Promoting effect of neutrophils on lung tumorigenesis is mediated by CXCR2 and neutrophil elastase

Background Tumor cells produce various cytokines and chemokines that attract leukocytes. Leukocytes can amplify parenchymal innate immune responses, and have been shown to contribute to tumor promotion. Neutrophils are among the first cells to arrive at sites of inflammation, and the increased number of tumor-associated neutrophils is linked to poorer outcome in patients with lung cancer. Results We have previously shown that COPD-like airway inflammation promotes lung cancer in a K-ras mutant mouse model of lung cancer (CC-LR). This was associated with severe lung neutrophilic influx due to the increased level of neutrophil chemoattractant, KC. To further study the role of neutrophils in lung tumorigenesis, we depleted neutrophils in CC-LR mice using an anti-neutrophil antibody. This resulted in a significant reduction in lung tumor number. We further selectively inhibited the main receptor for neutrophil chemo-attractant KC, CXCR2. Similarly, this resulted in suppression of neutrophil recruitment into the lung of CC-LR mice followed by significant tumor reduction. Neutrophil elastase (NE) is a potent elastolytic enzyme produced by neutrophils at the site of inflammation. We crossed the CC-LR mice with NE knock-out mice, and found that lack of NE significantly inhibits lung cancer development. These were associated with significant reduction in tumor cell proliferation and angiogenesis. Conclusion We conclude that lung cancer promotion by inflammation is partly mediated by activation of the IL-8/CXCR2 pathway and subsequent recruitment of neutrophils and release of neutrophil elastase. This provides a baseline for future clinical trials using the IL-8/CXCR2 pathway or NE inhibitors in patients with lung cancer.


Background
Lung cancer is the leading cause of cancer death worldwide accounting for 29% of all male and 26% of all female cancer deaths in 2012 [1]. Cigarette smoking (CS) is the principal cause of lung cancer, and CS induced lung cancer is characterized by a deregulated inflammatory microenvironment [2]. In addition, the association between chronic obstructive pulmonary disease (COPD), an inflammatory disease of the lung, and lung cancer has been demonstrated in population-based studies [3,4]. Smokers with COPD have an increased risk of lung cancer compared to smokers with comparable cigarette exposure but without COPD [5,6]. Importantly, among former smokers with COPD, even following withdrawal of cigarette smoke, inflammation persists and lung function continues to deteriorate as does the increased risk of lung cancer [7]. These facts suggest a strong link between airway inflammation and lung cancer promotion.
Tumor cells produce various cytokines and chemokines that attract leukocytes including neutrophils, dendritic cells, macrophages, lymphocytes and mast cells [6]. It is becoming increasingly clear that tumor-associated neutrophils (TANs) play a major role in cancer promotion [8]. During immune responses, neutrophils are among the first cells to arrive at sites of inflammation. The increased number of TANs is linked to poorer outcomes in patients with bronchioloalveolar carcinoma [9], and many patients with advanced cancer show high levels of blood neutrophils [10]. Furthermore, in histopathologic specimens of distal lung and in bronchoalveolar lavage fluid (BALF) from COPD patients, neutrophils are prominent [11,12], and shown to cause COPD progression [13]. We have previously established a COPD-like mouse model of airway inflammation induced by repetitive exposure to an aerosolized lysate of non-typeable Haemophilus influenzae (NTHi) [11], which is the most common bacterial colonizer of airways in COPD patients [14]. Then we showed that this type of airway inflammation promotes lung cancer in a K-ras mutant mouse model of lung cancer (CC-LR) [15]. This was associated with severe neutrophilic influx due to an increased level of neutrophil chemoattractant, KC, which was partially inhibited by using a natural non-specific anti-inflammatory agent, curcumin, and resulted in significant tumor suppression [16]. Therefore, we further dissected the role of neutrophils in lung tumorigenesis by selectively targeting neutrophils, its chemokine receptor (CXCR2) and its specific enzyme (neutrophil elastase). Neutrophil depletion, CXCR2 inhibition, and lack of neutrophil elastase (NE) all resulted in significant tumor reduction in our K-ras mutant mouse model of lung cancer.

