|Home | About | Journals | Submit | Contact Us | Français|
An anti-tumor activity associated with several bacterial pathogens, including Salmonella enterica serovar Typhimurium, has been reported; however, the underlying immunological mechanism(s) that leads to an anti-tumor effect is currently unclear. Further, such pathogens cannot be used to suppress tumor growth because of their potential for causing sepsis. Recently, we reported the characterization of S. Typhimurium isogenic mutants from which the Braun lipoprotein genes (lppA and B) and the multicopy repressor of high temperature requirement (msbB) gene were deleted. In a mouse infection model, two mutants, namely, lppB/msbB and lppAB/msbB, minimally induced proinflammatory cytokine production at high doses and were non-lethal to animals. We demonstrated that immunization of mice with these mutants, followed by challenge with the wild-type S. Typhimurium could significantly suppress tumor growth, as evidenced by an 88% regression in tumor size in lppB/msbB mutant-immunized animals over a 24-day period. However, the lppAB/msbB mutant alone was not effective in modulating tumor growth in mice, while the lppB/msbB mutant alone caused marginal regression in tumor size. Importantly, we demonstrated that CD44+ cells grew much faster than CD44− cells from human liver tumors in mice, leading us to examine the possibility that S. Typhimurium might down-regulate CD44 in tumors and splenocytes of mice. Consequently, we found in S. Typhimurium-infected mice that tumor size regression could indeed be related to the down-regulation of CD44high and CD4+CD25+ Treg cells. Importantly, the role of lipopolysaccharide and Braun lipoprotein was critical in S. Typhimurium induced-anti-tumor immune responses. Taken together, we have defined new immune mechanisms leading to tumor suppression in mice by S. Typhimurium.
Salmonella enterica serovar Typhimurium has been shown to possess anti-tumor activity in a mouse model, the capability to selectively amplify within tumors, and express therapeutic proteins 1, 2. Reports have indicated that S. Typhimurium selectively infects and preferentially colonizes solid tumors of cancer patients 3–5, making it potentially useful as a vehicle to target human tumors in vivo 1, 2, 6–8. Further, S. Typhimurium has a significant ability to infect non-phagocytic cells via expression of a type-III secretion system (T3SS), which facilitates bacterial penetration of host cells 9.
S. Typhimurium releases lipopolysaccharide (LPS) during both in vitro and in vivo growth 10, and its release is significantly enhanced during bacterial lysis following exposure to antibiotics or human serum. This enhanced LPS release causes septic shock 11. Likewise, Braun lipoprotein (Lpp) is also a critical bacterial component in the induction and pathogenesis of septic shock. Like LPS, it induces production of tumor necrosis factor (TNF)-α and interleukin (IL)-6 in mouse and human macrophages ex vivo 12, 13, and leads to lethal shock as a result of these cytokine production in both LPS-responsive and non-responsive mice 14. More importantly, Lpp synergizes with LPS to induce production of pro-inflammatory cytokines in mice, because Lpp binds to toll-like receptor (TLR)-2, whereas LPS binds to TLR-4 and CD14 to activate host cells 15. The biological potency of LPS can be significantly reduced when the multicopy repressor of a high temperature requirement (msbB) gene is deleted from S. Typhimurium 2. This gene codes for an enzyme which is responsible of adding myristic acid to the lipid A moiety of LPS 16. Consequently, in our recent studies, we characterized isogenic mutants of S. Typhimurium that were deleted for the Braun lipoprotein (lppA and B) genes in conjunction with the msbB gene 16. We provided evidence that such mutants were highly attenuated in a mouse model of salmonellosis and produced minimal levels of pro-inflammatory cytokines and chemokines 17. We predicted that the lpp/msbB mutants of S. Typhimurium would be excellent live-attenuated vaccine candidates. Indeed, mice immunized with such mutants were solidly protected against challenge with lethal doses of wild-type (WT) S. Typhimurium, in that the lppB/msbB and lppAB/msbB mutants provided maximum protection. These two mutants were further characterized in terms of their immune responses in a mouse model 17 and examined in this study for their anti-tumor activity.
