in the absence of FAP-expressing
stromal cells showed significant increases in hypoxia-induced tumor necrosis when
compared to FAP+ mice. Similarly, we have also shown modest additive effects when vaccination is
combined with the drug gemcitabine (27), while others reported increased anti-tumor responses by
inhibiting the tolerogenic molecule IDO (49). These studies emphasize that successful outcomes of
immunotherapy will likely require overcoming tumor-induced immunosuppression.

In several genetic models of conditional STAT3 deletion, subsequent immune-enhancement enabled
dramatic inhibitory effects on tumor growth. Mechanism-based studies revealed changes in
cytokine profile, T cell subsets, and signal transduction modifiers that all contributed to the blunting
of tumor growth as a result of a reduction or elimination of STAT3 expression (14, 34, 35). These
elegant studies have lead to preliminary therapeutic strategies employing small molecule inhibitors
and RNA interference by a variety of approaches that have in common direct intratumoral
administration. These approaches have shown moderate efficacy, but in every case there is tumor
breakthrough within 20-25 days post-administration. An alternative strategy has been the approach
of tumor-targeting Salmonella delivery of shRNA eukaryotic expression plasmids by i.v. injection.
In contrast to our findings, others have found efficacy through intratumoral administration of
shSTAT3 alone (18). Nonetheless, the growth attenuation was transient and its translational
potential, as an intratumoral therapy, remains in doubt. Although systemic administration of
YS1646-shSTAT3 may require more diligent efforts to determine its specific cell targets, this
obstacle does not detract from the translational feasibility of the approach for treatment of solid
tumors. Our Salmonella approach and work published by others was similar in that the STAT3-
specific shRNA sequence only had a single target. In contrast, a CpG DNA chimera with an RNAi
sequence that was administered intratumorally had multiple off-target sequences >100 in the mouse
genome that temper the interpretation of the results (17).

We show for the first time that an intravenously administered shRNA against STAT3 acts
synergistically with an oral Salmonella-based vaccine against SVN in a therapeutic setting, resulting
in suppression of subcutaneous B16F10 melanoma growth. We conclude that the in vivo
suppression of B16F10 tumor growth is the result of increased tumor cell apoptosis, as determined
by Annexin V staining, possibly caused by an increased level of tumor-specific CD8+ T cells within
the tumor. We saw no changes in tumor-expressed SVN during the treatments (Supplemental Figure
2) that might explain the increase in apoptosis or eventual escape from control (50) . The higher
Ki67+ levels also indicated that these intratumoral T cells were actively proliferating, thereby
supporting the notion that shRNA against STAT3 attenuated immuno-suppression within the tumor
microenvironment. Moreover, the fact that neither the vaccine nor shRNA against STAT3 alone
was effective to control tumor growth suggests that the combined treatments acted synergistically.
These data support that implementing successful immunotherapy may be futile without a receptive
tumor microenvironment generated through additional modalities such as shRNA to inhibit
immunosuppression.

Figures:

Fig. 1. Construction and validation of SVN expression vectors.

Fig. 1. Construction and validation of SVN expression vectors.

A, the expression vectors 2810, 3342, and 3342Max were constructed to encode HA-tagged LisA, SVN, or SVN codon-optimized for Salmonella (CO-SVN), respectively, using the low copy plasmid backbone pWSK29. Each of these proteins was fused to the SPI-2 protein sseF and its expression is dependent on the SPI-2 promoter sseA. Each construct was then electroporated into MVP728 (23).

B, SVN expression from MVP728 harboring 3342 and 3342Max (Max) constructs was detected by western blotting (WB) of bacterial lysates cultured in inducing conditions (low phosphate media, PCN–P) or non-inducing conditions (high phosphate media, PCN+P). Fusion protein was detected using anti-SVN antibody.

C, RAW264.7 cells infected with MVP728 alone, MVP728-3342, or -3342Max were fixed and
permeabilized and then stained with the conjugated antibodies LPS-FITC, HA-PE, and the nuclear
stain DAPI. Cells were imaged under 100X oil immersion using an Axiovert 200. Scale bars, 5 µm.

Fig. 2. Codon optimization of SVN enhances suppression of B16F10 melanoma growth.

Inset of A, B16F10 melanoma cell lysates were analyzed by WB for SVN expression.

A, C57BL6 mice (n=5) were injected subcutaneously (s.c.) with B16F10 on day 0 and then vaccinated with MVP728-2810, -3342, -3342Max, or PBS on days 3 and 7. Tumor volume was monitored daily.

B, Following s.c. injection of tumor on day 0, mice bearing palpable B16F10 tumor were vaccinated twice with MVP728-3342Max (days 3 and 7) and then depleted of immune subpopulations (day 5) using 200
ug of anti-CD8 mAb (clone H35), anti-CD4 mAb (clone GK1.5), or anti-NK polyclonal Ab (antiasialo
GM1) with a maintenance dose every 3 days thereafter (27).

