The lung is a common site of metastatic and primary tumor growth, and has been shown to be an immunosuppressive environment. We tested the impact of the lung environment on the development of tumor-specific T cell responses against the CMS5 fibrosarcoma, and found a deficit in the efficacy of naive tumor-specific DUC18 T cells against tumors established in the lung. One hundred-fold more naive tumor-specific T cells were required to protect against tumor development or reject established tumors in the lung than an identical tumor challenge delivered s.c. in the flank. Importantly, CMS5 growing in the flank facilitated the rejection of tumors present in the lungs. In the presence of flank tumors, transferred T cells were not phenotypically altered but were present in much greater numbers in the parabronchial lymph nodes, bronchoalveolar lavage, and lung parenchyma than in mice bearing lung tumors alone. We hypothesized that APC present in the lung and skin draining lymph nodes were differentially initiating T cell proliferation, leading to differences in the size of the final effector populations. A direct comparison of DUC18 T cell proliferation against APC from flank or lung draining lymph nodes showed profoundly greater proliferation to flank draining lymph node APC. The impaired stimulation of naive T cell proliferation by lung draining APC provides one mechanistic explanation for the lower overall immune response, and inability to effectively reject tumors, in the lung.
Tumors develop in and metastasize to a wide variety of tissues. To date most research in tumor immunology has focused on studying the direct interactions of the tumor and the host immune response. There has been relatively little study on the effects of the surrounding normal tissue on the initiation of tumor-specific T cell responses, although T cell trafficking and accessibility to tumor Ags has been shown to be key in the initiation of a successful tumor-specific immune response (1, 2, 3).
A number of models have been developed to study the interactions between tumors and host immune systems. Most of these models involve transferring T cell clones or in vitro activated transgenic T cells into mice to treat transplantable tumors that have been transfected with cognate Ag. A number of these studies have shown an inability of these tumor-specific T cells to reject established tumors, or to protect against subsequent tumor challenge. In a variety of models using different transgenic T cells and various transplantable tumors, T cells have shown anergy, ignorance, and exhaustion, but very rarely complete efficacy (4, 5, 6, 7, 8). A few groups have had success treating established tumors using either in vivo activated CTL or T cell clones (9, 10, 11, 12). Specifically, Dobrzanski and colleagues (10, 11) can retard the growth of B16-OVA that have been established in the lung for as long as 14 days using activated, polarized OT1 T cells. Ryan et al. (13) were able to decrease tumor burden in mice bearing 3-day established CMS4 in the lung by using a CTL line specific for an endogenous rejection Ag. Recently, ex vivo expanded tumor infiltrating lymphocytes have been used in conjunction with IL-2 and lymphoablation in patients to cause significant regression of metastatic melanoma (13). In all of these studies, however, naive cells either could not be obtained (9, 13) or were of no effect (11), leaving the question of what factors affect the initiation of a tumor-specific T cell response unanswered.
Our laboratory has developed a system in which naive T cells can have an effect against tumors in vivo. CMS5 is a transplantable fibrosarcoma specifically recognized and destroyed by C18 T cells. The C18 CD8+ T cell clone is specific for an epitope generated by a mutation in the extracellular signal-regulated kinase 2 (Erk2)3 gene in CMS5 termed tErk (14). The TCR from C18 was cloned and used to generate a transgenic mouse, DUC18. This mouse is resistant to s.c. challenge with CMS5, and transfer of naive DUC18 T cells to syngeneic BALB/c mice can both protect against subsequent tumor challenge and reject established tumors (15). Because naive cells are readily available from DUC18 transgenic mice, and the CMS5 tumor can grow in various tissues, we found this an ideal system to study the effects of the normal tissue environment on the initiation of a tumor-specific T cell response.
The lung is a site of both primary and metastatic tumor growth. It has also been shown to be a unique and immunosuppressive environment, with detectable background levels of IL-10, and a unique population of APCs (16, 17). The lung therefore seemed an appropriate environment to study the effects of normal tissue environment on the initiation of a tumor-specific immune response.
