|Home | About | Journals | Submit | Contact Us | Français|
Recently, we identified a lung adenocarcinoma signature that segregated tumors into three clades distinguished by histological invasiveness. Among the genes differentially expressed was the type II transforming growth factor beta receptor (TGFβRII), which was lower in adenocarcinoma mixed subtype and solid invasive subtype tumors compared with bronchioloalveolar carcinoma. We used a tumor cell invasion system to identify the chemokine CCL5 (RANTES) as a potential downstream mediator of TGF-ß signaling important for lung adenocarcinoma invasion. We specifically hypothesized that RANTES is required for lung cancer invasion and progression in TGFβRII repressed cells. We examined invasion in TGFβRII deficient cells treated with two inhibitors of RANTES activity, Met RANTES and a CCR5 receptor blocking antibody. Both treatments blocked invasion induced by TGFβRII knockdown. In addition, we examined the clinical relevance of the RANTES-CCR5 pathway by establishing an association of RANTES and CCR5 immunostaining with invasion and outcome in human lung adenocarcinoma specimens. Moderate or high expression of both RANTES and CCR5 was associated with an increased risk for death, P = .014 and .002, respectively. In conclusion our studies indicate RANTES signaling is required for invasion in TGFβRII deficient cells and suggest a role for CCR5 inhibition in lung adenocarcinoma prevention and treatment.
Lung adenocarcinoma, the most frequent histological type of non-small cell lung carcinoma is heterogeneous. Histologically, subclassification of adenocarcinoma is based upon World Health Organization (WHO) criteria that is determined predominantly by cell morphology and growth pattern and comprised of a spectrum that includes non-invasive bronchioloalveolar carcinoma (BAC), pure invasive adenocarcinoma (IAC) and adenocarcinoma with mixed subtypes (AC-Mixed) (Brambilla et al., 2001). Genomically, we and others have used microarray based approaches to subclassify lung adenocarcinoma into clusters associated with clinical outcome, lung terminal differentiation state, or with differentiation patterns similar to those of other non-small cell carcinoma histological cell types (reviewed in (Borczuk & Powell, 2007). Recently, we identified a lung adenocarcinoma genomic signature that segregated tumors into three clades distinguished by histological invasiveness and thus paralleled the WHO subclassifications of solid adenocarcinoma; mixed-subtype adenocarcinoma; and BAC or microinvasive BAC (defined as BAC with an invasive component less than 5mm) (Borczuk et al., 2005). Among the genes differentially expressed in the progression from BAC to invasive tumors was the type II transforming growth factor beta receptor (TGFβRII), which was lower in AC-Mixed and solid invasive tumors compared with BAC. This finding, which suggested that TGFβRII repression was required for lung adenocarcinoma invasion, was confirmed using qRT-PCR and immunohistochemistry, and by in vitro studies indicating that TGFβRII expression was inversely correlated with lung cancer cell invasion.
The importance of TGF-β signaling in mediating tumor invasion, which is the first step of the metastatic process, is recognized. However, downstream signaling mechanisms through Smad mediated or non-canonical pathways remain unclear and models support both the prometastatic and anti-metastatic properties of TGF-β (Gupta & Massague, 2006). Targeted deletion of TGFβRII in established cancer models of the breast and colon consistently shows that repression of TGFβRII mediated by Smad independent pathways is associated with tumor progression and metastasis (Biswas et al., 2004; Forrester et al., 2005; Ijichi et al., 2006). The phenotypes of the Tgfbr2 deficient cancer models clearly demonstrate the importance of TGF-β pathway signaling in tumor invasion, yet the downstream signaling mechanisms are undefined.