Fig. 3. Targeted silencing of STAT3 using YS1646-shSTAT3 results in significant suppression
of tumor growth when combined with 3342Max.

A, WB of STAT3 protein expression from B16F10 lysates after stable transfection of shRNA contructs (#58-61) with potential for silencing STAT3. b-tubulin is used as a loading control.

B, silencing of STAT3 expression in B16F10 tumor following i.v. injection of YS1646-shSTAT3. Mice bearing palpable B16F10 tumors were i.v. injected with 10⁷ cfu of YS1646-shSTAT3 twice, 4 days apart. Mice (n=3) were sacrificed on d3, d7, or d10 after first injection and tumor lysates were subjected to RNA extraction for qPCR analysis of STAT3 transcripts.

C, YS1646-shSTAT3 rescues anti-tumor effects of MVP728-3342Max in B16F10 model. C57BL/6 mice bearing B16F10 tumors (  ~50 mm³) were treated with either PBS, YS1646-shSTAT3 or -scrambled by intravenous injection. Four days following treatment, mice were then vaccinated with either PBS, MVP728-2810, or -3342Max and then monitored for tumor growth.

Fig. 4. YS1646-shSTAT3 treatment followed by 3342Max vaccination attenuates STAT3 activation in resident tumor macrophages and enhances infiltration of T lymphocytes.

B16F10 tumor-bearing mice (  ~50 mm³, n=5) were injected i.v. with 10⁷ cfu of YS1646-scrambled, -
shSTAT3, or PBS. Four days later, mice were then gavaged with 10⁷ cfu of MVP728-3342Max, -
2810, or PBS. B16F10 tumors were excised from mice seven days after vaccination and then
homogenized for staining and flow cytometry.

A, Comparison of phospho-STAT3 levels in F4/80+ macrophage for each treatment group. Phospo-STAT3 expression is presented as mean fluorescence intensity (MFI) and error bars represent standard error of the mean (SEM).

B, Frequency of CD4+ and CD8+ cells found in the tumor for each treatment group. Data represent absolute number of cells/mm3 tumor. CD4+ (C) and CD8+ (D) T cells were also analyzed for the expression of the
proliferation marker Ki-67.

Fig. 5. YS1646-shSTAT3 enhances tumor-specific cytotoxic responses.

B16F10 tumor-bearing mice (n=5) received combined treatment as described in Fig. 4.

A, Individual histograms of FITCAnnexin V stained tumor homogenates for a representative mouse from each treatment group.

B, tumor homogenates (n=5) from each group were stained with FITC-Annexin V and analyzed by
flow cytometry. Mean fluorescence intensity (MFI) of Annexin V represents cells gated from total
tumor CD45- cells. Error bars represent SEM.

C, tumor homogenates (used in A) were stained with PE-Granzyme B and PeCy7-CD8 and then analyzed by flow cytometry. Data represent mean percentages of Granzyme B+CD8+ cells out of total CD8+ cells.

D, splenocytes from mice in A (n=4) were isolated to generate effectors for use in a chromium release assay against B16F10 targets. To generate effectors, splenocytes were incubated for 7 days with RMA-S cells initially loaded with total human SVN library (15mers, overlapping by 11). Effectors were then incubated in
a 4-hour ⁵¹Cr release assay with ⁵¹Cr-loaded B16F10 targets at E:T ratios of 100:1, 20:1, and 4:1, in
triplicate. Percent specific lysis: (experimental release-spontaneous release)/(maximal releasespontaneous
release) x 100%.

Supplemental Figure 1. Measuring phospho-STAT3 levels in specific immune subsets for various treatment groups.

B16F10 tumor-bearing mice (~50 mm³, n=5) were injected i.v. with 10⁷ cfu of YS1646-scrambled, -shSTAT3, or PBS. Four days later, mice were then gavaged with 10⁷ cfu of MVP728 3342Max 2810 or PBS B16F10 tumors were excised from MVP728-3342Max, -2810, PBS. mice seven days after vaccination and then homogenized for staining the indicated immune subsets plus phospho-STAT3 for flow cytometry.

Supplemental Figure 2. Survivin expression in tumors early during combined treatment and later following escape from immune control.

B16F10 tumor-bearing mice (  ~50 mm³, n=5) were injected i.v. with 10⁷ cfu of YS1646-scrambled, -shSTAT3, or PBS. Four days later, mice were then gavaged with 10⁷ cfu of MVP728-3342Max, -2810, or PBS. For early detection of survivin, B16F10 tumors were excised from mice seven days after vaccination (d15) and then homogenized and lysed for western blot analysis. For detection of survivin after escape (esc.) in the shSTAT3+3342Max group, tumors were excisedi from mice 2 weeks after vaccination (d23) and then homogenized and lysed for western blot analysis. Equal amounts of each lysate were loaded and b-tubulin was used as a loading control.