Neutrophil depletion inhibits lung cancer promotion
To test the effect of neutrophil depletion on lung cancer development, we treated the CC-LR mice with mLy-6G Ab 5 mg/kg i.p. twice a week. Two groups (N = 8) of 10week-old CC-LR mice were treated with mLy-6G Ab for 4 weeks, with one of these groups exposed to the NTHi lysate once a week for 4 weeks for induction of a COPDtype inflammatory lung phenotype. Two other (N = 8) groups of mice were treated with isotype control while one of them was exposed to NTHi lysate. All groups were sacrificed one day after the fourth NTHi exposure.
We and others have shown that expression of Kras G12D within the airway epithelium of mice induces the production of chemokines which leads to the accumulation of inflammatory cells, particularly macrophages and neutrophils, within the lung [15,17,18]. In the BALF of non-NTHi exposed Ab treated CC-LR mice, the total white blood cells decreased mostly due to complete depletion of neutrophils by the mLy-6G Ab ( Figure 1A). The macrophage and lymphocyte counts were slightly reduced as well, because the mLy-6G Ab can non-specifically affect Gr-1 + monocytes/ macrophages and lymphocyte subpopulations [19]. Surprisingly, the mLy-6G Ab was not able to completely deplete the neutrophils from the BALF of CC-LR mice after repetitive NTHi exposure, while the macrophages had a 2.4 reduction ( Figure 1B).
Secreted cytokines and chemokines could both cause the recruitment of leukocytes and also help to identify the leukocyte phenotypes. Treatment with mLy-6G Ab in CC-LR mice non-exposed or exposed to NTHi resulted in a reduction in the level of neutrophil chemoattractant, KC in BALF (Table 1 and data not shown) which is consistent with reduced numbers of inflammatory cells in the BALF. CCL2 and CCL5 are phenotypic markers for TANs [20], which are protumorigenic (N2 phenotype). mLy-6G Ab inhibited these N2 type chemokine particularly CCL5 secretions (Table 1 and data not shown). Low iNOS and high arginase 1 expressions are two other important indicators of N2 type neutrophils [20][21][22]. Real-time Q-PCR analysis of the RNA extracted from the whole lung of CC-LR mice showed that treatment with mLy-6G Ab reduced the relative expression of arginase 1 in presence or absence of NTHi exposure ( Figure 1C), while it had the opposite effect on relative expression of iNOS (Additional file 1: Figure S1A). This was associated with decreased protein expression of arginase 1 in western blot (WB) analysis of the whole lung from CC-LR mice treated with mLy-6G Ab ( Figure 1C). Co-staining of lung tissue with Ly6G and arginase 1Abs showed the expression of arginase 1 in neutrophils (Additional file 1: Figure S1B). Levels of other essential proinflammatory cytokines (IL-6, IL-17, TGF-β) were slightly suppressed by mLy-6G Ab (Table 1 and data not shown). These results indicate that mLy-6G Ab could have an effect on tumor promotion by inhibiting neutrophil infiltration into the tumor microenvironment. Furthermore, mLy-6G Ab might be able to edit neutrophils phenotype from a pro-tumor phenotype (N2) toward an anti-tumor phenotype (N1).
The effects of the neutrophil depletion using mLy-6G Ab on lung tumor promotion was analyzed by determining the number of tumors visible on the pleural surface of the lungs from CC-LR mice. mLy-6G Ab significantly reduced the tumor numbers on the lung surface of CC-LR mice not exposed to NTHi lysate by~68% (3.2-fold, 51 ± 2 without mLy-6G versus 16 ± 2 with mLy-6G) and after NTHi exposure by~80% (4.9-fold, 108 ± 11 without mLy-6G versus 22 ± 2 with mLy-6G) ( Figure 1E). Histopathologic analysis showed that mLy-6G Ab treatment prevented lung cancer progression, with most of the lung tumors remaining at the early stage with less inflammatory cell infiltration both in the presence and absence of NTHi exposure ( Figure 1F).
We further stained for lung parenchymal and tumor infiltrating neutrophils using P40 Ab by IHC. We did not find any neutrophils infiltrating the lung of Ab treated non-NTHi exposed CC-LR mice compared to the lung of non-Ab treated mice (data not shown). NTHi exposure resulted in robust neutrophil infiltration which was significantly (1.6 fold) decreased by mLy-6G Ab treatment ( Figure 2A). Interestingly, there were more neutrophils infiltrated in the tumor site than the normal lung tissue, and these neutrophils in the tumor site decreased 2.1 fold in response to treatment with mLy-6G Ab ( Figure 2B).
In order to evaluate the mechanism of tumor suppression by mLy-6G Ab, we studied the rate of proliferation, angiogenesis and apoptosis. The relative number of cells showing positive staining for the cell proliferation marker, Ki-67, was measured in tumor tissues from the four groups by immunohistochemistry ( Figure 2C). The Total and lineage-specific leukocyte number in BALF of CC-LR mice treated or non-treated with mLy-6G Ab at the age of 14 weeks (mean ± SE; * = P ≤ 0.05 for CC-LR vs CC-LR plus mLy-6G). (B) Total and lineage-specific leukocyte number in BALF of NTHi-exposed CC-LR mice treated or non-treated with mLy-6G Ab collected 1 day after last NTHi aerosol exposure at the age of 14 weeks (mean ± SE; * = P ≤ 0.05 for CC-LR vs CC-LR plus mLy-6G with NTHi exposure). (C) Real-time Q-PCR expression analysis of arginase 1 on the RNA extracted from whole lung tissue (normalized to GAPDH expression level, mean ± SE; * = P ≤ 0.05 for CC-LR vs CC-LR with NTHi exposure; # = P ≤ 0.05 for CC-LR with NTHi exposure vs CC-LR with NTHi exposure plus mLy-6G treatment). (D) Western blot analysis of arginase 1 on the protein extracted from whole lung tissue. (E) Lung surface tumor number after mLy-6G Ab treatment in NTHi exposed or non-exposed 14-week-old CC-LR mice. (mean ± SE; * = P ≤ 0.05 for CC-LR vs CC-LR plus mLy-6G treatment; # = P ≤ 0.05 for CC-LR with NTHi exposure vs CC-LR with NTHi exposure plus mLy-6G treatment). (F) Histopathological appearance of lung tissue after treatment with mLy-6G Ab in NTHi exposed or non-exposed CC-LR mice. (4× magnification, scale bar = 50 mm, applicable to all panels). mLy-6G Ab effectively suppressed the expression of Ki-67 in tumor tissues compared with non-Ab treated mice from both NTHi exposed and unexposed control groups ( Figure 2D). Interestingly, in the NTHi exposed group, the infiltrating neutrophils also showed a high level of proliferation, suggesting the recruitment of immature neutrophils to the tumor site which are still going through proliferation ( Figure 2D, Additional file 1: Figure S1C). This was suppressed with mLy-6G Ab treatment. In order to study the effect of mLy-6G Ab on angiogenesis, we performed the CD31 immunostaining, which is expressed on a large portion of endothelial cells [23]. CD31 expression followed the same pattern as Ki-67 ( Figure 2E). The percentage of positive cells in the Ab treated group decreased (78%) compared to the control group in the absence of NTHi exposure. There was no significant change between the Ab treated and non-treated groups after NTHi exposure ( Figure 2F). IHC analysis of tumor cell using cleaved caspase-3 marker did not show apoptosis among different groups (data not shown).