We used a tumor cell invasion system to identify and characterize downstream mediators in TGFβRII repressed cells important for lung adenocarcinoma invasion. Candidate targets were identified using DNA microarray gene expression signatures of adenocarcinoma tumor specimens and of TGFβRII knockdown cells in vitro (Borczuk et al., 2005). Among potential mediators was the chemokine CCL5 (RANTES), which was upregulated in invasive tumors and in TGFβRII knockdown cells. RANTES (Regulated on Activation, Normal T-cell Expressed, and presumably Secreted) is involved in immunoregulatory and inflammatory processes and is transcribed and secreted not only by T cells, other inflammatory cells and stromal cells, but also by tumor cells and normal bronchial epithelium. RANTES is a ligand for chemokine receptors CCR1, CCR3, CCR4, and CCR5, which are expressed on epithelial cells, macrophages, lymphocytes, dendritic cells and stromal cells (van Deventer et al., 2005). Representing a potential therapeutic target important for tumor cell motility and chemotaxis, RANTES was assigned priority for validation and characterization as a mediator of lung adenocarcinoma invasion. We specifically hypothesize that invasion in TGFBRII repressed human lung adenocarcinoma tumors requires RANTES. To test this hypothesis we examined invasion in TGFβRII deficient cells treated with two inhibitors of RANTES activity, Met RANTES and a CCR5 blocking antibody. We show that these inhibitors block invasion induced by TGFβRII knockdown. In addition, we examined the clinical relevance of the RANTES-CCR5 pathway by establishing an association of RANTES and CCR5 expression with invasion and with clinical outcomes in a large panel of human lung adenocarcinoma specimens.
We have previously used RNAi to demonstrate that reduced expression of TGFβRII is associated with increased invasion of H23 and SKLU lung cancer cells(Borczuk et al., 2005). In the current work, we introduce a third lung adenocarcinoma cell line and a second TGFβRII siRNA construct to better control for potential off-target effects of RNA interference. As indicated in Supplementary Figure 1, both siRNA constructs effectively repressed TGFβRII expression, as measured by immunoblot and quantitative real time PCR.
Microarray data from these previous experiments indicated that TGFβRII expression was inversely correlated with the expression of the chemokine CCL5, thus identifying RANTES as a potential downstream effector of invasiveness in TGFβRII knockdown cells. This is consistent with recent reports suggesting a role for RANTES in mediating invasion of breast carcinoma cells(Azenshtein et al., 2002). Microarray data indicating that expression of CCL5 was increased in TGFβRII knockdown H23 cells were confirmed by quantitative real time PCR in H23, SKLU and H522 cells (Figure 1a). Next, we used ELISA assays to confirm that RANTES secretion increased in response to TGFβRII repression. Small amounts of RANTES were detectable in the media of control cells. After TGFβRII knockdown, RANTES secretion increased 2.8, 9.2 and 2.0 fold over controls in the H23 (P=2×10−2), SKLU (P=1×10−5), and H522 (P=1×10−3) cells, respectively (Figure 1b).
To determine if RANTES secretion is functionally important in mediating lung cancer invasion, we measured invasion in wild-type cells treated with exogenous RANTES, using physiologically relevant concentrations detected in the supernatant of siRNA treated cells. We detected increases in invasion of 5.3, 1.7 and 2.5 fold in H23, SKLU and H522 cells, respectively (Figure 1c). Based upon these results taken together, we hypothesized that RANTES is required for invasion in TGFβRII deficient lung cells. The cause of the differential in baseline values of invasion and response to RANTES among the cell lines is unclear but was not explained by detectable differences in expression of CCR5, the primary chemokine receptor for RANTES (data not shown).
