Podoplanin expression by cancer associated fibroblasts predicts poor prognosis of lung adenocarcinoma

Abstract

Recent studies have reported increased podoplanin expression by cancer cells and stromal cells, but little is known about its expression and biological significance in adenocarcinoma of the lung. We examined podoplanin expression by both cancer cells and stromal cells in 177 consecutive lung adenocarcinoma cases and analyzed relations between podoplanin expression and both clinicopathological factors and outcome. Podoplanin expression was observed on the apical membrane of the cancer cells in only 9 of the 177 (5.1%) cases. By contrast, cancer-associated fibroblasts (CAFs) were found to express podoplanin in 54 cases (30.5%). Podoplanin (+) CAFs were found only in invasive adenocarcinoma and none were found in noninvasive adenocarcinoma. Conventional prognostic factors were significantly correlated with podoplanin expression by CAFs. The univariate analyses and log-rank test showed that podoplanin expression was significantly associated with shorter survival time (p < 0.001 and p < 0.001, respectively). We divided the cases into 3 groups according grade based on the proportion of CAFs expressing podoplanin [a grade 0 group (n = 123), a grade 1 group (n = 36) and a grade 2 group (n = 18)]. The result showed that conventional prognostic factors were significantly correlated with the grade of podoplanin expression by CAFs. Furthermore, the grade 2 group tended to have a shorter survival time than the grade 1 group (p = 0.092). The results of this study highlight the importance of podoplanin expression by CAFs and provide new insights into the biology of the cancer microenvironment in adenocarcinoma of the lung. © 2008 Wiley-Liss, Inc.

Primary adenocarcinoma of the lung is known to have a poor prognosis, even if surgery is successful.1 Primary adenocarcinoma of the lung is increasing in frequency and now account for almost half of all nonsmall cell lung carcinomas. Improved knowledge of the molecular mechanisms underlying the development and progression of lung adenocarcinoma is essential to the development and establishment of effective therapeutic modalities.

Human podoplanin is a type-1 transmembrane sialomucin-like glycoprotein consisting of an extracellular domain, a single transmembrane portion and a short cytoplasmic tail for protein kinase C and cAMP phosphorylation.2, 3 Because podoplanin is expressed on lymphatic vessel endothelium and not on blood vessel endothelium, it is widely used in histopathology as a specific marker for lymphatic endothelium.3, 4 However, recent studies have reported increased podoplanin expression in follicular dendritic cells,5 stromal fibroblasts in chronic pleuritis,6 squamous cell carcinoma,2, 4, 7, 8 malignant mesothelioma9 and solitary fibrous tumor.10 Expression of podoplanin has also been reported to be associated with the outcome of oral and uterine cervix squamous cell carcinoma4, 8 suggesting that podoplanin contributes biological behavior during cancer progression.

Cancer cells coexist with several stromal cell types that together create the microenvironment of the cancer. The main constituents of stromal cell types are inflammatory cells, including lymphocytes, granulocytes and macrophages, the endothelial cells of blood and lymph vessels, pericytes and fibroblasts. Fibroblasts recruited into cancer tissue, called cancer-associated fibroblasts (CAFs), can produce collagens and extracellular matrix proteins in response to several extracellular stimuli and influence the cancer cell progression.11–16 The precise mechanisms underlying communication among stromal cells, including CAFs, cancer cells, and the environment, have not been completely identified, and analysis of the phenotypes of both stromal cells and cancer cells may be crucial to identifying them.

Although podoplanin expression has been reported to be biologically significant in squamous cell carcinoma,4, 7, 8 the significance of its expression in adenocarcinoma of the lung has not been clarified. Podoplanin expression has also been observed in fibroblasts during chronic inflammation.6 To determine the biological significance of podoplanin expression in adenocarcinoma of the lung, we examined its expression by both cancer cells and stromal cells in 177 consecutive cases of adenocarcinoma of the lung.