Selective CXCR2 inhibition suppresses lung cancer promotion
As described above, KC is a very important chemo attractant for neutrophils. It could bind to the CXCR2 receptor, which will cause neutrophil recruitment [24], hence making it a good target for lung cancer treatment. We treated CC-LR mice with a selective CXCR2 inhibitor, SB332235Z (SBZ) 50 mg/kg orally, twice daily by gavage. Two groups (N = 8) of 10-week-old CC-LR mice were treated with SBZ for 4 weeks, with one of these groups exposed to NTHi lysate once a week for 4 weeks. Two other groups (N = 8) of mice were treated with vehicle control while one of them was exposed to NTHi lysate. All groups were sacrificed one day after the fourth NTHi exposure. Analysis of the BALF from SBZ treated non-NTHi exposed mice showed significant inhibition of neutrophil recruitment ( Figure 3A). However, it did not effectively suppress neutrophil recruitment after 4 weeks of NTHi exposure ( Figure 3B). Similarly, BALF KC levels also decreased by SBZ treatment in the non-NTHi exposed group, while remaining high in the NTHi exposed group (Table 1 and data not shown). Meanwhile, the levels of proinflammatory cytokines were suppressed by SBZ only in the non-NTHi group (Table 1 and data not shown). This was associated with decreased arginase 1 mRNA expression ( Figure 3C) and increased iNOS mRNA expression (Additional file 2: Figure S2A) in the whole lung of CC-LR mice after SBZ treatment similar to what we found in samples treated with mLy-6G Ab. Furthermore, WB analysis of the whole lung from CC-LR mice treated with SBZ showed decreased protein expression of arginase 1 ( Figure 3D).
Treatment with this selective CXCR2 antagonist reduced the tumor number on the lung surface of CC-LR mice not exposed to NTHi lysate by~67% (3-fold, 54 ± 5 without SBZ versus 18 ± 2 with SBZ) and after NTHi exposure by~23% (1.3-fold, 108 ± 11 without SBZ versus 83 ± 21 with SBZ) ( Figure 3E). Histopathologic analysis of lung tissue from SBZ treated CC-LR mice showed that CXCR2 inhibition prevented lung cancer progression, with most of the lung tumors remaining at the early stage with less inflammatory cell infiltration both in the presence and absence of NTHi exposure ( Figure 3F).
IHC staining of the lung tissue from mice treated with SBZ with P40 antibody showed similar results as the mice treated with mLy-6G Ab ( Figure 4A). Not many infiltrating neutrophils in the lung parenchyma were found in both non-NTHi groups. However, there was a 1.8 fold reduction in neutrophil infiltration by SBZ treatment in NTHi group ( Figure 4B). There were more neutrophils found in the tumor site than the normal lung tissue, and infiltrating neutrophils in the tumor site decreased 1.6 fold in response to SBZ treatment ( Figure 4B).
To evaluate the effect of SBZ on cell proliferation, Ki-67 was measured in tumor tissues from the four groups by immunohistochemistry ( Figure 4C). CXCR2 inhibition with SBZ effectively suppressed the expression of Ki-67 in tumor tissues compared to both NTHi exposed and nonexposed control groups. Again, high numbers of proliferating neutrophils were found infiltrated in the tumor site ( Figure 4D, Additional file 2: Figure S2B). To study the effect of CXCR2 inhibition on tumor angiogenesis, CD31 IHC staining was performed ( Figure 4E). SBZ treatment resulted in a 53% decrease in CD31 expression in the non-NTHi exposed group, and a 68% decrease in CD31 expression in the NTHi exposed group compared to controls ( Figure 4F). Similar to mLy-6G Ab treated groups, no apoptotic tumor cells was found after SBZ treatment (data not shown).