CD4+CD25+ T regulatory (Treg) cells have been shown to be controlling self-reactive T cells by helping to maintain immunological self-tolerance 18, 19 and hence are a major obstacle in effective anti-tumor immunotherapy 20, 21. CD4+CD25+ Treg cells have been reported to be specifically recruited to tumor sites and to effectively block anti-tumor cytotoxic T-lymphocyte (CTL) responses. Hence, the targeted removal or inactivation of the CD4+CD25+ Treg cells in animal models could lead to improved tumor immuno-surveillance, better vaccine efficacy, and enhanced anti-tumor immunity 22–24. Further, CD44 is the principal receptor of the hyaladherin receptor family, and CD44-hyaluronan interactions mediate cell adhesion and migration in various physiological and patho-physiological processes 25–28. CD44 is expressed on many tissues in developing and adult humans, and it was described originally as a homing receptor required for binding of lymphocytes to high endothelial venules 29, 30 . CD44 has also been shown to be involved in lymphocyte activation in humans 31, 32and in a mouse model 33. CD44 expression by tumor cells can increase their interaction with endothelial cells and transmigration across an endothelial monolayer. Several reports have shown that CD44 inhibitors, including anti-CD44 monoclonal antibody and hyaladherin inhibitors, can be used to block tumor cell growth, invasion, and metastasis 34, 35.
In this study, we showed that lpp/msbB mutants alone were not effective in suppressing tumor growth in mice. However, we observed a significant regression in the size of tumors implanted in mice first immunized with the lpp/msbB mutants of S. Typhimurium and then challenged with WT S. Typhimurium. Further, we noted a decrease in CD4+CD25+ Treg cell numbers and down-expression of CD44high in the spleen of mice infected with S. Typhimurium, which for the most part could be mediated through LPS and Lpp.
The WT S. Typhimurium 14028 strain and its various mutants used in this study are listed in Table 1. The organisms were grown in Luria-Bertani (LB) broth or on LB agar plates in the presence of the appropriate antibiotics. For the growth of msbB mutants, we used a special MsbB medium as previously described 17. The MsbB medium/liter consisted of the following: 10 g tryptone, 5 g Yeast Extract, 1 ml 1 M MgSO4, and 1 ml 1 M CaCl2. The bacteria were cultivated at 37°C overnight with shaking at 200 rpm, harvested by centrifugation (6,000 rpm for 5 min), washed with phosphate-buffered saline (PBS), and resuspended in a minimal amount of PBS. Bacteria were counted by determining colony-forming units (cfu) in triplicate, and expressed as cfu/ml.
We used 6- to 8-week-old C57BL/6, BALB/c, and/or Swiss-Webster female mice (Taconic Farms, Germantown, NY). For selected studies, we used breeding pairs of IL-10−/− mice in a C57BL/6 background (The Jackson Laboratory, Bar Harbor, ME), and IL-10−/− aged mice (2 years old). We used IL-10+/+ and IL-10−/− mice as IL-10 is a pleiotropic cytokine with anti-inflammatory and anti-angiogenic properties in in vitro and in vivo models 36, 37. High IL-10 levels result in smaller tumors, their lesser metastases and reduced angiogenesis in humans, and low IL-10 production is associated with increased risk of prostate cancer 38. Mice were kept under specific pathogen-free conditions in filter-topped cages with sterile bedding and fed sterile food and water.
Mice were infected/immunized via the intraperitoneal (i.p.) route with: WT S. Typhimurium14028, lppB/msbB mutant, and/or the lppAB/msbB mutant (Table 1) at doses ranging from 0.5 × 102−2 × 103 cfu/100 µl, and deaths were recorded for a 30-day period. In some cases, mice were initially immunized with the mutant S. Typhimurium strains and then rechallenged with the WT S. Typhimurium strain (2 × 103 cfu) 30 days after the initial immunization. The animal experiments were performed under the approved Institutional Animal Care and Use Committee and death was the end point of the study when animals were infected with the WT bacteria. However, those animals that were immunized with the mutant S. Typhimurium strains and then challenged with the WT bacterium survived and were humanely euthanized after 30 days.
