The RET/PTC3 (RP3) fusion protein is an oncogene expressed during the development of thyroid cancer and in thyroid epithelial cells of patients with Hashimoto’s thyroiditis. RP3 has two immunological properties: 1) it encodes a chimeric protein including peptides that may be targets of antitumor immune responses and 2) it is a tyrosine kinase that can activate NF-κB transcriptional programs, induce secretion of proinflammatory mediators, and stimulate innate immunity. To distinguish the antigenic properties of the RP3 oncoprotein from its signaling function, a transplantable tumor system was developed. Tumors expressing the functional, but not mutant, form of RP3 show enhanced infiltration of CD8+ lymphocytes, myeloid-derived CD11b+Gr1+ cells, and enhanced growth in immunocompetent mice. In contrast, RP3 signaling mutant-expressing tumors maintained enhanced infiltration of CD8+ lymphocytes did not enhance recruitment of CD11b+Gr1+ cells and showed a decreased tumor incidence. These results implicate a role for RP3 function in enhancing a tumor-suppressive innate inflammatory response. These experiments support a mechanism whereby oncogenes can directly recruit and activate innate and adaptive immune cells, resulting in enhanced tumor progression.
When examining cancer development from a genetic standpoint, the result of either the loss of tumor suppressors or the activation of oncogenes most often results in a predicable outcome of augmented cancer progression (1). However, this cancer progression does not occur autonomously, but rather evolves in close contact with both immune cells of the host, as well as nontransformed local stromal cells. The pathologist Rudolf Virchow (2) is credited with being the first to suggest a possible functional relationship between malignant cell growth and invading inflammatory cells as far back as 1863. Until more recently, the nature of this relationship has remained as a clinical association observed with a high frequency in malignancies. Mechanistic studies made feasible through lymphocyte transfers into immunodeficient mice and the use of conditional transgenic mouse models have suggested that cancer inflammation is a complex relationship, relying on complex supportive interactions between host immune cells and nascent cancer; with the role of cancer-associated inflammation varying depending on the tissue of origin (3, 4).
From a historical perspective, the study of immune responses to tumors has been primarily focused on the qualities of tumor antigenicity and the identification of tumor-associated Ags capable of serving as antigenic targets of antitumor immune responses (5, 6, 7). These perspectives and the identification of unique tumor Ags as targets in developing cancers are the defining elements that make up the immune selection hypothesis. This hypothesis states that a combination of innate and adaptive host immune responses are the defining factors controlling progression (8). Despite the availability of a unique tumor Ag as a antigenic target and demonstration of an antitumor T cell response, cancer often persists in mouse tumor models as well as in cancer patients (9, 10). Furthermore, early expression of known oncogenes can induce cancer without evidence of specific T cell responses. These seemingly contradictory observations create discordance with a strict role of the immune system as anticancer and have led to a new focus on the potential for inflammatory cells as a means of tumor support and a necessity for cancer progression.
Inflammatory cells are capable of providing essential cytokines, growth factors, proangiogenic mediators, and activating antiapoptotic pathways that provide a supportive environment for cancer cell survival and tumor development (11, 12). These cytokines and growth factors not only provide paracrine signals to developing neoplastic cells, but also promote recruitment and establishment of nontransformed host-derived fibroblast and endothelial cells, comprising a supportive framework for cancer cells (13, 14). In addition to this cooperative role, chronically activated inflammatory cells located within tumors can play a protective role in suppressing adaptive antitumor immune responses and thereby increasing the likelihood of cancer progression (15). For example, a heterogeneous innate cell population of myeloid-derived suppressor cells (MDSCs),4 originally characterized in mice as CD11b+Gr1+ monocyte and myeloid precursor cells (16), may play an important role in suppressing T cell responses in tumor models through a direct inhibition of CD8+ T cell function (17, 18, 19, 20). These MDSCs also influence T cell activation in human cancers as well (21, 22). These observations of a supportive role of cancer-associated inflammatory responses reinforces clinical studies identifying chronic inflammation as a predisposing factor in cancer development (11). Paradoxically, the cancers demonstrating significant inflammatory infiltrates are more often less aggressive, more differentiated, and carry a better prognosis then non-inflammation-associated cancers (23, 24). One type of cancer that exemplifies this paradox is differentiated carcinomas of the thyroid. Papillary thyroid carcinoma displays a well-characterized clinical correlation with preexisting inflammatory disorders. Both of the autoimmune conditions of Hashimoto’s thyroiditis and Graves’ disease show an increased incidence of papillary thyroid carcinoma (PTC) (25). Expression of 1 of 11 family members of the RET/PTC family of oncogenes is a specific marker for papillary thyroid cancers derived from the follicular cells of the organ (26), and the expression of either the RET/PTC1 or the RET/PTC3 (RP3) oncogenes is sufficient for thyroid cell transformation in vitro and development of thyroid carcinoma in transgenic mice (27, 28, 29) The expression of RP3 is frequently associated with occult or microcarcinomas of the thyroid compared with poorly differentiated thyroid carcinoma, suggesting that RP3 is important at early points during thyroid tumorigenesis (30, 31, 32, 33). At these early stages of neoplastic transformation, tumor growth in response to inflammatory infiltrates often sways the balance between disease progression or regression (34). Conversely, expression of RET/PTC oncogenes has been identified in patients diagnosed with thyroid autoimmunity (Hashimoto’s thyroiditis) with no evidence of thyroid cancer (35, 36, 37). In these cases, the expression of the RET/PTC oncogenes in human disease helps to mechanistically link thyroid autoimmune disease (Hashimoto’s thyroiditis) with thyroid neoplasia even though these diseases rarely appear at the same time.