Lack of neutrophil elastase (NE) inhibits lung cancer promotion
NE is an elastolytic enzyme, which is a potential regulator of the inflammatory process [25]. Many literatures verified that NE contributes to different stages of tumorigenesis from promotion to metastasis [26]. It also participates actively in the development of COPD [27] by activating proteolytic cascades. NE knock out (KO) mice were crossed with CC-LR mice to generate mice with activating mutant K-ras in the airway epithelium with global lack of NE (CC-LR-NEKO). Lack of NE slightly changed the BALF inflammatory cell component of the CC-LR mice, mostly macrophages ( Figure 5A). However, it significantly inhibited lung cancer development by~43% (1.7-fold, 53 ± 5 in CC-LR vs 31 ± 4 in CC-LR-NEKO) ( Figure 5B). This was associated with decreased levels of IL-6 and TGF-β, ( Figure 5D). There were no changes in tumor numbers and inflammatory cell infiltration of NTHi exposed CC-LR-NEKO mice (Additional file 3: Figure S3A, and S3B), suggesting roles for other recruited inflammatory cells, such as macrophages and their products (e.g. macrophage elastase) in promotion of lung cancer by COPD-type airway inflammation.
Histopathologic study of the lung from CC-LR-NEKO mice showed that absence of NE not only caused less tumor number but also it delayed lung cancer progression from early stage hyperplastic lesions to advanced adenoma and adenocarcinoma ( Figure 5C). This was not the case for NTHi exposed CC-LR-NEKO mice compared to the NTHi exposed CC-LR mice (Additional file 3: Figure S3B). IHC analysis for tumor cell proliferation showed less Ki-67 expressing cells (69%) in the tumors from CC-LR-NEKO mice ( Figure 5E and F). In addition, there was significant decrease (48%) in the number of CD31 positive cells in the CC-LR-NEKO compared to CC-LR group ( Figure 5G and H), which indicates less angiogenesis in CC-LR-NEKO mice. Reduction of VEGF level in the BALF ( Figure 5I) and less lung tissue VEGF positive staining ( Figure 5J) in CC-LR-NEKO mice further verify the effects of NE on angiogenesis. Interestingly, we found a 1.2 fold increase in the number of cleaved-caspase-3 expressing cells in CC-LR-NEKO compared to control CC-LR mice ( Figure 5K) with most of apoptosis happening in the early stage tumor sites ( Figure 5L).