We found that CMS5 tumors growing in the lung following i.v. tumor delivery were poor targets for naive DUC18 T cells. Many more T cells were required to protect mice from i.v. tumor challenge than an equivalent s.c. challenge. The simultaneous challenge of mice with s.c. and i.v. tumors showed that naive T cells were not rendered dysfunctional by the tumors in the lung, as the s.c. tumors were rejected. On the contrary, the presence of s.c. CMS5 facilitated the rejection of the tumors in the lung. We observed a greater presence and organization of adoptively transferred T cells in the lung tissue of mice bearing lung and flank tumors. T cells from mice challenged with lung and flank tumors vs lung tumors alone were phenotypically similar. However, mice challenged with lung and flank tumors and treated with T cells had parabronchial lymph nodes (PBLN) 3–4 times the size of the PBLN in mice challenged with lung tumors alone. Double-challenged mice also had many more tumor-specific lymphocytes present in their bronchoalveolar lavage (BAL), indicating a much more robust immune response in the lung. The lung appears to be an inefficient site of naive T cell priming, allowing some percentage of tumors to escape complete CTL lysis, whereas the skin is an optimal site for T cell priming, enabling the rejection of even a second tumor challenge in a remote site.
Materials and Methods
BALB/c mice were purchased from the National Cancer Institute (Bethesda, MD). DUC18 transgenic mice were generated by our laboratory (15) and maintained in the Washington University School of Medicine (St. Louis, MO) specific pathogen-free animal facility. BALB/c Thy 1.1 mice were the kind gift of Hyam Levitsky (Johns Hopkins University, Baltimore, MD).
CMS5 (18) was freshly thawed from a frozen stock for every experiment and expanded through no more than three passages in RPMI 1640 supplemented with 10% FCS (HyClone Laboratories, Logan, UT), 1 mM glutamax (Life Technologies, Grand Island, NY), and 50 μg/ml gentamicin (Life Technologies).
For adoptive transfer, 0.5–1 × 106 splenocytes were stained for CD8 and Thy 1.2 using directly conjugated primary Abs. For ex vivo phenotypic analysis, samples were stained for Vβ8.3, CD62L, CD25, or CD69 using directly conjugated mAbs, or CD102 or P-selectin ligand using biotinylated primary and streptavidin-PE secondary, or a murine P-selectin/human IgG fusion protein and a PE-conjugated goat anti-human IgG secondary, respectively. All Abs were purchased from BD PharMingen (San Diego, CA), with the exception of the PE-conjugated polyclonal goat anti-human IgG (ICN, Aurora, OH). Data were collected on a FACScan or a FACSCalibur (BD Biosciences, San Jose, CA) and analyzed using CellQuest software. A total of 2 × 104 live lymphocytes were gated using forward light-side light scatter parameters for adoptive transfer analysis. For phenotypic analysis 5 × 105 events or the entire sample were collected.
CMS5 were harvested by trypsinization, counted, and resuspended in HBSS. A total of 3 × 106 CMS5 cells were injected either i.v. or s.c. Naive DUC18 T cells were obtained as fresh DUC18 transgenic splenocytes, which were mechanically disrupted and analyzed by flow cytometry for the percentage of Vβ8.3+ CD8+ cells.
All T cell transfers were i.v. via the lateral tail vein. Except where specifically mentioned, a dose of 3 × 106 cells was used for all T cell transfers.
Tumor incidence scoring
Mice were challenged with tumors and treated as described. Five weeks following i.v. tumor transfer mice were sacrificed and their lungs perfused with 10% buffered formalin (J.T. Baker, Phillipsburg, NJ). Lungs were removed and immersed in 10% buffered formalin for at least 24 h. Lungs were embedded in paraffin, cut to 10-μM sections, and stained using H&E. Coded lung sections were scored microscopically for the presence of visible tumor masses and scored accordingly as tumor free or tumor bearing. Tumors that were s.c. were measured every other day for the duration of the experiment. Tumor area was calculated as the product of two orthogonal diameters as measured using calipers. Mice were sacrificed after 21 days or if tumor area exceeded 400 mm2.
BALB/c-Thy 1.1 mice were challenged with CMS5 as described and treated 3 days afterward with DUC18-Thy 1.2 splenocytes. Five days after the transfer of DUC18 splenocytes, mice were sacrificed and the lungs perfused with HBSS and harvested. The lungs were submerged in OCT and snap frozen in liquid nitrogen-cooled dimethylbutane. Sections (8-μM) were cut and stained using biotinylated anti-Thy 1.2 mAb (BD PharMingen) and streptavidin-PE secondary (Caltag Laboratories, Burlingame, CA). Photographs were taken on a Nikon E800 microscope using Act-1 software.