At indicated time intervals following the infection/immunization of mice, spleens were removed and homogenized in PBS, and serial dilutions of the homogenates were plated on LB and Salmonella-Shigella agar plates to determine bacterial counts. The plates were incubated at 37°C for 24 to 48 h and colonies counted.
The B16F1 melanoma cell line from ATCC (American Type Culture Collection, Manassas, VA) was cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, 100 µg/ml streptomycin, and 2 mM of L-glutamine before injections into mice under an approved animal protocol. These cells were grown for 6 days at 37°C in the presence of 5% CO2 before injections.
Human liver cancer cells (from the Sealy Cancer Center, University of Texas Medical Branch, Galveston, TX) were stained with mouse anti human CD44 PE antibody (PE, clone 515) (BD Bioscience, Franklin Lakes, NJ). After washing with PBS, CD44+ and CD44− tumor cells were isolated by using a FACSAria (BD Bioscience). Each cell clone population was expanded in DMEM supplemented with 10% FBS, 100 U/ml penicillin, 100 µg/ml streptomycin, and 2 mM of L-glutamine in 96-well tissue culture microtiter U-bottom plates.
Six- to eight- week-old female C57BL/6 mice were subcutaneously (s.c.) injected with tumor cells 5 × 105/mouse/100 µl. Personnel blinded to the study design measured tumor growth by microcaliper 3 times a week. Tumor volume was calculated according to the formula: (length)×(width)×(width)/2.
Spleen cells were obtained from mice infected with WT or mutant strains of S. Typhimurium. Splenocytes (105) were suspended in PBS and 1% FBS, and incubated with mouse CD16/CD32 monoclonal antibodies (0.25 µg/100 µl) (BD Bioscience) for 15 min at room temperature. These antibodies were added to block antibody binding to mouse Fc-γ receptor-bearing cells. Then, the cells were washed twice with PBS plus 1% FBS. To quantitatite T cells, the cell suspension was incubated for 60 min at 4°C with anti-CD4 (BD Bioscience, clone GK1.5) conjugated to fluorescein isothiocyanate [FITC] (0.25 µg/100 µl), anti-CD44 (BD Bioscience, clone IM7) antibody conjugated to phycoerythrin (PE) (0.1 µg/100 µl), and anti-CD25 (BD Bioscience, clone 3C7) antibody conjugated to anti-allophycocyanin (APC) (0.1 µg/100 µl). Subsequent to incubation, the cells were washed twice with PBS and then analyzed by FACScan flow cytometry using CellQuest (Becton Dickinson, Mountain View, CA) software. The corresponding isotype controls (rat IgG1, IgG2a, and IgG2b) were purchased from eBioscience (San Diego, CA) and BD Bioscience, respectively.
At least three independent experiments were performed, and the data were analyzed using Student's t test. P values of ≤0.05 were considered significant. Statistically significant values are referred to as follows: *, p<0.05; **, p<0.01; ***, p<0.001.
We opted to immunize mice with various S. Typhimurium strains before transplantation of tumors, as we did not intent to develop tumor-antigen-specific immune/memory response, which could potentially interfere with S. Typhimurium-induced strong cellular and humoral immune responses. Further, as discussed in the next section, it was important to immunize animals with the mutant S. Typhimurium strains as we could use high doses of them without inducing any lethality and then challenge mice with a high dose of WT bacteria, which otherwise will be lethal to animals in an un-immunized group. Finally, people traveling from the developed world to developing countries are invariably vaccinated with the live-attenuated Salmonella strain and therefore it appeared logical to immunize animals with S. Typhimurium .