Discussion
Tumor-induced inflammation is now a cancer hallmark, and it is considered an enabling characteristic due to its contributions to the acquisition of core hallmark capabilities [28,29]. Tumors are complex tissues composed of multiple distinct cell types that interact with each other. The recruited stromal and inflammatory cells, which form the tumor microenvironment, are active participants in tumorigenesis and significantly influence most of the cancer hallmarks in the malignant cancer cells. In summary, the presence of these stromal and inflammatory cells within the tumor microenvironment is associated with enhanced tumor progression, cancer cell growth and spread, angiogenesis, invasion and tumor immunosuppression [30].
Many environmental causes and risk factors for cancer are associated with chronic inflammation. According to epidemiological studies, up to 20% of cancers are linked to chronic infections and inflammation [31]. Cigarette smoking generates an inflammatory microenvironment which can promote lung cancer [2]. On the other hand, COPD, which is a disorder characterized by an abnormal local and systemic inflammatory response, is also strongly associated with lung cancer [32]. In histopathologic specimens of distal lung and in BALF from COPD patients, neutrophils and macrophages are prominent [33]. Neutrophils make up a significant portion of the inflammatory cell infiltrate found in a wide variety of murine models and human cancers [34,35]. They are among the first cells to arrive at sites of inflammation and, as mentioned before, the increased number of TANs is linked to poorer outcomes in patients with lung cancer. Neutrophils are present within the alveolar airspaces and within the tumor parenchyma during neoplastic development in a urethane-induced mouse model of lung cancer [36]. We have previously shown that bacterial (NTHi) lysate induced COPD-like airway inflammation promotes lung cancer in a K-ras mutant mouse model of lung cancer (CC-LR) [15]. This was associated with severe neutrophilic influx due to the increased level of neutrophil chemoattractant, KC. Furthermore, we and others found that Kras mutation per se induces an inflammatory tumor microenvironment predominated by neutrophils and macrophages [15,17]. In this study, we provided more evidence to support the essential role for neutrophils in lung cancer promotion.
mLy-6G Ab is used to deplete neutrophils in murine disease models. It binds to Ly6G, which is present on neutrophils, and to Ly6C, which is expressed on neutrophils, dendritic cells, and subpopulations of lymphocytes and monocytes [19]. In our study, depletion of neutrophils by mLy-6G Ab suppressed lung tumor promotion which was associated with decreased tumor cell proliferation and angiogenesis. Similar to tumor associated macrophages (TAMs), TANs also can be divided into an antitumorigenic (N1) phenotype and a protumorigenic (N2) phenotype [20]. Analysis of cytokines and chemokines in BALF of CC-LR mice showed suppression of proinflammtory cytokines (IL-6, IL-17, TGF-β) by mLy-6G, with further inhibition of N2 type cytokines and molecules (CCL5, arginase 1). Surprisingly, mLy-6G Ab was not able to suppress the number of neutrophils in BALF of NTHiexposed mice which can imply that there exists a pool of neutrophil that can be mobilized. However, mLy-6G Ab reduced the numbers of macrophages in the BALF and the number of neutrophils infiltrated in the lung parenchyma. Interestingly, many of these infiltrating neutrophils were found to be stained by Ki-67 in the tumor site, showing a proliferating state for these cells. These cells mostly express pro-tumor N2 phenotype markers and can be viewed as myeloid-derived suppressor cells (MDSCs), which represent a heterogeneous group of immature myeloid cells comprising macrophages, monocytes, neutrophils, and DCs [37] or myeloid progenitor cells at different stages of development that have been reported to be recruited along with PMNs to sites of inflammation and proliferate in tissue [38].
Interleukin-8 (IL-8), which is known as keratinocytederived chemokine (KC) in the murine, is a proinflammatory CXC chemokine associated with the neutrophil chemotaxis and degranulation [24]. It is also reported to be a transcriptional target of Ras Signaling, but its role in Ras-induced tumorigenesis has not been fully defined [39].  [40][41][42]. According to our previous study, the expression of KC in the BALF of the CC-LR mice was higher compared to the WT mice [15]. The application of curcumin, which is a naturally occurring polyphenolic phytochemical isolated from the rhizome of the medical plant Curcuma longa, with non-specific anti-inflammatory effect, could suppress lung tumor promotion partly by suppressing KC production and subsequent inhibition of neutrophil recruitment to the lung [16]. In the current study, KC level was also suppressed by mLy-6G Ab in the non-NTHiexposed group.