Investigations into the properties of the RP3 fusion protein have revealed dual immunological features. First, constitutive activation of the RET kinase domain present in the RP3 oncogene is capable of inducing the nuclear transactivation of the NF-κB protein complex (38, 39, 40), thus initiating the expression of multiple proinflammatory mediators (41, 42). Consistent with these studies, forced expression of RP3 in either thyroid or nonthyroid cells results in the ectopic secretion of high levels of Mcp1 and Gmcsf protein (43). RP3 expression in transplanted thyroid cells as well as forced expression in mouse thyroids caused recruitment of CD11b+ macrophages, IL-6 production, and larger tumors (43, 44). The second immunological feature of the RP3 oncoprotein relates to its qualities as a novel protein expressed ectopically in thyroid follicular cells at early stages of differentiated thyroid carcinoma development (30, 31, 32, 33) and in patients diagnosed with thyroid autoimmunity (Hashimoto’s thyroiditis) (35, 36, 37). The immunogenicity of the RP3 fusion protein resides within the highly conserved tyrosine kinase domain of the protein (45). Evidence for demonstrative anti-RP3 immune responses during human disease remains to be determined; however, these studies confirmed the potential for protein antigenicity.
We hypothesized that the combined immune and neoplastic properties of RP3 play a role in tumor progression and that these properties are dependent on the immune system. In this study, we examined the multiple immunological functions of RP3 in a model of established cancer by measuring the recruitment of inflammatory cells in the tumor microenvironment and the effects on tumor growth. We demonstrate that oncogene function alone can directly account for some of the observed effects on the immune system. We contend that RP3 oncoprotein function can alter the nature of the immune response, resulting in differences in tumor growth. A better understanding of the role oncoproteins play in stimulating immune functions will help to design improved strategies for cancer therapy.
Materials and Methods
Mice and tumor cell lines
Six- to 8-wk-old female C3H/HeJ mice (The Jackson Laboratory) and C3H.C-Prkdcscid/lcrSmnHsd (C3H/SCID) mice (Harlan) were used for in vivo assays. Animal care and experiments were performed in accordance to protocols approved by the Thomas Jefferson University Animal Use Committee. Mice were maintained at the Kimmel Cancer Institute animal facility at Thomas Jefferson University. The 4102-PRO and 6132-PRO cell lines were provided by Dr. H. Schreiber (University of Chicago, Chicago, IL). The 4102-PRO and 6132-PRO C3H/HeN progressive tumor cell lines were characterized previously (46, 47). All cell lines were cultured in DMEM (Mediatech) supplemented with 10% FBS and 100 U/ml penicillin/streptomycin at 37°C in 5% CO2.
RP3 tumor cell model
Recombinant retroviral technology was used to deliver the RET/PTC3 gene to tumor cells using a triple plasmid transient transfection. The human RET/PTC3 cDNA was cloned into the MigRI MSCV-based retroviral vector containing an internal ribosome entry site (IRES) GFP cassette (provided by Warren Pear, University of Pennsylvania, Philadelphia, PA) (48). The RP3 tyrosine 588 to phenylalanine (Y588F) mutant construct was made by mutating the RP3/MigRI construct with a QuikChange XL Site-Directed Mutagenesis Kit (Stratagene). The RP3/MigRI constructs along with the pVSVg vector (49) and the Moloney murine leukemia virus gag-pol expression plasmid pCgp were transiently transfected into 293FT cells (Invitrogen) using FuGENE 6 transfection reagent (Roche). Following infection, tumor cells were sorted as a population for intermediate enhanced GFP (eGFP)-positive florescence profiles on an EPICS Elite flow cytometer and the final tumor cell transfectant populations were analyzed for purity of eGFP-positive cells by flow cytometry using an XL-MCL automated analytical flow cytometer (Beckman Coulter).