To determine if RANTES is required for mediating invasion in lung cancer cells with altered TGF-ß signaling, we tested two inhibitors of RANTES, Met-RANTES and a CCR5 blocking antibody. Met-RANTES is a methionylated molecule that is a direct binding antagonist for both CCR1 and CCR5 (Proudfoot et al., 1996). Using Met-RANTES, the invasion induced by knockdown of TGFβRII was completely abrogated in all three cell lines tested (Figure 2), suggesting that invasion by TGFβRII deficient adenocarcinoma cells could be prevented or reversed by antagonists of RANTES. To determine the specificity for signaling through the CCR5 receptor and because the therapeutic potential of Met-RANTES is limited by reports of agonist activity in vivo (Culley et al., 2006), we also tested a CCR5 specific monoclonal blocking antibody (α-CCR5). Using α-CCR5 we detected decreases in invasion of TGFβRII knockdown cells of similar magnitude to that observed with Met-RANTES, suggesting that more specific targeting of CCR5 also has potential utility in reversing or preventing adenocarcinoma invasion. Reduction of invasiveness using either Met-RANTES or α-CCR5 in TGFβRII deficient cells was reproducibly detected using both siRNA-TGFβRII constructs in all three cell lines, with P < .01 in drug treated cells vs. control in each instance (Table 1). Taken together these results suggest that RANTES signaling is required for invasion in TGFβRII deficient lung adenocarcinoma cells.
To determine if our findings from in vitro studies are applicable to human tumors, we examined the intensity and distribution of immunostaining for RANTES and CCR5 in a large panel of human lung adenocarcinoma specimens (Figure 3a). We hypothesized that immunostaining for RANTES and CCR5 would be increased in invasive adenocarcinoma tumors (Figure 3b). Moderate or strong staining for RANTES was detected in 58% of the invasive tumors (out of a total of 148) but in only 27% of the BAC tumors (out of a total of 33). Conversely, while absent or weak staining was detected in 42% of IAC tumors, 72% of BAC tumors had absent or weak staining. The probability of a difference this great or greater occurring by chance, according to the test for comparison of two proportions, is .003. Thus the immunostaining results confirm the gene expression studies. Compared with RANTES, immunostaining for CCR5 demonstrated a larger difference in BAC vs. invasive tumors with moderate or strong staining in 15% of BAC tumors and 70% of IAC, P = 1.60 × 10−8.
Since RANTES is expressed and secreted by several cell types and may act in both an autocrine or paracrine fashion, we also measured its expression in fibroblasts, macrophages and stromal collagen. Although we did not detect differences between IAC and BAC for RANTES and CCR5 staining in macrophages or stromal collagen, we did detect an increase in staining intensity of fibroblasts in IAC for both RANTES and CCR5 when compared with BAC tumor fibroblasts. This finding supports the role for tumor-fibroblast interactions mediated by TGF-β and RANTES in the progression of lung adenocarcinoma.
To more quantitatively assess invasion, we examined the association between the maximal measured length of invasion and immunostaining intensity for RANTES and CCR5 in lung adenocarcinoma specimens. Length of invasion was positively correlated with tumor cell immunostaining for both RANTES (Spearman coefficient r= .25; P<.001) and CCR5 (Spearman coefficient r= .465; P<.0001). Interestingly, in fibroblasts, immunostaining for only CCR5 but not RANTES was correlated with invasion length. This suggests that the ligand and receptor, as expressed by fibroblasts and tumor cells, have complementary roles in mediating invasion. The importance of CCR5 in fibroblast mediated tumor invasion is supported by a recent study indicating that adoptive transfer of CCR5 expressing stromal fibroblasts into CCR5 knockout mice increased melanoma metastases(van Deventer et al., 2005).
Extent of basement membrane invasion in lung adenocarcinoma is associated with tumor recurrence and survival (Sakurai et al., 2004). We examined the correlation between the biomarkers of invasion (RANTES and CCR5) and survival in a cohort of 162 patients with lung adenocarcinoma resected between 1997 and 2000. Using Kaplan-Meier Cox proportional hazard analysis, moderate or high staining for both RANTES and CCR5 in tumor cells was associated with increased risk for all cause mortality (P=0.014 for RANTES and P=.002 for CCR5, Figure 3C). These results support the clinical relevance of /RANTES as a biomarker of lung adenocarcinoma invasion and progression and support the potential clinical utility of CCR5 antagonists for lung cancer prevention and treatment.