Expression and secretion of IL-8 by tumor cells enhance proliferation and neutrophil infiltration into the tumor
The protein encoded by the CXCR2 gene is a member of the G-protein couple receptor family. It is the receptor for IL-8. This receptor mediates neutrophil migration to sites of inflammation and is essential for neutrophil recruitment [43,44]. There is a remarkable up-regulation of CXCR2 in airway epithelial cells in COPD, and this correlates with the increased number of neutrophils in the airways [45]. Neutrophil recruitment into tumors also appears to be dependent on chemokines that bind to CXCR2 expressed by neutrophils [40]. Furthermore, CXC chemokines are potent promoters of angiogenesis, and mediate their angiogenic activity via signal-coupling of CXCR2 on endothelium [41,46,47]. It has recently been shown that CXCR2 expression in tumor cells is a poor prognostic factor and promotes invasion and metastasis in lung adenocarcinoma [48]. We found that inhibiting CXCR2 by a selective antagonist could efficiently suppress neutrophilic inflammation and lung tumor progression. This was associated with significant reduction in tumor cell proliferation and angiogenesis. This is similar to the findings from other groups showing that CXCR2 inhibition using a neutralizing antibody inhibits the progression of premalignant alveolar lesions by reducing vascular density in another mouse model of lung cancer [49]. Analysis of the cytokines and chemokines in the BALF of CC-LR mice showed that the secretion of pro-inflammatory cytokines (IL-6, IL-17, TGF-β) decreased after treatment with CXCR2 inhibitor with further reduction in levels of N2 type markers, CCL-5. This was associated with decreased expression of arginase 1, while the expression of iNOS increased. These results indicate that CXCR2 inhibition not only suppresses the recruitment of neutrophils but also may shift their differentiation from a N2 phenotype to a N1 phenotype. There was a much larger reduction in angiogenesis with CXCR2 inhibition as opposed to neutrophil depletion with mLy-6G Ab treatment. This could be due to inhibition of additive direct effect of CXCR2 axis activation on tumor endothelial cells that leads to less tumor angiogenesis.
Similar to mLy-6G Ab treatment, neutrophils in BALF of NTHi-exposed mice were not affected by the CXCR2 inhibitor, but it reduced the number of infiltrating neutrophils with MDSC/progenitor cell characteristics in the lung parenchyma and tumor sites. Failure of the antineutrophil Ab and CXCR2 inhibitor in inhibiting BALF neutrophilic influx in response to repetitive exposure to NTHi indicates involvement of other chemokines and inflammatory signaling pathways besides the IL-8/CXCR2 pathway in this phenomenon.
Products secreted from neutrophils, such as, reactive oxygen species and proteinases, have defined and specific roles in regulating tumor cell proliferation, angiogenesis, and metastasis [34]. Among them, NE is generally considered the major effector of neutrophil function, making up approximately 2% of the dry weight of a neutrophil [50]. It contributes to cigarette smoke-induced emphysema in mice [51]. NE KO mice are significantly protected from the development of CS-induced emphysema and recruitment of both neutrophils and macrophages. We have also found that lack of NE reduces lung macrophages and inhibits lung cancer promotion in our K-ras mutant mouse model indicating an essential promoting role for this enzyme in lung tumorigenesis. This is consistent with the study from Houghton et al., where they found the similar promoting effect for NE through degradation of IRS-1 in another mouse model of lung cancer [52]. NE exposure can activate PI3K/Akt pathway resulting in tumor cell proliferation and survival [52]. Lack of NE in our K-ras mutant model was associated with decreased proliferation and increased apoptosis of tumor cells which could be secondary to lack of PI3K/Akt pathway activity in absence of NE. Tumor regression was also accompanied with decreased levels of TGF-β similar to the finding by Kang-Yun et al., showing that NE up-regulates TGF-β gene expression and release via the MyD88/IRAK/NF-κB pathway [53]. It has been previously shown that TANs were involved in tumor angiogenesis by the production of proangiogenic factors such as VEGF and IL-8 [8], proteases such as MMPs [54] and elastases [55]. We have also found decreased expression of VEGF and numbers of CD31 positive cells in our K-ras mutant mouse with lack of NE, indicating a role for NE in tumor angiogenesis. It has also been shown that NE can cleave VEGF to generate a diffusible VEGF fragments that stimulate inflammatory cell recruitment and activation via VEGF receptor 1/Akt pathway [56] that could be considered as a tumor promoting mechanism of NE in our model. Similar to our findings Wada et al. have demonstrated that NE released from activated neutrophils stimulates the growth and progression of cancer cells by releasing the growth factors such as VEGF on the cell surface and that sivelestat, a specific NE inhibitor, blocks these processes [57].
It is known that K-ras induces the secretion of the cytokine IL-6 in various epithelial cell types including lung epithelial cells [58]. We have previously revealed increased levels of IL-6 and subsequent activation of the STAT3 pathway in our K-ras induced mouse model of lung cancer in absence and presence of COPD-like airway inflammation [59]. We further demonstrated that genetic ablation of IL-6 results in significant tumor suppression in this model in vivo, indicating an essential role for IL-6 in lung cancer promotion in this model. Therefore, reduced levels of IL-6 in response to neutrophil depletion, CXCR2 inhibition, or lack of NE could be considered as another mechanism of tumor regression. However, this could simply be a reflection of the reduced production of this cytokine from K-ras mutant cells due to reduction in tumor burden.