Western blot analysis
Tumor cell lysates were prepared and analyzed as detailed previously (44) Membranes were immunoblotted with goat anti-RET TK (C-19) or goat anti-pRET (Tyr-1062; Santa Cruz Biotechnology). Following incubation, membranes were washed with PBS plus 1% Tween 20 and treated with HRP-conjugated anti-goat IgG (Amersham Biosciences). Protein bands were visualized using SuperSignal West Pico (Pierce Biotechnology) chemiluminescent reagent and exposed to x-ray film.
NF-κB activity assay
To measure NF-κB transactivation activity, 4102-PRO tumor cell transfectants were transiently transfected with 1 μg of the pNFκB-TA-Luc vector (Clontech) containing the firefly luciferase reporter gene and 0.1 μg of pRL-TK vector (Promega) containing Renilla luciferase as a control for transfection efficiency. FuGENE 6 transfection reagent was used to cotransfect both plasmids and cells were incubated for 48 h at 37°C in 5% CO2. Renilla and firefly luciferase activity was analyzed using a Dual-Luciferase Reporter Assay System (Promega) and detected on a Lumat LB 9507 luminometer (PerkinElmer-Wallac). Firefly luciferase activity reported was normalized to the Renilla luciferase activity.
In vitro cell proliferation
Cell proliferation was measured using the CellTiter 96 AQueous One Solution Cell Proliferation Assay according to manufacturer’s directions (Promega). Tumor cell transfectants were plated in 96-well flat-bottom plates and grown for 4 days using one plate per day. Reagent absorbance was recorded at 490 nm minus absorbance at the reference wavelength of 655 nm. Average background absorbance of medium-only controls was subtracted from all final calculated absorbance values.
Tumor growth in vivo
To generate sufficient tumor tissue to use for each tumor challenge, 4102-PRO tumor cell transfectants were first grown in vitro, washed two times with Dulbecco’s BPS (Mediatech), and 1 × 107 cells were injected s.c. into the flanks of 6- to 8-wk-old C3H/SCID and grown for 3–4 wk. Tumors were then resected, cut into 1-mm3 tumor fragments and four fragments were transplanted into either C3H/HeJ or C3H/SCID mice using a 10-gauge trochar (Innovative Research of America). Transplants into C3H/SCID mice were done last to control for tumor viability during the procedure. Tumor size was measured every 7 days using a caliper and followed for 4 wk after challenge. Tumor volume in mm3 was calculated as (abc)/2, where a, b, and c are the orthogonal diameters. Assessment of tumor progression vs regression was made 4 wk following tumor challenge, with evidence of measurable tumor considered a progressive tumor. Cell lines from progressor tumors following tumor challenge were adapted to culture by mincing progressor tumor tissue and culturing in DMEM (Mediatech) supplemented with 10% FBS and 100 U/ml penicillin/streptomycin and grown at 37°C in 5% CO2.
Tumor tissues were collected for histological analysis 2 wk after initial tumor challenge and processed as detailed previously (44). Tissue endogenous peroxidase activity was blocked using 0.3% H2O2 followed by incubation with 10% rabbit serum (Vector Laboratories). Sections were stained with anti-mouse CD4 (L3T4) or anti-mouse CD8 (Ly-2) Abs (1/200; BD Pharmingen). Sections were reacted with a dilution of biotinylated rabbit anti-rat IgG (1/200; Vector Laboratories) followed by HRP/avidin D (Vector Laboratories). Color was detected using 3,3′-diaminobenzidine (Vector Laboratories). Sections were counterstained with Harris hematoxylin (Sigma-Aldrich). Positive staining cells were counted in 10 arbitrary microscopic fields at ×40 magnification using an ImageJ cell counter (National Institutes of Health, Bethesda, MD). Pooled infiltrate numbers represent two separate tumors stained independently (20 fields total).