This work was supported by 5P01-CA030206-28-Prj3, NIH-RAID Administrative Supplement of 3P01-CA030206-28S2 and ThinkCure Foundation (DJD).
A Minority Supplement of 3P01-CA030206-28S3 (EM).
COH Cancer Center is supported by 5P30-CA033572-27.

Acknowledgements:

The authors thank D. Castanotto and M. Kortylewski for advice on shRNA and STAT3.
 

This pioneering study by Edwin Manuel , Céline Blache, Rebecca Paquette, Teodora Kaltcheva, Hidenobu Ishizaki, Joshua Ellenhorn, Michael Hensel, Leonid Metelitsa, and Don Diamond reveals some of the details needed in designing an effective vaccine against an individual patient's neoplasm. In this study of mouse melanoma, and important and necessary component is a specific short RNA species, reminiscent of the immune RNAs discovered previously in T-lymphocyte-mediated immune reactions.

Additional References:

1. Kern DH, Chow N, and Pilch YH, "Lymphocyte Populations Participating in Cellular Antitumor Immune Responses Mediated by Immune RNA", J. Natl. Cancer Inst. 60, 335-344 (1978).

2. Dudley ME, Gross CA, Langhan MM, Garcia MR, Sherry RM, Yang JC, Phan GQ, Kammula US, Hughes MS, Citrin DE, Restifo NP, Wunderlich JR, Prieto PA, Hong JJ, Langan RC, Zlott DA, Morton KE, White DE, Laurencot CM, and Rosenberg SA
"CD8+ Enriched “Young” Tumor Infiltrating Lymphocytes Can Mediate Regression of Metastatic Melanoma".

3. Phan GQ, Yang JC, Sherry RM, Hwu P, Topalian SL, Schwartzentruber DJ, et al.Cancer regression and autoimmunity induced by cytotoxic T lymphocyte-associated antigen 4 blockade in patients with metastatic melanoma. Proc Natl Acad Sci U S A 2003;100:8372–7.

4. Coughlin CM, Vance BA, Grupp SA, and Vonderheide RH, "RNA-transfected CD40-activated B cells induce functional T-cell responses against viral and tumor antigen targets: implications for pediatric immunotherapy", Blood First Edition Paper, prepublished online November 20, 2003;
DOI 10.1182/blood-2003-07-2379.

5. Liao X, Li Y, Bonini C, Nair S, Gilboa E, Greenberg PD, and Yee C, "Transfection of RNA Encoding Tumor Antigens Following Maturation of Dendritic Cells Leads to Prolonged Presentation of Antigen and the Generation of High-Affinity Tumor-Reactive Cytotoxic T Lymphocytes"., Molecular Therapy, vol. 9, no. 5, pp. 757-764 (May, 2004).

1. Each cell retains all of its embryonic genes for a lifetime.

2. Controls for embryonic genes are often absent in adults.

3. Uncontrolled embryonic genes can replicate wildly.

4.  Replicating genes participate in  intra-cellular competition.

5.  The basis for gene competition is selective transcription.

6.  MicroRNAs can reprogram embryomic transcription.

7.  Gene reprogramming can produce normal phenotypes.

8.  Normal phenotypes can by-pass chromosomal lesions.

9.  MicroRNA therapy may need to be permanent.

10. Transplantation of microRNAs could be preferred.

http://www.embryomas.net/

1. Pathways within cell genomes involve a flow of information.

2. Information can flow by direct contact or by third parties.

3. Direct contact within whole genomes is difficult to regulate.

4. DNA-DNA direct contects are influenced by agents.

5. Nuclear agents include hydrophilic ionic and hydrophobic conforming ligands.

6. Third parties within genomes involve RNAs and proteins.

7.  RNAs and proteins are easy to regulate or reverse.

8.  Information can be shared, lost, or transformed.

9. System information can be hidden during system isolation.

10.  Local information can be permanently lost during system entropy.

http://www.cancerbiophysics.net/

Links to Current Research in Euchromatin:
Links to Euchromatin Activator RNA Reviews:
Links to Euchromatin Activator RNA Research:
Links to Ultrastructural Probes of DNase I-Sensitive Sites:
Links to RNA as a Therapeutic Agent:
Links to Hodgkin Lymphoma Immuno-Pathology:
Links to Activated T-Lymphocyte Immunotherapy:
Links to Medical Systems Biology:
Links to Selective Gene Transcription:
Links to RNA-Induced Epigenetics:
Links to RNA-Induced Embryogenesis:
Links to RNA and Biological Causality:
Links to Reprogramming and Neoplasia:

A Brief History of Activator RNA:

"Ultrastructural Probes of Active DNA Sites, and the RNA Activators of DNA".
(PowerPoint Presentation).

Top of Page - Euchromatin Network -  Euchromatin Research -  Research in Quantitative Radiology

For Further Information and Feedback:

Jeannette A. Hovsepian, M.D.
E-mail: frensasc@ix.netcom.com
Phone:  +1 650 367 6483