Conclusions
Taken together, we confirmed that neutrophils play a very important role in the promotion of lung cancer which is strongly mediated through the IL-8/CXCR2 pathway and release of NE and development of a type 2 protumor microenvironment. This provides a baseline for future clinical trials and/or prevention trials using the IL-8/CXCR2 pathway or NE inhibitors in patients with lung cancer and high risk COPD patients.

Materials and methods
Animals CCSP Cre /LSL-K-ras G12D mice (CC-LR) were generated as previously described [15]. Briefly, this is a mouse generated by crossing a mouse harboring the LSL-K-ras G12D allele with a mouse containing Cre recombinase inserted into the Clara cell secretory protein (CCSP) locus [15]. CC-LR mice were crossed with neutrophil elastase (NE) knock out (KO) mice [52,60], which were kindly provided by Dr. Steven Shapiro (University of Pittsburgh) to generate CC-LR-NEKO mice. All mice were housed in specific pathogen-free conditions and handled in accordance with the Institutional Animal Care and Use Committee of M. D. Anderson Cancer Center. Mice were monitored daily for evidence of disease or death.

NTHi lysate aerosol exposure
A lysate of NTHi strain 12 was prepared as previously described [15], the protein concentration was adjusted to 2.5 mg/ml in phosphate buffered saline (PBS), and the lysate was frozen in 10 ml aliquots at −80°C. To deliver the lysate to mice by aerosol, a thawed aliquot was placed in an AeroMist CA-209 nebulizer (CIS-US, Bedford, MA) driven by 10 liter/min of room air supplemented with 5% CO 2 for 20 min.

Histochemistry
The tracheas of euthanized mice were cannulated with PE-50 tubing and sutured into place. The lungs were infused with 10% buffered formalin (Sigma, St. Louis, MO) and then removed and placed in 10% buffered formalin for 18 hours. Tissues then were transferred to 75% ethanol, embedded in paraffin blocks, and sectioned at 5-mm thickness. The sections on glass slides were dried at 60°C for 15 minutes and then were deparaffinized and stained with hematoxylin and eosin (H&E) by incubating the tissues in Harris hematoxylin (Sigma) followed by serial eosin (Sigma) and graded ethanol steps. The H&E-stained slides were examined by a pathologist blinded to genotype and treatment, and the proliferative lesions of the lungs were evaluated in accordance with the recommendations of the Mouse Models of Human Cancer Consortium [61]. Fluor 594 (Invitrogen). All antibodies reacted with mouse antigens. Heat-induced antigen retrieval was performed using 10 mmol/L of citrate buffer (pH 6) in a pressure cooker for 20 min. Endogenous peroxidase was quenched with 3% hydrogen peroxide for 30 min at room temperature. Blocking was performed with nonimmune normal serum. Immunoreactivity for immunohistochemistry was detected using biotinylated IgG secondary antibodies specific for each primary antibody followed by incubation with ABC kit (Vector Laboratory, Burlingame, CA) for 30 min, and stained with diaminobenzidine chromogenic substrate for 4-10 min. Slides were counter-stained with Harris hematoxylin in for 30 sec, followed by dehydration and mounted with cytoseal 60 (ThermoFisher Scientific, Cheshire, UK). Images were obtained by an OLYMPUS BX 60 inverted microscope at ×4, ×20 or ×40 magnification with Image-Pro Plus, version 4.5.1.22. The numbers of labeled positive cells for any of these markers were quantitated as a fraction of total tumor nuclei per high power field (40×) in 10 fields from three mice of each group. Results were expressed as percentage of positive cells ± SE.