To delineate whether the WT S. Typhimurium or its mutant strains suppressed the growth of melanomas, mice were infected with 50 cfu/mouse of either the WT S. Typhimurium or its lppAB/msbB or lppB/msbB mutant strain. Mice injected with PBS served as a negative control. After 30 days, B16 cells (5 × 105) were subcutaneously injected into mice. The mean tumor size for various bacterial strains after 24 days was: WT S. Typhumirum, 640 +/−52 mm3; lppB/msbB mutant, 798 +/− 65 mm3; lppAB/msbB mutant, 871+/−62 mm3; and PBS, 912+/−74 mm3 (Figure 1A). As can be noted from this figure, tumor size was inhibited in WT S. Typhimurium-infected mice by a statistically significant 30% compared to the rate in animals that were given PBS alone. The decreased tumor size was only marginal (12%), however, in mice infected with the lppB/msbB mutant after 24 days, and non-existent for up to 21 days in animals infected with the lppAB/msbB mutant when compared to the PBS controls. A slight, but statistically insignificant decrease in tumor size occurred after 24 days with the lppAB/msbB mutant (Figure 1A).
Consequently, we further investigated, using a different approach, the anti-tumor effects of WT S. Typhimurium and its mutants. In this set of experiments, mice were immunized with the lppAB/msbB and lppB/msbB mutants of S. Typhimurium at a dose of 2 × 103 cfu. Our previous studies indicated that mice were fully protected against mortality by i.p. doses of up to 1 × 106−1 × 107 cfu of these mutants and demonstrated that animals immunized with the above mutants at doses of 1 × 106− 1 × 107 cfu were 100% protected against similar doses of the WT bacteria after rechallenge 17. After 30 days of immunization of these mice with the above mutants, we rechallenged them with the WT S. Typhimurium at a dose of 2 × 103 cfu. Two weeks after rechallenge of mice with the WT bacteria, B16 cells (5 × 105) were implanted s.c. into the animals. The mean tumor sizes in mice immunized with the mutants, followed by WT infection, were as follows: lppAB/msbB mutant alone, 863+/−71 mm3; lppB/msbB mutant alone, 660+/−49 mm3, lppAB/msbB + WT, 470+/−31 mm3, and lppB/msbB + WT, 110+/−9 mm3 (Figure 1B). The tumor size in mice injected with PBS alone was 903+/−86 mm3 and served as a positive control. It is evident from these data that tumor size regressed in the animals first immunized with the mutants and then challenged with the WT S. Typhimurium, and that lppAB/msbB and lppB/msbB mutants, respectively, provided 48% and 88% protection. The protection afforded by the lppB/msbB and lppAB/msbB mutants alone at a dose of 2 × 103 cfu was 27% and 5%, respectively (Figure 1B).
To define the underlying mechanism(s) involved in S. Typhimurium-induced tumor growth suppression, we infected Swiss-Webster mice with WT S. Typhimurium or its lppAB/msbB and lppB/msbB mutants (1 × 103cfu/mouse). After 5 days of infection, spleens from infected mice were stained with anti-CD4 FITC and anti-CD25 APC, and the splenocytes analyzed by flow cytometry. We found that the CD4+CD25+ Treg cells decreased significantly in the WT group of infected mice (0.5%) compared to uninfected animals (1.6%, a decrease of 69%) (Figure 2A). The number of CD4+CD25+ Treg cells decreased slightly by 31% in mice infected with the lppB/msbB mutant (1.1% versus 1.6% for control), and minimally affected (1.5%) in the lppAB/msbB-infected mice compared to the controls (1.6%). The percentage of CD4+ cells were slightly decreased in animals infected with the WT S. Typhimurium (18%) when compared to animals infected with either the lppB/msbB (21%) or the lppAB/msbB (22%) mutant and/or those mice that were only given PBS (24%). In addition we found that the expression of gene-encoding Forkhead box P3 (Foxp3), which controls Treg cell development and their functions 39, was down-regulated in the splenocytes of mice infected with WT S. Typhimurium (0.6%), when compared to splenocytes of control uninfected mice (1.8%). As a positive control, we used splenocytes treated with S. Typhimurium lipopolysaccharide (3 µg/ml), which, as expected, decreased the expression of the gene-encoding Foxp3 significantly (0.31%) compared to the controls (Figure 2B). Some down-regulation of T cell activation (20%) was also seen in splenocytes infected with the WT S. Typhimurium ex vivo.