Lymphocyte phenotypic analysis
DUC18 T cells were transferred as previously described into mice bearing 3-day established lung and flank tumors or lung tumors alone. Five days following T cell transfer, mice were sacrificed and spleens, left inguinal lymph node (LILN, contralateral to flank tumors), and the primary PBLN were removed. A BAL was extracted by washing the lungs three times with 1 ml of HBSS. Lungs were harvested and digested in collagenase IA for 90 min at 37°C. Supernatant was separated from the remaining tissue pieces and centrifuged to obtain cells. Lymph node and spleen samples were individually counted using a hemocytometer. BAL and lung infiltrating lymphocytes were counted by running the entire sample through the FACSCalibur. For all organs tested, samples from five different mice per treatment were pooled and stained for FACS analysis.
Groups of three or more BALB/c mice were challenged with 3 × 106 CMS5 i.v. or s.c. Three days later, the draining lymph nodes were removed, pooled from each treatment group, and dissociated to form a single cell suspension. Cells were stained with 14-4-4 Ab (I-Eα specific) and FACS analyzed for MHC class II expression. Unfractionated lymph node cells were then plated in 96-well plates to the indicated number of MHC class II+ cells per well. A total of 1 × 105 DUC18 splenocytes were also plated in each well. Cells were cultured in RPMI 1640 supplemented with 10% FCS (HyClone), 1 mM glutamax (Life Technologies), 5 × 10−5 M 2-ME, and 50 μg/ml gentamicin (Life Technologies). After 48 h in culture, cells were pulsed with 0.4 μCi [3H]thymidine (Perkin-Elmer Life Science, Boston, MA), and harvested the next day using a Skatron 96-well harvester.
Where appropriate, group data were compared by Fisher’s Exact test or Student’s t test using www.matforsk.no/ola/fisher.htm or Microsoft Excel, respectively.
Our laboratory has previously characterized a tumor model in which naive transgenic, CMS5-specific, DUC18 T cells cannot only protect BALB/c mice from subsequent tumor challenge, but can even reject s.c. CMS5 fibrosarcomas established for up to 5 days (15). We wished to compare the efficacy of T cell responses vs s.c. tumors to tumors growing in other anatomical locations.
A number of tumor lines, including fibrosarcomas, have been shown to form lung nodules following i.v. injection. We found that 90–100% of mice challenged with 3 × 106 CMS5 i.v. developed multifocal tumors in the lungs visible upon gross or histological examination 5 wk after challenge, and this dose was used in all subsequent experiments. Tumors occur as multifocal masses of varying size and number. Approximately 20% of the mice developed additional tumor growth in the heart. Liver, spleen, brain, and kidney showed no signs of tumor growth either grossly or histologically. DUC18 T cells were capable of specifically eliminating i.v. transferred CMS5, as DUC18 transgenic mice were completely resistant to i.v. CMS5 challenge, but not to challenge with an irrelevant fibrosarcoma, Meth A (data not shown).
Naive DUC18 T cells are less effective in the lung than in the flank
To determine whether more DUC18 T cells were required to protect against i.v. tumor challenge than s.c., 3 × 104−3 × 106 naive DUC18 T cells were transferred into BALB/c mice, which were then challenged i.v. or s.c. with CMS5 cells 24 h later. Transfer of 3 × 106 DUC18 naive T cells was sufficient to provide some protection against subsequent tumor challenge (Fig. 1,A). This finding indicates that naive DUC18 T cells are effective against i.v. transferred CMS5. Although naive cells can protect against i.v. challenge, a s.c. challenge requires 10–100-fold fewer T cells for protection (Fig. 1 B). Thus, when the T cells are present before the tumor challenge, there is a difference in the efficacy of the DUC18 T cells against i.v. transferred CMS5 and s.c. tumor challenge.
Naive cells do not efficiently reject established lung tumors
To determine whether there was also a relative deficiency in the ability of DUC18 T cells to reject established CMS5 in the lung, mice were challenged with CMS5 i.v., then received 3 × 106 naive DUC18 T cells i.v. 1, 3, or 5 days later. At 5 wk after tumor transfer, mice were sacrificed and tumor incidence was assessed. T cell transfer within 24 h of tumor challenge led to the rejection of all tumors in 90–100% of mice. T cells were also able to eliminate the lung tumors in 70% of the mice treated 3 days post tumor challenge (Fig. 1 C). Previous studies have shown that 100-fold fewer T cells have the same level of efficacy against comparable s.c. tumors (15). These data further reinforce the disparity between T cell efficacy against tumors developing in the lung vs the skin.