Assessment of lung tumor burden and inflammation
On the first day after the final NTHi exposure, animals were euthanized by i.p. injection of a lethal dose of avertin (Sigma). In all mice (n = 8 per group per time point), lung surface tumor numbers were counted and then in some of them (n = 8 per group per time point), the lungs were prepared for histologic analysis as described above. In other mice (n = 8 per group per time point), bronchoalveolar lavage fluid (BALF) was obtained by sequentially instilling and collecting 2 aliquots of 1 mL PBS through a tracheostomy cannula. Total leukocyte count was determined using a hemacytometer, and cell populations were determined by cytocentrifugation of 300 mL of BALF followed by Wright-Giemsa staining. The remaining BALF (1,400 mL) was centrifuged at 1,250 × g for 10 min, and supernatants were collected and stored at −80°C.

Cytokines and chemokines measurement
The levels of KC, TNF, IL-6, IL-10, IL-17A, CCL2, CCL5 and VEGF in the BALF were assessed using MCYTOMAG-70 K assay (Millipore, St Charles, MO), according to the manufacturer's instructions). Due to reagent incompatibilities, TGF-β was assayed separately from the other cytokines using the TGFB-64 K-01 (Millipore, MO, USA) assay. Data were collected using a Luminex 100 (Luminex Corporation, TX, USA). Standard curves were generated using a 5-parameter logistic curvefitting equation weighted by 1/y (StarStation V 2.0; Applied Cytometry Systems, CA, USA). Each sample reading was interpolated from the appropriate standard curve.

Quantitative RT-PCR analysis
Total RNA was isolated from whole lung according to the TRIzol reagent protocol (Invitrogen, NY, USA) and purified by E.Z.N.A. total RNA kit I (OMEGA, GA, USA). Reverse transcription PCR was performed using the qScript cDNA SuperMix (Quanta biosciences, MD, USA). Real-time PCR was carried out according to a standard protocol using the PerfeCTaFast Mix II (Quanta) with the TaqMan probe for arginase 1 and iNOS (ABI, NY, USA), and products measured on an ABI Viia 7 PCR system (ABI). Relative expression of each gene was calculated and graphed for comparison among the treatment groups. GAPDH RNA was measured for reference.

Western blot analysis
Total proteins were prepared from each group of mouse lungs. Lung samples were removed and immediately placed in RIPA buffer and a protease inhibitor mixture. The samples were then homogenized and centrifuged at 14000 g for 20 min at 4°C. The supernatants were collected as the total proteins. Equal amounts (40 μg) of the total proteins were boiled for 5 min in the presence of loading buffer, loaded on each lane, and separated by 10% SDS-PAGE. The gels were then transferred to nitrocellulose membranes. Equal amounts of protein loading for each lane was checked by Ponceau (Sigma Chemical Co., MO, USA) staining. The anti-arginase 1 (Thermo Scientific, clone PA5-22009, MA, USA) antibody was diluted to the concentration according to commercial recommendations (1:1000). Immunoreactive bands were detected with Pierce ECL Western Blotting Substrate (Pierce, IL, USA).

Statistical methods
Summary statistics for cell counts in BALF, IHC positive cells in lung tissue, and mRNA expression were computed within treatment groups, and analysis of variance with adjustment for multiple comparisons was conducted to examine the differences between the control group and each of the treatment groups in the presence or absence of NTHi exposure. For tumor counts, comparisons of groups were made using Student's t test. Differences were considered significant for P < 0.05. statistical analysis, and participated in preparing the figures and drafting the manuscript. All authors read and approved the final manuscript.