An important question which should be considered is whether there are increases in T- and/or NK cell lytic activity against B16 melanoma cells in mice which were first immunized with the S. Typhimurium mutant (e.g., lppAB/msbB) and then challenged with the WT S. Typhimurium compared to those mice that were immunized with the mutant but challenged with the same lppAB/msbB mutant. The above-mentioned scenarios would lead to reduced tumor growth. Therefore, we immunized C57BL/6 mice with the lppAB/msbB mutant at a dose of 2 × 103 cfu. After 30 days, we challenged them with the WT S. Typhimurium at a similar dose of 2 × 103 cfu. After two weeks, B16 cells (5 × 105) were implanted s.c. into the animals, and three weeks later, tumor cells from these mice were stained with CD4-PE-Cy7, CD25-APC, NK1.1-FITC, CD11b-PE, CD3-PE, interferon (IFN)-γ–FITC antibody and analyzed by flow cytometry. Our data indicated that the number of CD4, CD4+CD25+, and NK cells as well as macrophages were increased to 1.8%, 3.2%, 2.6%, and 7%, respectively, in tumors of mice that were first immunized with the lppAB/msbB mutant before WT S. Typhimurium challenge. These numbers were in contrast to 0.1%, 0.4%, 0.5%, and 1.5% in those animals that were immunized with the mutant and then challenged with the mutant. Likewise, animals that were given only PBS had the following percentages of CD4, CD4+CD25+, NK, and macrophages: 0.1%, 0.3%, 0.2%, and 1.6%, respectively. Importantly, the percentage of IFN-γ producing T cells that infiltrated into tumor of mice that were first immunized with the lppAB/msbB mutant and then challenged with the WT S. Typhimurium was also much higher compared to in tumors of mice that were immunized with the lppAB/msbB mutant and subsequently challenged with the mutant. Animals given only the PBS but injected with B16 cells had baseline level of IFN-γ (Figure 3A–C).
We cultured B16 cells with WT S. Typhimurium or its lppAB/msbB and lppB/msbB mutants at a multiplicity of infection of 1 in DMEM at 37°C in the presence of 5% CO2. After 24 h of incubation, the cells were harvested and stained with mouse anti-CD44 PE and analyzed by flow cytometry. As noted from Figure 4A, WT S. Typhimurium-infected B16 cells exhibited a significant decrease in the expression of CD44 marker (69%) compared to that of control (83%). Likewise, the expression of CD44 showed a downward trend, with decreases to 74% and 71% in B16 cells infected with the lppAB/msbB and lppB/msbB mutant, respectively, compared to uninfected B16 cells (83%).
We also infected Swiss-Webster mice with WT S. Typhimurium or its lppAB/msbB and lppB/msbB mutants (1 ×103 cfu/mouse). After 5 days of infection, spleen cells were stained with anti-CD44 PE, and analyzed by flow cytometry. As noted from Figure 4B, the splenocytes of WT S. Typhimurium-infected group exhibited a significant decrease (3.4% versus 12.6%; indicated by an arrow) in the expression of CD44high, followed by the lppB/msbB (6.5%) and lppAB/msbB (9.7%) mutants compared to the control splenocytes from uninfected mice.