Lung cancer metastasis represents the final step of a complex sequence comprised of invasion (loss of cell-cell adhesion, increased cell motility, and basement membrane degradation); vascular intravasation and extravasation, establishment of a metastatic niche and angiogenesis. Our research has focused on characterizing the molecular mechanisms important for invasion, the initial step of metastasis. In lung adenocarcinoma, loss of TGFβRII expression with concomitant altered TGF-β signaling is an important initiating event of invasion. Yet, the downstream signaling events are complex and not fully defined. To determine these events in lung adenocarcinoma tumor cells, we used genomics and an in vitro based invasion assay to identify and characterize genes upregulated in invasive tumor specimens and in cells with reduced expression of TGFβRII. Among these genes was CCL5 which encodes the CC chemokine RANTES.
Microarray data indicating CCL5 expression was increased in TGFβRII deficient cells were confirmed by qRT-PCR and by ELISA. The functional importance of this protein in invasion was examined using exogenous RANTES, which increased invasion 2–5 fold. We established the requirement for RANTES in mediating invasion in lung adenocarcinoma cells by abrogating invasion in TGFβRII deficient cells by use of selective and non-selective RANTES antagonists. Taken together, our results suggest that lung adenocarcinoma invasion associated with TGFβRII repression requires (RANTES).
Complex networks of chemokines and chemokine receptors expressed on tumor cells, macrophages, fibroblasts and endothelium interact in tumor progression and tumor metastasis (Balkwill, 2004). In the lung, CCR5 (the primary receptor for RANTES) is constitutively expressed by lung epithelial cells, granulocytes, dendritic cells, macrophages, lymphocytes and stromal cells. It plays an important role in the tissue inflammation, protease production and tissue remodeling that is characteristic of diseases such as emphysema, rheumatoid arthritis, sarcoidosis and organ transplant rejection (Ma et al., 2005; Nissinen et al., 2003). Based upon its role in inflammatory lung diseases, it is not unexpected that the (RANTES)/CCR5 pathway is important for mediating physiological processes required for tumor invasion and progression. Indeed, increased RANTES signaling is associated with advanced tumor stage in carcinoma of the breast, prostate, ovary, and squamous cell carcinoma of the lung (Luboshits et al., 1999; Remmelink et al., 2005; Vaday et al., 2006). Our studies in an epithelial cell autonomous system indicate that RANTES expression by tumor cells acts directly to promote invasion in part in an autocrine fashion, as has been suggested by others (Azenshtein et al., 2002).
Interestingly, other reports suggest a role for RANTES as an inhibitor of tumor progression. Co-transduction of RANTES and GM-CSF into mouse WEHI3B leukemia cells increased recruitment of CD4 lymphocytes and inhibited subcutaneous tumor growth (Nakazaki et al., 2006). Moran and colleagues identified increased RANTES expression as a marker of increased tumor lymphocytic response that was associated with longer survival than in patients with tumors expressing lower RANTES levels and with an absent lymphocytic response (Moran et al., 2002). Because RANTES is expressed by both tumor cells and by immune cells, it is unclear whether differential RANTES mRNA expression was intrinsic to the tumor or was predominantly determined by the increased infiltration of inflammatory cells in good prognosis vs. poor prognosis tumors. Nevertheless, these reports indicate the importance of tumor-immune cell interactions in cancer progression and metastasis.
The advantage of an epithelial cell autonomous system is the ability to isolate tumor signal transduction events so as to characterize the importance and sequence of pathways altered by TGF- β and RANTES in adenocarcinoma invasion. To our knowledge, our report is the first to link alterations in TGF-β signaling with the RANTES/CCR5 pathway and lung carcinoma progression. TGF-β has been found both to induce and to suppress RANTES production through the TAK1 (Transforming growth factor Activated Kinase 1) pathway in microglial cells (Jang et al., 2002). In the context of our prior results showing p38 activation in TGFβRII knockdown cells, and those of others demonstrating that activated p38 enhances binding of the CREB and ATF2 transcription factors to the RANTES promoter (Gustin et al., 2004), potential pathways linking TGFβRII repression with TAK1, p38, CREB, and increased RANTES expression are suggested.