Material and methods

Patients

During the period from January 1994 to December 1996, a total of 409 patients with primary lung carcinoma were treated by surgical resection at the National Cancer Center Hospital East, Chiba, Japan, and we reviewed the cases of 188 consecutive patients in whom complete resection of adenocarcinoma of the lung has been archived as potential candidates for inclusion on this study. All signed an Institutional Review Board-approved informed consent. The tumors were staged according to the International Union Against Cancer's tumor-node-metastasis classification and were histologically subtyped and graded according to the third edition of the World Health Organization (WHO) guidelines. After excluding 11 patients because of the poor quality of the specimen obtained, 177 patients were ultimately included in this study, and median follow-up period for these patients was 9.8 years. Survival time was measured from the date of surgery.

Pathological studies

All surgical specimens were fixed with 10% formalin and embedded in paraffin. The tumors were cut at approximately 5-mm intervals, and serial 4-μm sections were stained with hematoxylin and eosin, and by the Alcian blue-periodic acid Schiff method to visualize cytoplasmic mucin production and the Verhoeff-van-Gieson method to visualize elastic fibers. Lymphatic invasion and pulmonary metastasis were evaluated in sections stained with hematoxylin and eosin. Vascular and pleural invasion was evaluated by the Verhoeff-van-Gieson method. Two observers (A.K. and G.I.) who are unaware of the clinical data independently reviewed all pathological slides. The histological diagnoses were based on the revised WHO histological classification. Tumor size was measured as the maximal diameter on the cut section of the lung. The pathological stage was determined according to the classification of the Union Internationale Contre le Cancer.

Evaluation of clinicopathological factors

Clinical characteristics were retrieved from the clinical records available. The following clinicopathological factors were investigated retrospectively to assess their impact on patient survival : gender, age (<70 years vs. >= 70 years), smoking history (never-smokers vs. smokers), tumor size (=<3 cm vs. >3 cm), pathological nodal involvement (positive vs. negative), pathological stage (I vs. II, III, or IV), grade of differentiation (well differentiated vs. moderately or poorly differentiated), vascular invasion (present vs. absent), lymphatic permeation (present vs. absent) and pleural invasion (present vs. absent).

Antibodies and immunohistochemistry

After reviewing the hematoxylin- and eosin-stained slides of the surgical specimens, the block containing the most extensive tumor component was selected from each specimen. Sections, 4-μm each, were cut from the paraffin blocks and mounted on silanized slides. The sections were deparaffinized in xylene, dehydrated in a graded ethanol series, and after washing with distilled water, the sections were placed in 0.1 M citric acid buffer. For antigen retrieval, the slides were heated twice at 95°C for 20 min in a microwave oven (H2800 Microwave Processor, Energy Beam Sciences, East Granby, CT) and then allowed to cool for 1 hr at room temperature. Next, the slides were washed 3 times in phosphate-buffered saline (PBS) and immersed in a 0.3% hydrogen peroxide solution in methanol for 15 min to inhibit endogenous peroxidase activity. After washing the slides 3 times in PBS, nonspecific binding was blocked by preincubation with 2% normal swine serum in PBS (blocking buffer) for 30 min at room temperature. Individual slides were then incubated overnight at 4°C with mouse anti-podoplanin antibody (D2-40, Signet Laboratories, Dedham, MA) at a final dilution of 1:50 in the blocking buffer, and mouse anti Human CD31 (clone JC70A, DakoCytomation Denmark A/S, Produktionsvej 42, DK-2600 Glostrup, Denmark), at a final dilution of 1:40 in the blocking buffer. The slides were again washed 3 times with PBS then incubated with EnVision™ (DAKO, Denmark) for 1 hr at room temperature, and after extensive washing with PBS, the color reaction was developed for 3 min in 2% 3, 3′-diaminobenzidine in 50 mM Tris-buffer (pH 7.6) containing 0.3% hydrogen peroxidase. Finally, the sections were counterstained with Meyer's hematoxylin, dehydrated and mounted.10 When at least 10% of the cancer cells or stromal fibroblasts showed an unequivocally strong reaction that was the same as the strength of the reaction of type I pneumocytes or lymphatic endothelial cells, the case was classified as positive (Fig. 1).