Assay for cytotoxicity
Female C3H/HeJ mice were immunized i.p. with 5 × 106 irradiated (1 × 104 rad) cells followed by a booster immunization 7 days later. Fourteen days following priming, splenocytes from immunized mice were isolated and restimulated for 6 days in vitro with irradiated tumor cell transfectants used for immunization. Tumor cell transfectants were labeled with 51Cr (sodium chromate; GE Healthcare Bio-Sciences) at 100 μCi/2 × 106 cells and incubated with T cells for 4.5 h at 37°C in 5% CO2. Supernatant radioactivity was measured using a beta counter (MicroBeta; PerkinElmer-Wallac) and the percentage of specific lysis was calculated by the following formula: percent specific lysis = [(experimental release − spontaneous release) ÷ (total release − spontaneous release)] × 100. Maximum release was determined by lysis of target cells in 1% Triton X-100 detergent.
Flow cytometric analysis of tumor infiltration
Freshly excised tumor tissue (0.2 g/tumor) was digested in digestion buffer (RPMI 1640, 100 U/ml penicillin/streptomycin, 5 mg/ml collagenase type IV, and 0.15 mg/ml deoxyribonuclease type IV (Sigma-Aldrich). Cells were rinsed in FACS buffer (PBS with 1% BSA (Fisher Scientific) and 0.01% NaN3 (Sigma-Aldrich). Cells were blocked with anti-mouse CD16/32 FcBlock (BD Pharmingen), followed by staining with primary Ab. Cells were then washed and stained with avidin-allophycocyanin (BD Pharmingen). Primary Abs were fluorescein (FITC)-conjugated anti-CD11b (M1/70), R-PE- conjugated anti-Gr-1 (RB6-8C5; BD Pharmingen), and biotinylated anti-CD45 (30-F11; eBioscience). Cells were analyzed on a FACScan flow cytometer using CellQuest software (BD Biosciences).
Analysis for statistical significance for Table I was determined using Fisher’s exact test. Analysis for statistical significance between tumor volumes after 35 days (for details, see Fig. 4A) was determined using 1-way ANOVA analysis followed by Turkey’s multiple comparison test. Analysis for statistical significance for cellular infiltration (for details, see Figs. 5B and Fig. 7, B and C) was determined by a two-tail unpaired t test. All tests were performed using Prism 4 (GraphPad Software) and in accordance with the recommendations of a biostatistician.
|Tumor Challenge .||Tumor Incidence (%)a .|
|Total C3H Wild-type Mice .||Total C3H SCID Mice .|
|4102 vector (−)||20/35 (57)||11/11|
|4102 RP3b||27/37 (73)||11/11|
|4102 RP3Y588Fb||7/19 (37)||12/12|
|Tumor Challenge .||Tumor Incidence (%)a .|
|Total C3H Wild-type Mice .||Total C3H SCID Mice .|
|4102 vector (−)||20/35 (57)||11/11|
|4102 RP3b||27/37 (73)||11/11|
|4102 RP3Y588Fb||7/19 (37)||12/12|
The results are derived from five independent experiments.
The differences between the two ratios is significant by Fisher’s exact test (p < 0.02).
RP3+ tumor cells show activation of NF-κB but grow similar to controls in vitro
To measure the influence of the RP3 kinase on tumor growth in vivo, we obtained a mouse fibrosarcoma tumor cell line (4102-PRO) well characterized for its progressive growth and host immune cell dependency in immunocompetent mice (46, 47). Therefore, to separate RP3’s antigenicity (protein structure) from its function (kinase signaling), we relied on previous studies that identified the intracellular tyrosine residue at position 588 as a critical Shc docking site necessary for propagating signaling along the RAS and PI3K pathways (38, 50). RP3 or the signaling-deficient mutant (RP3Y588F) were stably expressed in the 4102-PRO fibrosarcoma tumor cell line (46) using an IRES-GFP reporter retroviral expression vector (48). FACS analysis was used to isolate single 4102-PRO cells expressing equal mean levels of GFP fluorescence, corresponding to equal levels of RP3 expression for each construct (Fig. 1). We confirmed expression of RP3 protein and elimination of tyrosine phosphorylation at position 588 by Western blot analysis using a phosphoprotein-specific RP3 Ab (Fig. 2,A). The phospho-RP3-specific band appears as a doublet in the RP3 lane (the lower band of which is the correct band, asterisk; Fig. 2,A). This double band is only present in the RP3 lane, whereas a nonspecific band (upper band of the doublet, Fig. 2 A) is seen in the adjacent lane. Stable and high levels of RP3 and GFP protein expression was confirmed in the sorted populations when cultured for up to 15 passages in vitro (data not shown).