These studies provide insights into the molecular pathways that mediate progression of adenocarcinoma from noninvasive BAC to invasive adenocarcinoma and thus are of high clinical significance. Immunohistochemical analysis of tumors suggest that these pathways are operative in human lung adenocarcinoma and indicate that increased expression of RANTES and CCR5 in both tumor cells and in tumor associated fibroblasts is associated with increased tumor invasion, which previously has been associated with lung adenocarcinoma recurrence. Importantly, our correlative studies in resected lung adenocarcinomas indicate that increased expression of RANTES and CCR5 protein is associated with increased risk for death and suggest that tumor immunostaining for RANTES and CCR5 are potential prognostic biomarkers that may distinguish individuals with increased risk for death following resection of lung adenocarcinoma. Our in vitro studies indicate that invasion mediated by the RANTES/CCR5 pathway is reversible in part and support the importance of studies to determine if CCR5 inhibition may prevent progression or prevent metastasis of lung adenocarcinoma. This suggests the exciting prospect of Phase II trials using novel oral small molecule inhibitors of CCR5, such as Maraviroc which recently received United States Food and Drug Administration approval for AIDS treatment as a HIV re-entry inhibitor (Fatkenheuer et al., 2005).
Deciphering the molecular processes underlying the acquisition of invasiveness promises to have increasing importance as we anticipate a rise in the detection of early stage lung adenocarcinoma as a result of lung cancer screening with low-dose CT scans. The heterogeneity in clinical outcomes for patients with screen detected lung adenocarcinoma is likely attributable in part to histological heterogeneity and invasiveness. As screening for lung cancer becomes more widespread, we are likely to see a shift in the epidemiology of lung cancer away from more advanced disease and towards early, and in some cases, not yet invasive disease. An improved understanding of the biological properties of these tumors and discovery of novel targeted therapeutics promises to significantly enhance our treatment approach to lung cancer.
RNAi was performed using two predesigned annealed anti-TGFβRII siRNAs (Ambion, Austin, TX) denoted as siRNA TGFβRII-A (sense 5′-GGUCGCUUUGCUGAGGUCUTT-3′), and siRNA TGFβRII-B (sense 5′-GGAAGUCUGUGUGGCUGUATT-3′) and a control mixture of four non-targeting siRNAs (Dharmacon, Lafeyette, CO). 1×105 cells were transfected for 48 hours with 100nM annealed siRNA using Lipofectamine 2000 system (Invitrogen, Basel, Switzerland).
Cells were incubated with Met-RANTES at 1ng/ml (R&D Systems, Minneapolis, MN) or with anti-CCR5 monoclonal antibody (BD Pharmingen, San Diego, CA) at 10ug/ml or control anti-mouse IgG 10ug/ml (Vector Laboratories, Burlingame, CA) for 24 hours prior to seeding in transwell chamber. Experimental drug concentrations were determined by 50% inhibition of calcium mobilization as described previously (Schwabe et al., 2003).
RANTES in undiluted supernatant was analyzed using the Endogen Human RANTES ELISA kit (Pierce-Endogen, Rockford, IL) following the manufacturer’s protocol and read with LabSystems Multiscan MCC/340 (Fisher-Thermo Electron Corporation, MA). All assays were performed in triplicate.
Immunostaining for RANTES (15μg/ml), and CCR5 (2μg/ml) was performed on 196 lung adenocarcinomas from tissue sections and tissue microarray (149 IAC and 33 BAC) (Supplementary Methods).
This work was supported in part by NIH (1RO1CA120174), American Cancer Society (RSG0524801CNE), Joan’s Legacy Foundation, and Flight Attendants Medical Research Institute.
The authors thank Dr. Samuel Silverstein and Dr. Yens Huseman, Department of Physiology, Columbia University College of Physicians & Surgeons, New York, NY, for assistance and advice.