Figure 1. Immunohistochemical analysis of podoplanin expression in adenocarcinoma of the lung, and immunoblot and immunofluorescence analysis of podoplanin expression by CAFs. (a) Cases of lung adenocarcinoma in which the cancer cells expressed podoplanin. The arrow head pointed to the internal positive control (lymphatic vessel) (b) Podoplanin was expressed in the apical membrane of cancer cells. (c) Podoplanin expression by the cancer stroma of lung adenocarcinoma. (d) Podoplanin expression by stromal spindle cells that were morphologically identified as fibroblasts. (e,f) Result of CD31 staining at the same site as c and d. CD31 expressed in endothelium only lymphatic or blood vessels. (g) Western blotting of NCAFs and cancer associated fibroblasts (CAFs). NCAFs were negative for the anti-podoplanin antibody (D2-40) in 3 cases. In 2 cases (case 2 and 3), CAFs were positive for anti-podoplanin antibody (D2-40). (h) Results of double immunofluorescence staining. The upper left panel shows cells immunostained with anti-podoplanin antibody (D2-40). The upper right panel shows cells in the same area immunostained with α-SMA antibody. The lower left panel shows cells stained with DRAQ5™ to identify nucleated cells. The lower right panel shows a composite of images obtained with both fluorophores.

Immunofluorescence and confocal microscopy

Paraffin-embedded specimens were cut into 4-μm thick sections, and the sections were subjected to a microwave-based antigen-retrieval technique. After incubating the sections with the primary antibodies, including mouse anti-podoplanin antibody at a 1:50 dilution; and rabbit polyclonal anti-α-SMA (Actin, Smooth Muscle, LAB VISION, Corp., Fremont, CA) at a 1:50 dilution, the sections were washed and either Alexa Fluor 488 goat anti-mouse IgG or Alexa Fluor 546 goat anti-rabbit IgG (Molecular Probes) was used as the secondary antibody. Before mounting, all sections were stained with DRAQ5™ (Alexis Biochemical, Lausen, Switzerland) to identify nucleated cells. After mounting, the sections were examined with an LSM5 Pascal confocal imaging system (Carl Zeiss, Jena, Germany), and them with an inverted microscope at an excitation wavelength of 488 nm for Alexa Fluor 488, 568 nm for Alexa Fluor 546 and 633 nm for DRAQ5™. Confocal images were stored as digital files and viewed with Photoshop software (Adobe, Mountain View, CA).

Fibroblast culture

CAFs and non-cancer-associated fibroblasts (NCAFs) were prepared from human lung cancer tissue and non-cancerous lung tissue in the same sample. Briefly, an approximately 5-mm³ sample of carcinoma from each tissue specimen was cut into about 15subdivisions and placed in αMEM culture containing with 10% heat-inactivated FBS and antibiotics (penicillin and streptomycin). The medium was changed every other day until the tissue was surrounded by adherent fibroblasts. After 10 to 20 days of growth, the fibroblasts were separated from contaminating epithelial and endothelial cells by differential trypsinization. When the cells reached 80% confluence they were harvested with 0.25% trypsin and 1 mmol/l ethylene-diamine-tetra-acetic acid and replated at a density of 1 × 10⁴ cells/cm². All specimens were collected after the subjects gave their written informed consent and it was approved by the Institutional Review Board of the National Cancer Center.

Western blotting

Western blot analysis was performed as follows. Cells were lysed in whole-cell extraction buffer (20 mM Hepes-NaOH, 0.5% NP-40, 15% glycerol) containing Complete, a protease inhibitor cocktail tablet (Roche Diagnostics, Mannheim, Germany). Proteins were separated on 12% SDS-polyacrylamide gel and transferred to an Immobilon-P, PVDF (polyvinylidene fluoride) membrane (MILLIPORE, Billerica, MA). Blots were saturated with blocking buffer (5% skim milk in TBS-T) for 1 hr at room temperature and then incubated overnight at 4°C with anti-human mouse monoclonal podoplanin antibody (D2-40), or polyclonal goat actin antibody (Santa Cruz, Santa Cruz, CA). After washing in TBS-T, membranes were incubated for 1 hr at room temperature with HRP-Rabbit Anti-mouse IgG or HRP-rabbit Anti-goat IgG (Zymed, San Francisco, CA). ECL Western Blotting Detection Reagents (GE Healthcare, Buckinghamshire, UK) were used to develop high-performance chemiluminescence film (GE Healthcare, Buckinghamshire, UK).