Splenocytes from naive C57BL/6 and IL10−/− aged mice were stained with mouse anti-CD4 FITC, anti-CD25 APC and anti-CD44 PE antibodies and analyzed by flow cytometry. R1 cells were gated as lymphocytes and R1+R2 as CD44high cells, while R1+R3 represented CD44med cells (Figure 5A). CD44high cells contained much higher levels of CD4+CD25+ Treg cells (1.82%), compared to the levels in the CD44med (1.13%) and control splenocytes (1.61%) of naive C57BL/6 mice (Figure 5B). In IL-10−/− aged mice, CD44high cells contained much higher levels of CD4+CD25+ Treg cells (2.3%) compared to the levels in CD44med (0.7%) and control splenocytes (1.8%) (Figure 5C).
Splenocytes from C57BL/6 and IL-10−/− aged mice were stained with mouse anti-CD44 PE antibodies. The splenocytes from C57BL/6 aged mice expressed higher levels of CD44 than did those from the young mice (Figure 6A; 2.06% versus 3.97%). The splenocytes from IL-10−/− aged mice expressed even higher levels of CD44 than did the IL-10−/− young mice (Figure 6B; 2.5% versus 13.6%). We then subcutaneously implanted B16 tumor cells in IL-10−/− aged mice. After 24 days, mice with bigger or smaller tumors were killed and their splenocytes stained with mouse anti-CD44 PE antibodies. In general, the mice having the larger tumors expressed higher levels of CD44high (12.5% versus 8.4%) than did animals with smaller tumors (Figure 6C).
We sorted CD44+ and CD44− cells from human liver tumors and grew them in DMEM. After expansion of the cell colonies in 96-well microtiter plates, the cells were transferred to 25-cm2 cell culture flasks. Six days later, the CD44+ and CD44− tumor cells (5 × 105/100 µl) were injected s.c. into BALB/c mice. The mean tumor size for the CD44+ group was 1075 +/−175 mm3 compared to 62 +/− 42 mm3 for the CD44− group on day 20 after injection (Figure 7). These data indicated that CD44 was required for tumor growth and represented an important tumor marker.
The exact mechanism by which Salmonella strains result in an inhibition of tumor growth is still unclear. Some earlier studies reported the construction of mutant strains of S. Typhimurium, such as msbB and auxotrophic mutants that were nonpathogenic for mice, pigs, and humans. These were also demonstrated to have accumulated 1000-fold more in tumors than in other organs. However, the clinical application of such strains did not successfully curtail tumor growth 2, 40.
In this study, we showed that the lppAB/msbB mutant did not suppress tumor growth, while the lppB/msbB mutant did somewhat slow-down the growth of the tumors compared to those of controls in the 24 days’ observation (Figure 1A). Importantly, tumor size in mice immunized with either of the mutants and then infected with the WT bacteria was significantly reduced (Figure 1B). These data indicated that the S. Typhimurium mutants that were deleted for the msbB and lpp genes lost this tumor repressor function, and LPS and Lpp were critical for S. Typhimurium’s anti-tumor function.
The CD4+CD25+ Treg cells play a crucial role in tumor immune pathogenesis, and they also modulate immune therapeutic efficacy 41–44. We explored the mechanism(s) underlying an S. Typhimurium-mediated, anti-tumor effect, and found that CD4+CD25+ Treg cells from spleens of mice were significantly decreased (69%) in the WT S. Typhimurium-infected mice and decreased by 31% in lppB/msbB mutant-infected mice (Figure 2A). This decrease in Treg cell development correlated with the decreased Foxp3 expression in the splenocytes of mice infected with the WT S. Typhimurium (Figure 2B). LPS has been shown to reduce expression of Foxp3 in lymphocytes 45, 46. Our data presented in Figure 2A and Figure 4B tend to suggest that the expression of CD44high on T cells could in part be related to both T cell activation and Treg cell population. We must emphasize here that for some experiments we preferred to use out-bred Swiss-Webster over in-bred C57BL/6 mice as the latter are more sensitive to S. Typhimurium infection, since as little as 50 cfu resulted in the deaths of 30–50% of these mice. Therefore, in those experiments in which higher doses of bacteria were used, we used Swiss-Webster mice. We should cautiously interpret data as different strains of mice might behave differently to S. Typhimurium infection as it relates to tumor growth.