We and others previously showed that RP3, but not RP3Y588F can induce nuclear translocation of NFκB as measured by EMSA and luciferase reporter assays (40, 43). To confirm similar signaling through this pathway in 4102-PRO tumor cell transfectants, we performed transient transfections of a NF-κB-dependent luciferase reporter plasmid. Consistent with previous results, RP3+, but not RP3Y588F+ 4102-PRO tumor cells, showed enhanced NF-κB activity (Fig. 2,B). In addition to activation of proinflammatory gene transcription, NF-κB is responsible for supporting cell growth and cell cycle progression (51, 52). However, since the 4102-PRO cell line is neoplastic, it was first necessary to confirm that expression of RP3 and subsequent activation of NF-κB did not alter cellular growth rates in vitro. Comparison of in vitro growth rates between 4102-PRO cells and each transfectant showed statistically identical rates in repeated experiments (Fig. 3).
RP3-mediated signaling enhances tumor growth in immunocompetent mice
Unlike surrogate or self-Ags used for studying tumor-induced immunity, the RP3 oncoprotein has immunological functions. One function relates to the activation of the NF-κB pathway via signaling through the tyrosine kinase domain and to a second function derived from epitopes within the protein that provide unique T cell Ags. Although it is well established that mutating tyrosine 588 in the RET kinase is capable of eliminating the transforming ability of the RP3 oncoprotein (53, 54), there have been no studies examining the immunological implications of this mutation on tumor growth in vivo. The 4102-PRO tumor line is highly sensitive to the growth-enhancing properties of an inflammatory response, resulting primarily from the provision of paracrine growth factors (55, 56). Therefore, using tumor tissue derived from the 4102-PRO tumor cell line provides an ideal model to study the influence of RP3 kinase signaling on immune cell activation and its effects on tumor growth in vivo.
Tumor take in immunocompetent mice (the total number of mice with continually growing implanted tumor fragments) was more frequent using RP3+ tumor tissue compared with control or RP3Y588F+ tumor tissue (Table I). RP3 expression in tumor tissue also conferred an enhancement of growth rates and total volumes compared with controls (Fig. 4,A). The increased growth rates and volumes of RP3+ progressor tumors were only measurable in normal mice, whereas all tumor groups showed equivalent growth rates and sizes in immunodeficient SCID recipients. Cells from progressor tumors were isolated from selected normal and SCID mice to confirm expression of RP3 and RP3Y588F using the coexpressed GFP marker (Fig. 4 B) and through Western blot analysis for RP3 (data not shown). Although the RP3- and RP3Y588F-expressing tumors showed a minor reduction in GFP fluorescence compared with control tumors, there was no difference between RP3 and RP3Y588F tumors. Thus, these data suggest a role for RP3 kinase function in the enhancement of tumor progression in normal, but not immunodeficient mice. These data also suggest that RP3 may be a tumor Ag since we observed a significant difference in tumor survival between RP3 and RP3Y588F groups and only a minor difference in tumor survival between RP3 and control tumors in immunocompetent mice. Furthermore, the reduction of gene expression levels for both RP3 and RP3Y588F was evident in all progressor tumors analyzed.
RP3 expression correlates with increased CD8+ T cell recruitment at an early stage of tumor progression
Previous studies demonstrated the importance of CD8+ T cell immunity in restraining UV-induced mouse fibrosarcoma growth in vivo (47, 56). The differences in number and size of RP3-expressing tumors in normal vs immunodeficient mice suggested that T cell recognition may be an important component in the mechanism driving oncoprotein-mediated tumor progression. Observation of tumor growth and rejection data (Fig. 4,A and our unpublished data) demonstrated that 2 wk after tumor challenge was a critical time to observe either tumor progression or regression and was coincident with maximal T and B cell infiltration into tumor tissue, suggesting that these cells may be involved in the growth enhancement of RP3+ tumors. To test this possibility, we examined tumor tissues by immunohistochemical analysis 2 wk after tumor challenge. In addition to infiltration by Gr-1+ granulocytes (data not shown), all tumor types showed infiltration of both CD4+ and CD8+ cells throughout the entire tumor tissue (Fig. 5,A). Analysis of CD4+ infiltrates revealed similar numbers of CD4+ lymphocytes in control, RP3-, and RP3Y588F-expressing tumors (Fig. 5,B). In contrast, tumors expressing RP3 and RP3Y588F showed significantly higher numbers of infiltrated CD8+ lymphocytes compared with controls (Fig. 5 C).