Statistical analysis

The correlations between podoplanin expression level in cancer cells or stromal cells and the clinicopathological factors were evaluated by the χ²-test or Fisher's exact test, as appropriate. The correlations between the grade of podoplanin expression by stromal cells and clinicopathological factors were evaluated by the Spearman's rank correlation coefficient. Overall survival was measured from the date of surgery to the date of death from any cause or the date the patient was last known to be alive. Survival curves were estimated by the Kaplan-Meier method, and differences in survival between subgroups were compared by the log-rank test. A p value less than 0.05 were considered significant. Statistical analysis software (SPSS, Version 11.0) was used to perform the analyses.

Results

Podoplanin expression by cancer cells and cancer-associated fibroblasts

A series of 177 specimens of adenocarcinoma of the lung were examined for podoplanin expression (Table I). Podoplanin expression at the apical membrane of the cancer cells was observed in only 9 of the 177 (5.1%) cases (Figs. 1a and 1b). By contrast, podoplanin expression was observed by stromal spindle cells that were morphologically identified as fibroblasts in 54 cases (30.5%) (Figs. 1c and 1d). Figures 1e and 1f showed the results of CD31 staining of the same area as 1c and 1d. These indicated that most of podoplanin-positive stromal spindle cells did not include the endothelium of lymphatic or blood vessels.

Table I. Immunostaining Results for the Anti-Podoplanin Antibody

	Cases	Cancer cells		Cancer associated fibroblasts
		Negative (%)	Positive (%)	Negative (%)	Positive (%)
All cases	177	168 (94.9)	9 (5.1)	123 (69.5)	54 (30.5)
Non-invasive	11	10 (90.9)	1 (9.1)	11 (100)	0 (0)
Invasive	166	158 (95.2)	8 (4.8)	112 (67.5)	54 (32.5)

To confirm that cancer associated fibroblasts (CAFs) express podoplanin, cell lysates from cultured fibroblasts isolated from lung cancer tissue and from non-cancerous lung tissues were run on SDS-PAGE and immunoblotted with the antibodies to podoplanin (Fig. 1g). Fibroblasts from the lung cancer tissue of 2 patients reacted strongly with anti-podoplanin Ab, indicating that CAFs express podoplanin. By contrast, none of the fibroblasts from non-cancerous lung tissue reacted strongly. Figure 1h shows the results of double immunofluorescence staining of CAFs. Some CAFs were positive for both podoplanin and αSMA indicating that myofibroblasts also express podoplanin.

Cases with podoplanin expression by CAFs were limited in invasive adenocarcinoma of the lung. In non-invasive adenocarcinoma (100%-BAC), all cases showed podoplanin-negative (Table I).

Relationship between podoplanin expression by CAFs and clinicopathological factors

Relationship between podoplanin expression by CAFs and clinicopathological factors is shown in Table II. Podoplanin expression by CAFs was significantly correlated with smoking history (p = 0.003), tumor size (p = 0.021), pathological nodal involvement (p = 0.033), pathological stage (p < 0.001), tumor differentiation (p < 0.001), vascular invasion (p < 0.001) and pleural invasion (p = 0.002).