Presence of s.c. tumor facilitates the rejection of pulmonary tumors
To assess whether the presence of lung tumors caused a global defect in the activation of naive DUC18 T cells, mice were challenged simultaneously with 2 doses of 3 × 106 CMS5 cells, one given i.v. and the other s.c. Mice received naive T cells 3 days later. Flank tumor size was recorded over the course of the experiment (Fig. 2,A). Ninety percent of the flank tumors were completely rejected, indicating that the naive T cells were still able to become activated in the presence of CMS5 in the lung. When the lungs of these mice were assessed for tumor growth, none of the mice that had been challenged with CMS5 both i.v. and s.c. showed any signs of tumor in the lungs (Fig. 2 B). Not only were the T cells able to reject the double tumor burden, it appeared that the presence of the s.c. tumor actually augmented the efficacy of the naive T cells against tumors in the lung.
Magnitude of the pulmonary immune response is increased in mice bearing lung and flank tumors
The presence and phenotype of transferred T cells was assayed to discern the mechanism for the augmentation of the pulmonary immune response in mice bearing i.v. and s.c. tumors. DUC18-Thy 1.2 splenocytes were transferred into BALB/c-Thy 1.1 congenic mice bearing 3 day established lung and flank tumors or lung tumors alone. Five days following T cell transfer, lungs were assayed for the presence of Thy 1.2+ cells by immunofluorescence. The lungs of mice with lung and flank tumors showed numerous large clusters of transferred T cells. Although there were transferred T cells present in the lungs of mice that had been challenged with lung tumors alone, the T cells were fewer in number and more sparsely distributed (Fig. 3). This pattern was observable in histological sections taken from all 14 mice examined in two independent experiments (seven mice total per challenge treatment).
To distinguish whether the flank tumors generated effector T cells with a different phenotype than lung tumors, the surface phenotype of the DUC18 T cells was analyzed by FACS. Five days following Thy 1.2 T cell transfer into Thy 1.1 mice bearing 3-day established lung tumors or both lung and flank tumors, the contralateral inguinal lymph node, spleen, and PBLN were removed for analysis. A BAL was taken, and the lung tissue removed for collagenase digestion and lymphocyte analysis.
The greater cellularity of the PBLN, BAL, and lung digests in the mice with lung and flank tumors indicated a more robust immune response taking place in the lungs of these mice (Fig. 4,A and data not shown). This increased response was limited to the tumor draining lymph nodes, as there was no difference in the cellularity of the contralateral inguinal lymph nodes of mice that received lung and flank tumors vs lung tumors alone (Fig. 4,B). Analysis of the cells in the BAL from mice bearing lung and flank tumors or lung tumors alone showed that a greater percentage of the lymphocytes present in the BAL of mice bearing lung and flank tumors were transferred Thy 1.2+ T cells (Fig. 4,C). FACS analysis of Vβ8.3 expression revealed that 100% of the transferred (Thy 1.2+) CD8 T cells present in all of the BAL samples and T cells extracted from the lung by trypsinization were DUC18 T cells (Fig. 5 and data not shown). To determine activation state, BAL cells were analyzed for surface activation markers CD25, CD62L, and CD69. All the markers tested were expressed at similar levels in all comparable cell populations in the mice challenged with lung and flank tumors, or mice with lung tumors alone (Fig. 5). Cells were also assayed for the surface levels of P-selectin ligand and CD102. CD102 and P-selectin ligand have been shown to be important for T cell trafficking to the lung during inflammatory responses, and P-selectin ligand also appears to play a role in the recruitment of tumor-specific T cells to draining lymph nodes (19, 20, 21). P-selectin ligand and CD102 expression were the same in both groups of mice (Figs. 4 C and 5, and data not shown).
The most striking differences between the mice challenged with lung and flank tumors vs lung tumors alone were the large differences in the overall cellularity of the lung draining lymph nodes (Fig. 4, A and B), and the much greater percentage of tumor-specific T cells in the BAL of the double-challenged mice (Figs. 4 C and 5). Overall, these findings support the hypothesis that the presence of the flank tumor allows more efficient priming of the DUC18 T cells than the presence of the lung tumors alone, leading to greater proliferation of naive cells and a larger pool of activated effector cells that can better reject developing lung tumors.