RP3 expression in tumors does not augment tumor antigenicity
We had previously demonstrated that RP3 protein is recognized by T cells following either immunization with RP3 protein or in the context of transplanted mouse RP3+ thyroid tumors (45). Increased infiltration of CD8+ cells into RP3+ and RP3Y588F+ tumors compared with controls suggested that oncoprotein antigenicity may be playing a direct role in promoting CD8+ T cell infiltration due to its properties as a unique protein Ag. To investigate this, we used a mouse fibrosarcoma cell line (6132-PRO) expressing RP3 or the RP3Y588F signaling-deficient construct generated in the same manner as the 4102-PRO transfectants (data not shown). The 6132-PRO line was used to verify the properties of RP3-induced inflammation in a distinct tumor system and to separate endogenous tumor-specific immune responses from those directed against RP3. This is important since the 4102-PRO and 6132-PRO lines express their own unique CTL-defined Ags and do not share a dominant T cell Ag (46, 56). Therefore, cross-recognition between the two RP3 transfectant cell lines would be the result of a RP3 Ag-specific response. Spleen cells from mice immunized and boosted with either 4102 vector (−), 4102 RP3, or 4102 RP3Y588F tumor cells were used as effector cells to detect RP3-specific recognition in 6132 RP3 transfectants. Fig. 6 shows that despite a strong anti-4102 response, neither 6132 RP3 nor 6132 RP3Y588F cells were killed by RP3- or RP3Y588F-immune splenocytes, demonstrating that RP3 was not recognized as a tumor cell Ag in the context of 4102-PRO. NIH3T3 cells were included as a non-MHC-matched control. Additionally, the degree of T cell recognition of 4102-PRO cells was equivalent to cells expressing RP3, indicating that a maximal immune response to the 4102-PRO tumor Ag occurred without significant RP3 Ag recognition.
RP3 signaling induces recruitment of CD11b+Gr1+ cells into tumors
Previous studies from our laboratory suggested that the signaling activity of the RP3 oncoprotein mediated recruitment of CD11b+ macrophages and the secretion of myeloid chemotactic factors from RP3-transfected cells (43, 44). Studies of the 4102-PRO tumor cell line demonstrated that progression of these fibrosarcomas was highly dependent on Gr1+ cells, with depletion of these cells in vivo resulting in reduced tumor growth in nude mice and an increased incidence of tumor rejection in wild-type-immunocompetent mice (55, 56). Of particular interest for our study was a population of immature monocytes displaying both the CD11b and Gr1 cell surface markers associated with cells that suppress CD8+ lymphocyte responses in other tumor models (18, 20). The reported effects of these MDSCs may help to explain why RP3+ tumors show increased growth even in the presence of a significant CD8+ T cell infiltrate (Table I and Fig. 4,A). To further quantify this cell population, single- cell suspensions were prepared from progressor tumor tissues and analyzed by flow cytometry 2 wk after tumor challenge. Flow cytometric analysis of CD45+ tumor-infiltrating lymphocytes showed an increased percentage of CD11b+Gr1+ cells in 4102-PRO C3H tumors expressing the active RP3 oncoprotein compared with control or RP3Y588F tumors (Fig. 7, A and C). The RP3-mediated increase in this cell population was also evident in 4102 RP3+ tumors derived from C3H/SCID mice at the 2-wk time point (Fig. 7 B). These data suggest that the RP3-induced tumor infiltration of CD11b+Gr1+ cells is dependent on RP3 function and, unlike the observed tumor incidence and growth, myeloid cell recruitment into these tumors does not require T cells, B cells, or their products.
In this study, we provide evidence that the signaling function of the RP3 oncoprotein is responsible for enhancing tumor progression through its influence on lymphocytes in vivo. We suggest that this enhancement is mediated by the function of the RP3 kinase to provoke recruitment of innate suppressive inflammatory cells.
The participation of immune cells in assisting cancer cell growth has been documented using various tumor models (2). Although immune recognition of tumor Ags can limit cancer growth through immune surveillance, tumor progression represents the failure of surveillance and/or a positive growth influence provided by soluble factors contributed by hematopoetic cells of the tumor microenvironment (12). We have explored the cause of tumor-induced inflammation through mechanistic studies of the RP3 oncogene, whose expression occurs in cases of thyroid cancer and in benign thyroid autoimmune diseases. Previous studies involving the RP3 thyroid oncoprotein showed that activating the RP3 kinase causes secretion of myeloid chemokines from thyroid epithelium that results in the attraction of CD11b+ macrophages (43, 44). In addition, RP3 can serve as a tumor-specific Ag for T cells (45). These immunogenic and proinflammatory qualities of RP3 suggest that it could initiate cancer-associated innate inflammation as well as provoking adaptive T cell responses. The results from this study illustrate the importance of oncogene-mediated inflammation in support of tumor development and progression in a mouse model of tumor growth.