S. Typhimurium-induced CD8 CTL and NK cells can kill tumor cells directly 47–49. Likewise, S. Typhimurium promotes the maturation of dendritic cells (DCs) and enhances the ability of antigens to present to DCs 50. It has been shown that splenic T cells are broadly activated in the host later during Salmonella infection 51. Further, salmonellae-induced IL-12 and IFN-γ are two critical anti-tumor cytokines 52, and IFNγ-producing T cells are significantly elevated after primary infection 53. Our results showed that CD4, CD4+CD25+, and NK cells as well as macrophages were significantly increased in the tumors of mice that were first immunized with the lppAB/msbB mutant and then challenged with the WT S. Typhimurium compared to those animals that were only immunized or given PBS (Figure 3). It is known that CD4 and NK cells could inhibit tumor growth 54, 55, but the function of CD4+CD25+cells in modulating the growth of tumors is not clear. Based on our data, CD4+CD25+ cells appeared to be recruited into tumors but they might not inhibit tumor growth directly but might prevent suppressor CD4 T cells to enter tumors. We found CD4+CD25+ cell numbers to be lower in smaller size tumors than in bigger size tumors after infection of mice with the WT S. Typhimurium (data no shown). Further, IFN-γ producing cells were significantly increased in tumors that were first immunized with the lppAB/msbB mutant and then challenged with the WT S. Typhimurium, suggesting that Th1 response might be important for the anti-tumor response.
CD44 has been shown to be highly expressed in many tumor cells 56 such as colorectal cancer stem cells, and expression of CD44 in immortalized cancer cell lines is of functional importance 57, 58. CD44 has recently been shown to be an important marker of cancer stem cells or cancer-initiating cells59, 60. Both CD44 and its variant (vCD44) when expressed in cancer cells, correlate strongly with invasiveness, metastasis, and tumor growth 61, 62. Likewise, CD44+ cancer stem cells have been shown to have enhanced potential for proliferation, migration, and invasion 63. We noted that CD44 was expressed at a significant level in B16 tumor cells. However, its expression was decreased in B16 tumor cells infected with S. Typhumirum, albeit to different levels with WT versus the mutants (Figure 4A).
WT S. Typhimurium strains used in this study did not only directly decrease CD44 expression of tumor cells in vitro but also CD44 expression of lymphocytes in vivo (Figure 4B). It has been shown in the literature that LPS induces CD44 expression in monocytes 64,65, however, then why did CD44 cells decrease after S. Typhimurium infection in our study? We believe that activation-induced death of splenic cells could contribute to the observed CD44 cell loss during S. Typhimurium infection. The role of CD44 expression in the development of Treg cells has been reported 66, and therefore, we analyzed CD44high cells of spleens for the presence of CD4+CD25+ Treg cells (Figure 5A). We found that CD44high cells contained much higher levels of CD4+CD25+ Treg cells than did the CD44med cells in C57BL/6 mice (Figure 5B). Likewise, CD44high cells contained much higher levels of CD4+CD25+ Treg cells than did CD44med cells in IL-10−/− mice (Figure 5C).
IL-10 has been shown to be associated with both immune stimulation and suppression of tumor growth. A high incidence of colorectal adenocarcinomas (60%) was observed in IL-10−/− aged mice. The IL-10−/− mouse is a model for colon cancer, which develops via the dysphasia sequence, a process similar to that seen in clinical inflammatory bowel disease-associated cancer 67, 68. Therefore, we next examined IL-10−/− aged mice and found increasingly higher CD44high levels in the splenocytes of aged IL-10−/− mice than young mice (Figure 6B). Moreover, when we injected B16 tumor cells s.c. in IL-10−/− aged mice and selected a few mice having either large or small tumors, we found a higher expression of CD44high in the spleens of mice having larger tumors than in those with smaller tumors (Figure 6C). Aged mice have reportedly expressed higher levels of CD44high than do the younger mice (about two fold) (Figure 6A). For example, CD44high expression was greater in aged SPA-1−/− mice developing late onset of myeloid leukemia 69. Thus, CD44high cells may contribute to tumor occurrence and development. Therefore, the relationship of CD44high expression of lymphocytes with that of tumor growth should be further studied in detail.