Diminished capacity of lung draining lymph nodes to induce DUC18 T cell proliferation
The greater overall magnitude of the immune response against flank and lung tumors vs lung tumors alone could be explained by greater proliferation of the naive DUC18 T cells stimulated by flank tumor-draining APC. To directly assess APC function in flank vs lung draining lymph nodes, draining lymph nodes from mice bearing 3-day established lung and flank tumors were harvested and used to stimulate DUC18 T cell proliferation in vitro. Flank draining lymph node cells had greater capacity to stimulate DUC18 splenocyte proliferation as compared with lung draining lymph nodes (Fig. 6). The number of APC were normalized to MHC class II+ cells by FACS analysis, therefore the difference in priming was not due to a difference in the numbers of “professional” APC. Additionally, there were equal or greater percentages of MHC class II+ cells in the lung draining lymph nodes (data not shown). Thus, there seems to be decreased overall APC function in the lung draining lymph nodes, a likely explanation of the less robust DUC18 T cell response in the mice challenged with lung tumors alone and then treated with DUC18 T cells, and the eventual outgrowth of lung tumors in a percentage of these mice.
We compared T cell-mediated rejection of the same tumor growing in the skin or the lung using T cells that have been shown to be effective in the rejection of established s.c. tumors. There is a profound difference in the efficacy of naive tumor-specific T cells in the lung vs the skin. Between 10- and 100-fold more T cells are required to protect against an i.v. tumor challenge vs an equivalent s.c. tumor challenge. A similar disparity was seen in the efficacy of naive T cells against established tumors. The difference in efficacy is even greater when one considers that the volume of tumor tissue present in a s.c. tumor is greater than that of the corresponding lung tumors, which are undetectable histologically until 2 wk post tumor transfer, whereas a s.c. tumor is 30–50 mm2 in size 1 wk after challenge. Mice challenged with both s.c. and lung tumors and treated with naive T cells were not only able to reject their s.c. tumors, they were also able to reject all of their lung tumors, despite being challenged with twice the total dose of tumor cells as the mice carrying only lung tumors. These results are direct evidence that the skin is a much better site of T cell priming than the lung, allowing not only local tumor clearance, but also overcoming the depressed pulmonary T cell response as well. When an efficient site of priming was provided, T cells were more efficiently activated and rejected all tumors present in the lungs. The mechanism responsible for this difference is unclear. We found that mice challenged with lung and s.c. tumors simultaneously had a more robust overall pulmonary immune response. That is, the number of tumor-specific T cells present in the PBLN, BAL, and lung tissue was at least equal to, and often far greater than, the number of tumor-specific T cells in the same sites in the mice bearing lung tumors alone. Although there were no profound differences in the surface phenotype of the T cells generated, the greater numbers of effector cells generated in the mice bearing lung and flank tumors vs lung tumors could be enough to ensure the rejection of all lung tumors present in those mice. The APC in the lung draining lymph nodes are far less capable of stimulating naive DUC18 T cell proliferation than are APC from the flank draining lymph nodes. This difference in the capacity to induce naive CD8+ T cell proliferation could be due to a number of factors, such as varying Ag drainage from the tumor bed, immunosuppressive cytokine environment, or differences in the populations of APC present in the different lymph nodes. This decreased proliferative capacity in the lung would lead to diminished overall pulmonary immune responses, resulting in smaller populations of effector cells, and eventual tumor outgrowth.
Although the inferior proliferation stimulated by the lung draining APC provides a likely explanation for our observations of diminished naive T cell function in the lung, it does not eliminate other possible factors that may also contribute to the disparity of T cell efficacy in the lung vs the flank. It is possible that the T cells are more effectively anergized or tolerized in the lung. The maintenance of tolerance has proven to be a great barrier to antitumor therapy, and overcoming this barrier either by vaccination or by neutralizing immunosuppressive moieties has proven to greatly facilitate antitumor immune responses (22, 23, 24, 25, 26). It is also possible that there is some difference in the homing capabilities of DUC18 cells from mice with lung tumors alone that make them less able to get to the lung draining lymph nodes to be primed and into the lung parenchyma to reject the tumors. Specific adhesion molecules and chemokine receptors have been shown to be up-regulated on T cells infiltrating the lung, especially during inflammation or infection (19, 20, 21, 27). Lastly, activation of innate immunity may differ between the lung and skin, not only because of different cell populations present, but also because of the tissue damage caused at the site of the introduction of a s.c. transplantable tumor (28).