Table II. Relationship Between Podoplanin Expression by CAFs and Clinicopathological Factors

Variables (n = 177)	Cases	Podoplanin expression		p value
		Negative (%)	Positive (%)
Gender
Male	86	54 (62.8)	32 (37.2)
Female	91	69 (75.8)	22 (24.2)	0.073
Age
<70	124	84 (67.7)	40 (32.3)
>70	53	39 (73.6)	14 (26.4)	0.48
Smoking history
Never-smokers	92	73 (79.3)	19 (20.7)
Smokers	85	50 (58.8)	35 (41.2)	0.003
Tumor size
≤3.0 cm	100	77 (77.0)	23 (23.0)
>3.0 cm	77	46 (59.7)	31 (40.3)	0.021
pN
pN0	123	92 (74.8)	31 (25.2)
pN1 or pN2	54	31 (57.4)	23 (42.6)	0.033
Pathological stage
I	110	89 (80.9)	21 (19.1)
II/III/IV	67	34 (50.7)	33 (49.3)	<0.001
Differentiation
Well	91	80 (87.9)	11 (12.1)
Moderately/poorly	86	43 (50.0)	43 (50.0)	<0.001
Vascular invasion
Absent	90	79 (87.8)	11 (12.2)
Present	87	44 (50.6)	43 (49.4)	<0.001
Lymphatic permeation
Absent	101	76 (75.2)	25 (24.8)
Present	76	47 (61.8)	29 (38.2)	0.07
Pleural invasion
Absent	117	92 (78.6)	25 (21.4)
Present	60	31 (51.7)	29 (48.3)	0.002

Relationship between podoplanin expression by CAFs and overall survival

Univariate analyses by the Cox proportional hazards model were performed to determine the prognostic value of podoplanin expression by CAFs (Table III). Podoplanin expression was significantly correlated with shorter survival time (p < 0.001). In addition, tumor size (p = 0.004), pathological nodal involvement (p < 0.001), pathological stage (p < 0.001), tumor differentiation (p < 0.001), vascular invasion (p < 0.001), lymphatic permeation (p < 0.001) and pleural invasion (p < 0.001) were also correlated with shorter survival time.

Table III. Prognostic Significance for Overall Survival (Univariate Analysis)

Variables (n = 177)	Cases	Hazard ratio	95% CI	p value
Gender
Male	86	1
Female	91	0.82	0.519–1.293	0.392
Age
<70	124	1
≥70	51	1.231	0.756–2.005	0.403
Smoking history
Never-smokers	92	1
Smokers	85	1.386	0.878–2.188	0.161
Tumor size
≤3.0 cm	100	1
>3.0 cm	77	1.978	1.250–3.130	0.004
pN
pN0	123	1
pN1 or pN2	54	3.998	2.516–6.352	<0.001
Pathological stage
I	110	1
II/III/IV	67	4.425	2.750–7.121	<0.001
Tumor invasion
Noninvasive	11	1
Invasive	166	2.984	0.732–12.166	0.127
Differentiation
Well	91	1
Moderately/poorly	86	2.71	1.678–4.375	<0.001
Vascular invasion
Absent	90	1
Present	87	3.621	2.180–6.014	<0.001
Lymphatic permeation
Absent	101	1
Present	76	3.092	1.921–4.978	<0.001
Pleural invasion
Absent	117	1
Present	60	2.484	1.572–3.925	<0.001
Podoplanin expression by CAFs
Negative	123	1
Positive	54	2.872	1.817–4.542	<0.001

The overall survival curves obtained by the Kaplan-Meier method, with statistical significance assessed using the log-rank test, are shown in Figure 2. The overall survival time of the patients whose CAFs were podoplanin-positive was significantly shorter in all pathological stages and in pathological stage I than that of the patients whose CAFs were negative (Figs. 2a and 2b) (p < 0.001 and p = 0.006, respectively), but the differences in overall survival time in pathological stages II/III/IV cases were not significant (Fig. 2c).

Figure 2. Overall survival curve of podoplanin expression by CAFs cases. (a) Overall survival curve. (b) Survival curve of pStage I cases. (c) Survival curves of pStage II/III/IV cases. (d) Relationship between the grade of podoplanin expression by CAFs and overall survival. (grade 0: Podoplanin-positive CAF area/stromal area × 100 = <10%, grade 1: Podoplanin-positive CAF area/stromal area × 100 = 10−50% and grade 2: Podoplanin-positive CAF area/stromal area × 100 = >50%).