We have used the growth of mouse fibrosarcomas in vivo to study the mechanism of tumor progression by focusing on two immunological properties inherent in the RP3 oncoprotein: 1) its capacity to provoke proinflammatory signaling and 2) its inherent antigenicity as a tumor-specific protein. Previous studies using the 4102-PRO mouse fibrosarcoma tumor model demonstrated growth sensitivity to changes in the inflammatory microenvironment due to paracrine cytokines and chemokines contributed by activated lymphocytes (55). Our previous studies correlated RP3- mediated transformation and coexistent inflammation with enhanced development of thyroid hyperplasia in mice transgenic for thyroid-specific expression of RP3 (44). In these mice however, the influence of RP3-mediated transformation was not separated from any positive growth influences contributed by local inflammatory infiltrates. In this study, the use of 4102-PRO progressor tumor model provided a means to analyze the influence of RP3 inflammatory functions on tumor progression independent of its function in cellular transformation. We showed equivalent in vitro growth rates of the 4102 tumor cells expressing RP3 or vector control constructs, indicating that RP3 kinase signaling does not interfere with the transformation and cell growth phenotype of the parental 4102 tumor cells. When transplanted into C3H/SCID-immunodeficient mice, 4102 tumors expressing RP3 progressed with a similar frequency and rate as tumors expressing mutant RP3Y588F or control tumors, corroborating the in vitro growth rate data and confirming that RP3 function did not enhance intrinsic cell growth rate.
We observed enhancement of both tumor incidence and growth rate only when tumors were grown in mice with an intact adaptive immune system, suggesting that the contribution of T cells is necessary. RP3 kinase function provided a progressive advantage for tumors, despite increased infiltration of CD8+ lymphocytes, suggesting that tumor-infiltrating lymphocytes may be playing a supportive role. We find that RP3 expression also correlates with an enhanced innate cell infiltrate. This observed infiltration is likely the result of the mediators synthesized from RP3-expressing cells consequent to the activation of NF-κB (43, 57). Even though enhanced tumor growth was only observed in wild-type mice with an intact adaptive immune system, enhanced infiltration of immature monocytes/macrophages into tumors expressing an active RP3 kinase in both wild-type and SCID mice implies that the function of the RP3 kinase is involved in the recruitment of innate immune cells as well. Infiltration of CD11b+Gr1+ cells occurred at an early time point following tumor challenge, suggesting that the interplay between innate and adaptive leukocytes activated by RP3 signaling may combine to provide tumor enhancement through either the suppression of antitumor T cell responses or by the production of cytokines and growth factors that have tumor growth-promoting effects.
Tumors expressing RP3 show a minor survival advantage compared with negative control tumors following tumor challenge, which was in contrast to the significant decrease of tumor survival when compared with tumors expressing the RP3Y588F mutant. This suggests that factors, in addition to RP3 signaling activity, may be influencing tumor growth in immunocompetent mice and that the role of RP3 as a tumor-specific Ag may be playing only a minor role in this model. Consistent with this notion, previous studies suggested that RP3 could be recognized as a T cell Ag in the context of a tumor rejection response (45). Demonstration of selection for reduced RP3 expression in both RP3 and RP3Y588F progressor tumor cells implies an in vivo selection against tumor cells expressing RP3, regardless of oncoprotein function. Thus, recognition of RP3 as a tumor Ag may account for the increase in CD8+ infiltration of both RP3 and RP3Y588F progressor tumors. A more detailed analysis of the contribution of RP3 to the antigenicity of the 4102-PRO tumor cells failed to show that immunizing with 4102 RP3-expressing tumors could induce a dominant anti-RP3-specific response. However, the 4102-PRO line was shown to express a dominant tumor-specific Ag detected by CD8+ cells in vivo (46, 47). Repeated immunization and in vitro restimulation using the 4102 RP3 tumor cells may potentially bias the response predominately toward this immunodominant tumor-specific response and overshadow any RP3-specific responses. However, the immune response to this tumor Ag in vivo may be less robust then repeated antigenic restimulations in vitro, thereby allowing some responses against a minor Ag such as RP3. Additionally, RP3 antigenicity was measurable when RP3 signaling function was eliminated, as exemplified by the detectable increase in rejection of RP3Y588F tumors compared with the 4102 vector control.