The skin is a very efficient organ for activation of immune responses, designed to be the first line of defense against pathogens. In contrast, the lungs appear to be a profoundly immunosuppressive environment. The lung has been shown to induce Th2 responses, even to challenges that normally induce vigorous Th1 responses (29). The lung has also been reported to be a site of constitutive IL-10 production, which could dampen local T cell responses, as well as being a Th2 cytokine. The cytotoxic CD8+ T lymphocyte 1 responses seem to be more effective in model tumor systems (10, 11, 30). The general immunosuppression associated with cytokines like IL-10, already known to be relatively highly expressed in the lung, could at least in part explain our observations. Additionally, the lung appears to generate distinct populations of APCs that differ from their counterparts found throughout the rest of the body (16, 17, 31). In preliminary experiments, we tried to compensate for deficient lung priming by supplementing the naive T cell transfer with activated bone marrow derived dendritic cells. We have seen no evidence that supplemental dendritic cells alone could overcome the priming deficiency in the lung, as some mice that received dendritic cells and naive T cells still developed lung tumors (data not shown). Other groups have been able to overcome immune tolerance and initiate effective antitumoral immune responses using dendritic cell-based vaccines (32, 33, 34). Based on our analysis of surface markers, it seems that the lung environment is not miscuing the tumor-specific T cells. They are simply not being activated as efficiently, or proliferating as vigorously, leading to decreased overall numbers of effector cells, an observation supported by the in vitro data showing poor T cell proliferation stimulated by PBLN cells. It is also possible that the T cells activated in the lung are less able to traffic into the lung parenchyma due to a difference in chemokine or cytokine signals during activation caused by the lung environment. Based on our own observations and numerous other pulmonary immunology studies already published, we feel that there are four likely mechanisms responsible for the suppression of the DUC18 T cell response in the lung: 1) inappropriate polarization caused by the lung environment (29); 2) high background levels of immunosuppressive cytokines, most notably IL-10 and TGF-β (16); 3) poor APC function (31); and 4) faulty trafficking of lung activated T cells could dampen the efficacy of the DUC18 T cell response, allowing some tumors to escape rejection. Additionally, various tumors, and specifically primary lung cancers, have been shown to accumulate populations of T regulatory cells that may contribute to the down-regulation of function seen against the lung established tumors (35). These mechanisms are not mutually exclusive, and future studies will be required to dissect the respective roles of each player in this pulmonary repression.
Tumor development has been documented in nearly every tissue of the body. Heterogeneity of immune priming could be responsible for at least some of the poor immune responses against newly developing tumors. Tumors are believed to be poor at initiating immune responses; they often express low MHC class I, no class II MHC, and no costimulatory molecules (36, 37, 38). They are essentially self tissue, and express none of the molecules important in the inflammatory activation of APCs. Anything less than an optimal priming situation could allow tumor growth to outstrip the immune response. It has been observed that there are distinct differences in lymphocyte characteristics even between an organ and its own draining lymph node (39), and immunization with peptide-pulsed dendritic cells has been shown to induce T cell responses with varying kinetics and tissue distribution depending on the route of vaccination (40, 41). Persistence of pulmonary melanoma metastases has been shown even in the presence of immunity against primary tumors, indicating that rejection of a tumor present in one site does not guarantee the rejection of remote tumors (42). It is reasonable to suppose, then, that heterogeneity in tumor-specific T cell activation as a result of a tumor’s tissue of residence could result in selective outgrowth of certain tumors that would have otherwise been rejectable. When one takes the data presented in this study in the context of the concept of “tumor immunoediting” (43, 44), that is, that the immune response surveys for and eliminates newly developing immunogenic tumors, it is likely that heterogeneity in the efficiency of immune priming could lead to selective outgrowth of tumors in immunosuppressive sites, such as the lung. In our system we have seen that the immune environment of the normal tissue surrounding a developing tumor can dictate the success or failure of tumor rejection, with direct implications for the outgrowth of primary pulmonary tumors, which then may be under very little immunological pressure due to the suppressive environment in which they develop. Pulmonary metastasis, however, may still be a reasonable target for the immune system, depending on where the primary tumor developed. As we have seen in our system, the lack of priming in the lung may be overcome in the presence of a primary tumor that develops in an immunologically robust site.
We thank Silvia Kang, Laura Mandik-Nayak, Ken Matsui, and Lyse Norian for critical reading of this manuscript, Holly Hanson and Craig Byersdorfer for technical assistance and advice, Darren Kreamalmeyer for maintaining our mouse colony, and Jerri Smith and Donna Thompson for assistance in preparation of this manuscript.
This work was supported by grants from the National Institutes of Health. L.A.O. is also supported by a Howard Hughes Predoctoral training grant.
Abbreviations used in this paper: Erk, extracellular signal-regulated kinase; BAL, bronchoalveolar lavage; PBLN, parabronchial lymph node; LILN, left inguinal lymph node.