Relationship between grade of podoplanin expression by CAFs and overall survival time

To confirm that the volume of CAFs expressing podoplanin influences the malignant behavior of cancers, we graded podoplanin expression by CAFs as follows: grade 0, podoplanin-positive CAF area/stromal area ×100 =<10%; grade 1, podoplanin-positive CAF area/stromal area ×100 = 10–50%; grade 2, podoplanin-positive CAF area/stromal area ×100 =>50% and analyzed grades of podoplanin expression by CAFs and clinicopathological factors for correlations (Table IV). The results showed that smoking history (p = 0.003), tumor size (p = 0.007), pathological nodal involvement (p = 0.008), pathological stage (p < 0.001), tumor differentiation (p < 0.001), vascular invasion (p < 0.001), lymphatic permeation (p = 0.026) and pleural invasion (p < 0.001) were significantly correlated with the grade of podoplanin expression by CAFs. The overall survival curves of each grade of podoplanin expression by CAFs of all pathological stage cases and outcome are shown in Figure 2d. The grade 1 cases and grade 2 cases had a significantly shorter overall survival time than the grade 0 cases (p = 0.002 and p < 0.001, respectively), and overall survival time in the grade 2 cases tended to be shorter than in the grade 1 cases (p = 0.092).

Table IV. Relationship Between Grade of Podoplanin Expression by CAFs and Clinicopathological Factor

Variables (n = 177)	Cases	Podoplanin expression			p value
		Grade 0 (%)	Grade 1 (%)	Grade 2 (%)
Gender
Male	86	54 (62.8)	21 (24.4)	11 (12.8)
Female	91	69 (75.8)	15 (16.5)	7 (8.7)	0.06
Age
<70	124	84 (67.7)	28 (22.6)	12 (9.7)
≥70	53	39 (73.6)	8 (15.1)	6 (11.3)	0.539
Smoking history
Never-smokers	92	73 (79.3)	13 (14.1)	6 (6.5)
Smokers	85	50 (58.8)	23 (27.1)	12 (14.1)	0.003
Tumor size
≤3.0 cm	100	77 (77.0)	18 (18.0)	5 (5.0)
>3.0 cm	77	46 (59.7)	18 (23.4)	13 (16.9)	0.007
pN
pN0	123	92 (74.8)	24 (19.5)	7 (5.7)
pN1 or pN2	54	31 (57.4)	12 (22.2)	11 (20.4)	0.008
Pathological stage
I	110	89 (80.9)	17 (15.5)	4 (3.6)
II/III/IV	67	34 (50.7)	19 (28.4)	14 (20.9)	<0.001
Differentiation
Well	91	80 (87.9)	9 (9.9)	2 (2.2)
Moderately/poorly	86	43 (50.0)	27 (31.4)	16 (18.6)	<0.001
Vascular invasion
Absent	90	79 (87.8)	10 (11.1)	1 (1.1)
Present	87	44 (50.6)	26 (29.9)	17 (19.5)	<0.001
Lymphatic permeation
Absent	101	76 (75.2)	20 (19.8)	5 (5.0)
Present	76	47 (61.8)	16 (21.1)	13 (17.1)	0.026
Pleural invasion
Absent	117	92 (78.6)	17 (14.5)	8 (6.8)
Present	60	31 (51.7)	19 (31.7)	10 (16.7)	<0.001

Discussion

In the present study, podoplanin expression by CAFs were found only in invasive adenocarcinoma of the lung, and they were never found in noninvasive adenocarcinoma. Podoplanin expression by CAFs was correlated with several biological factors: tumor size, pathological nodal involvement, pathological stage, tumor invasion, tumor differentiation, vascular invasion and pleural invasion, and its expression was a significant predictor of an unfavorable outcome. The results also showed that the grade of podoplanin expression by CAFs was correlated with smoking history, tumor size, pathological nodal involvement, pathological stage, tumor differentiation, vascular invasion, lymphatic permeation and pleural invasion. Taken together, these finding may mean that CAFs expressing podoplanin contribute to the prognosis of adenocarcinoma of the lung.