The 4102-PRO tumors have a dependence on Gr1+ granulocytes for improved in vivo tumor growth rate in athymic mice (55) and for preventing T cell-mediated tumor rejection in immunocompetent mice (56). This may be occurring in our studies more efficiently due to the signaling function of RP3, known to be a potent inducer of myeloid stimulants such as GM-CSF and MCP-1. Accumulation of CD11b+Gr1+ myeloid cells has been observed in human cancers of increasing size and their function has been linked to profound immunosuppression (17). Interestingly, enhanced infiltration of CD11b+Gr1+ cells over controls was observed in RP3 tumors derived from both wild-type as well as SCID mice, suggesting that the RP3 kinase can stimulate the recruitment of innate cells independent of adaptive immune cells. However, tumor growth enhancement is only observed when adaptive lymphocytes are present. Although the 4102-PRO tumors are capable of recruiting lymphocytes and cytokine secretion independent of RP3 expression, we propose that enhanced transactivation of NF-κB by the RP3 kinase increases the quantity of secreted cytokines (or changes the balance of cytokines expressed), thereby increasing the number of myeloid-derived CD11b+Gr1+ cells recruited into tumor tissue.
Although these studies only measured the frequency of CD11b+Gr1+ myeloid cells, other studies have shown that these cells possess strong suppressive activity on CD8+ tumor-specific responses (17). Therefore, enhancement of MDSC (CD11b+Gr1+) frequency in RP3-expressing tumors at an early stage of tumor progression may account for both the enhanced tumor incidence and larger tumor size observed despite increased numbers of CD8+ cells in RP3-expressing tumors. Others have reported that elimination of granulocytes in immunodeficient mice suppressed growth of 4102-PRO tumors (55), suggesting that granulocytes are directly responsible for supporting tumor growth through the release of growth factors. However, we did not observe a growth enhancement in SCID tumors that accumulated increased CD11b+Gr1+ cells. We only observed growth differences in wild-type mice when CD11b+Gr1+ cells were present along with adaptive immune cells, suggesting that other factors are playing a role in the tumor growth enhancement observed in RP3+ tumors. Recently, it was demonstrated that IFN-γ, derived from intratumoral-activated T cells, was required to activate the suppressive function of CD11b+ cells (58). These data reinforce the notion that the enhancement of T cell activation at the tumor site, when combined with localization of myeloid-derived CD11b+Gr1+ cells, may promote a favorable environment supporting cancer progression, rather than regression, in the context of a potential antitumor response. Additionally, we have shown that RP3 signaling from nonhematopoietic cells is capable of activating naive lymphocytes to secrete IFN-γ in vitro (J. S. Pufnock, manuscript in preparation), providing further evidence that the signaling function of RP3 may influence the activation state of both the adaptive and innate immune system in cells harboring this mutation. Although we suggest that immature CD11b+Gr1+ cells are acting in a suppressive manner on CD8+ cells, it requires further evaluation to determine whether this interaction provides direct help to growing tumors through the secretion of growth factors rather than, or in addition to, suppression of directed anti-tumor CD8+ cytolytic responses.
Indeed, other mechanisms may be operating to result in a positive influence of RP3 function on tumor enhancement in the context of an adaptive immune response. For example, tumor cells expressing functional RP3 may show an enhanced resistance to destruction by antitumor CD8+ lymphocytes resulting from the antiapoptotic activities associated with NF-κB signaling (59). Although our analysis showed no significant quantitative differences in CD4+ infiltration of tumors, it is possible that differential cytokine patterns released from RP3 tumors may preferentially promote the differentiation of CD4+ cells into suppressive regulatory T cells known to influence tumor growth in a similar fibrosarcoma model (9).
In this study, we have demonstrated that despite a minor antigenic role, the signaling function of the RP3 oncoprotein plays a significant role in promoting tumor progression in the context of a well-established tumor model. This suggests that RP3 enzymatic functions, independent of those required for cellular transformation, are critical to promote tumor progression from an early stage. These immunological functions exert a powerful influence on both innate and adaptive immune cells present at the site of the growing tumor. Furthermore, the function of the RP3 oncoprotein to enhance NF-κB activation may be the key pathway governing its immunomodulatory functions. Studying both the cell autonomous (oncogenic) and paracrine (inflammatory) function of oncogenes is necessary to understand the process of cancer progression and to develop and improve oncological therapeutics.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Pennsylvania Tobacco Health Research Grant 4100026302 and by grants from the National Institutes of Health (CA76259 and AI063065 to J.L.R. and T32-CA09683 to J.L.P.).
Abbreviations used in this paper: MDSC, myeloid-derived suppressor cell; PTC, papillary thyroid carcinoma; RP3, RET/PTC3; IRES, internal ribosome entry site; eGFP, enhanced GFP; MTS, (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethylphenyl)-2-(4-sulfophenyl)-2H-tetrazolium).