To confirm that CD44 is important for tumor growth, we sorted CD44+ and CD44− cells from human liver tumor samples. We found CD44+ tumor cells grew much faster compared to CD44− tumor cells in vivo (Figure 7); however, their growth rates were quite similar in vitro. These data suggested to us a dependence on CD44 in tumor development in vivo. CD44 is a cell surface proteoglycan, thought to be involved in cell-to-cell adhesion and cell matrix-adhesion interactions, lymphocyte activation, and homing and cell migration. We believe S. Typhimurium suppresses tumor growth through various direct and indirect pathways, and that our data provide indication that it directly down-regulated the expression of CD44 in tumors in vitro and in lymphocytes in vivo. Further, S. Typhimurium seemed to enhance the anti-tumor response by reducing the numbers of CD4+CD25+ Treg cells.
We noted that the lppAB/msbB and lppB/msbB mutants induced a very low Th1 response in mice, including minimal interferon (IFN)-γ and TNF-α production and higher levels of IgG1 antibody production, compared to that in the WT S. Typhimurium-treated mice which exhibited the opposite trend 17, 70. These data suggested that the Th1-induced cellular responses were more important than the antibody (humoral) responses in anti-tumor immune mechanisms. Would such mutants accordingly be beneficial in controlling tumor growth in humans? The advantage of immunizing first with the mutants is that subsequently higher doses of the WT bacteria can be given without danger of inducing sepsis in the host. Certainly, this adjunct therapy would then seem to be advantageous, as the virulence of the WT S. Typhimurium is likely seems to be essential in retarding tumor growth in the host. The lpp/msbB mutants in general and the lppAB/msbB mutant in particular caused a significantly increase in the down regulation of CD44 and the number of CD4+CD25+ Treg cells, when compared to the WT S. Typhimurium in the splenocytes of infected mice (Figure 2 and Figure 4). These data tend to suggest the role of LPS and Lpp in the down regulation of CD44 and Treg cells, with an ultimate reduction in the tumor suppression function of lpp/msbB mutants (Figure 1). An earlier study by Low et al. (2) reported that S. Typhimurium msbB mutant did not alter anti-tumor activity against s.c. implanted B16F10 melanomas 2. Therefore, Lpp might be more important in inducing an anti-tumor effect than LPS and would be studied in the future.
Recently, it was demonstrated that anti-tumor activity of Salmonella strains correlated with decreased angiogenesis and increased tissue necrosis within the tumor tissue 71. Their results showed that Salmonella could directly inhibit tumor growth. Our data pointed to the fact that S. Typhimurium did not only directly inhibit tumor growth but also inhibited tumor growth through indirectly pathways including changes in the host immune system, specifically by down-regulating CD44 and CD4 T reg cell numbers in the spleen of mice.
In this study, we used S. Typhimurium mutants lppAB/msbB and lppB/msbB as adjuncts to help WT S. Typhimurium suppress tumor cell growth. We found tumor growth suppression by 88% in mice that were first immunized with the lppB/msbB mutant and then challenged with 2 × 103 cfu of WT S. Typhimurium. Our data provided possible mechanisms by which S. Typhimurium could modulate tumor growth and include down regulation of CD44 and CD4+CD25+ Treg cell expression, which would, in turn, up-regulate host immune responses to inhibit tumor growth that is largely LPS- and lipoprotein- dependent. The use of mutant S. Typhimurium to suppress tumor growth might provide attractive alternatives for the development of immunotherapy for cancer.
We thank Ms. Mardelle Susman for assisting in manuscript preparation.
This study was supported by NIH grant AI064389 to A. K. Chopra.
The authors have no financial conflict of interest.