Podoplanin on cancer cells has been reported to act as a platelet-aggregating factor. Mahalingam et al. reported a high-distant metastasis rate by tumor cells that strongly aggregate platelets, a low metastasis rate by tumor cells that weakly aggregate platelets, suggesting that the platelet aggregation activity of tumor cells is correlated with the metastasis rate.17 Platelet activation by tumor cells is thought to protect them from immune surveillance, to promote tumor cell adhesion to the vascular endothelium, which is the first step in extravasation, and to provide permeability, which should facilitate extravasation and metastasis because aggregated platelets contacted tumor cells directly and surrounded them.18–20 The aforementioned findings suggest that platelets may play an important role in the promotion of tumor cell invasion.21 Suzuki-Inoue et al. reported that podoplanin expression in cancer cells promotes platelet aggregation and that it may also be involved in cancer cell migration, invasion, metastasis, and malignant progression.22 That may suggest that CAFs expressing podoplanin also have the ability to aggregate platelets and contribute to the aggressive behavior of cancer cells.

Expression of podoplanin by cancer cells of oral and uterine cervix squamous cell carcinoma has been reported to be associated with prognosis.4, 8 However, previous studies have found that adenocarcinoma cells rarely express podoplanin, consistent with the results of present study.7, 9, 23, 24 Podoplanin expression by cancer cells was found in only 9 of our 177 cases, and investigation of the relation between podoplanin expression by cancer cells and clinicopathological factors showed that podoplanin expression by cancer cells was present in a significantly higher proportion of the well differentiation cases than moderately/poorly differentiation cases (p = 0.035). Moreover, podoplanin was expressed by cancer cells in significantly more cases that were in negative for pleural invasion cases than in cases that were positive for pleural invasion cases (p = 0.029). The cases in which cancer cells were positive for podoplanin expression did not have a shorter overall survival time than the case that were negative (p = 0.472). Because cancer cells expressed podoplanin in only a small number of cases, it is impossible to draw any definite conclusion about the biological function of podoplanin in adenocarcinoma cells, and accumulation of more cases that are positive for podoplanin expression by cancer cells cases will be necessary to evaluate the biological function of podoplanin in adenocarcinoma cells. Alternatively, the role of podoplanin expression by adenocarcinoma and squamous cell carcinoma cells may be different.

Since no CAFs expressing podoplanin were found in the non-invasive adenocarcinoma cases or in fibroblasts in normal lung tissue in this study, and Noguchi et al. reported finding an association between the presence of active fibroblasts and invasive behavior by adenocarcinoma of the lung,25 podoplanin may be a biological marker of active fibroblasts. The fact that the fibroblasts in non-cancerous lung tissue did not express podoplanin suggests that podoplanin expression may be upregulated during the crosstalk between cancer cells and CAFs. Alternatively, podoplanin-expressing fibroblasts progenitors may be recruited from remote sites. Identifying the mechanism that regulates podoplanin expression by stromal fibroblasts may be important to understand the biological behavior of CAFs.

Although podoplanin expression by CAFs was significantly associated with a poor outcome, a multivariate analysis did not show that it was an independent prognostic factor (p = 0.091, data not shown), perhaps because of the strong relations between podoplanin expression by CAFs and both pathological nodal involvement and vascular invasion, which are robust independent prognostic factors in adenocarcinoma of the lung.

Targeting the tumor microenvironment should increase the effectiveness of cancer therapy. The biological function of podoplanin expression by CAFs is poorly understood and largely unknown. Identification of the function and mechanism of expression of podoplanin by CAFs may enable podoplanin or the podoplanin receptor to be used as a molecular target.

Acknowledgements

The technical support of Ms. Hiroko Hashimoto, Ms. Mai Okumoto, Ms. Naho Atsumi and Dr. Manabu Yamazaki is gratefully acknowledged.