INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Inflammation-induced fibrosis is considered a risk factor for carcinogenesis in multiple organs wherever etiologies exist, e.g., in diseases such as idiopathic pulmonary fibrosis, ulcerative colitis, and liver cirrhosis (1). Pneumoconiosis is an occupational lung disease characterized by irreversible fibrous lesions (silicotic nodules or mixed dust fibrosis) in the lung resulting from chronic inhalation of silica and silicates in dusty workplaces (4, 5). Since the first report by Harding and McLaughlin (6), it has been more widely described that about 10 to 20% of pneumoconioses are inclined to entail diffuse interstitial fibrosis (DIF) similar to interstitial pneumonia, both radiographically and pathologically, in addition to typical pneumoconiotic lesions (7). However, little is known about the characteristics of DIF in pneumoconiosis and its contribution to the risk of lung cancer. We have previously demonstrated that pneumoconiosis-related squamous cell carcinomas (SCCs) of the lung occur predominantly in peripheral epithelia, with concomitant frequent p53 expression in bronchiolar dysplasias from pneumoconiotic patients (11).

Assuming further that fibrogenesis and carcinogenesis are both prevalent over peripheral sites in pneumoconiosis (12), we were prompted to undertake subsequent clinicopathologic studies to explore the association between pneumoconiosis- associated DIF and lung cancer, especially peripheral-type SCC. Furthermore, p53 expression was investigated by immunohistochemistry (IHC) in lung cancers and dysplasias from patients with pneumoconiosis with and without DIF to determine whether p53 expression may be involved in the development of lung cancer in DIF-type pneumoconiosis.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

We investigated clinicopathologic characteristics of patients with nonasbestos pneumoconiosis who had been followed at Iwamizawa Rosai Hospital and who died during the period from January 1985 through December 1996. Pneumoconiosis was diagnosed and categorized for patients with a definite occupational history, according to the ILO International Classification of Radiographs of Pneumoconioses (13): Categories 1 to 3 stand for simple pneumoconiosis with increased profusion of small opacities less than 10 mm in diameter. Category 4 represents complicated pneumoconiosis with massive fibrosis more than 10 mm in diameter. Pneumoconiotic patients have been carefully followed by regular medical examinations in our hospital. Namely, chest CT was carried out at least every 6 mo. Other routine checkups, including chest radiography and sputum cytology, were performed at least every 3 mo. Accordingly, pneumoconiotic patients missing the above regular examinations for more than 3 mo prior to the date of death were considered ineligible for the analysis on the eventual concurrence of lung cancer or DIF, and they were excluded from the present study. On the basis of this criterion, a total of 563 patients with nonasbestos pneumoconiosis (404 coal miners, 105 metal miners, 18 tunnelers, and 36 other workers without asbestos exposure) were eligible for the current analysis during the above period. The existence of DIF and its grade were evaluated on the basis of chest radiography and chest CT: findings of "honeycomb lung" and/or reticulonodular shadows, predominantly in the bilateral middle to lower lung fields, were indicative of irreversible fibrosing lesions and judged to be DIF. DIF-positive patients were further classified into two grades: "limited DIF" represented the patients with DIF confined to middle to lower lobes, "extensive DIF" represented the patients with DIF extending to upper lobes. Histology of lung cancers was determined according to the World Health Organization criteria (14). A total of 247 patients with lung cancer without occupational exposure were collected from the Hokkaido University Medical Hospital and Iwamizawa Rosai Hospital and analyzed in relation to a total of 107 patients with lung cancer and pneumoconiosis in Iwamizawa Rosai Hospital in this study. Lung cancers, especially SCCs of the lung, were divided anatomically into the "central type" (with the origin proximal to segmental bronchi) and the "peripheral type" (with the origin distal to subsegmental bronchi) on the basis of clinicopathologic observations.

Bronchial and bronchiolar dysplasias were diagnosed, using authorized histologic criteria as described previously (15, 16). Briefly, bronchiolar dysplasias were considered atypical epithelial cells with or without squamous metaplasias in the small airways from membranous to respiratory bronchioles in the absence of inflammatory or erosive regenerative changes. The pathology of pneumoconiosis, lung cancers, and dysplasias was independently reviewed and agreed upon by three pathologists (autopsied specimens by K.H. and K.K.; surgically resected specimens by K.H. and K.O.) in the present study.

Immunohistochemistry

Of lung cancers from pneumoconiotic patients, autopsied and surgically resected tumors available were investigated for p53 expression by IHC. The patients with bronchial or bronchiolar dysplasia were subjected to IHC as well. Formalin-fixed, paraffin-embedded tissues of interest were serially sliced into 4-µm sections. After being deparaffinized in xylene and microwaved in 10 mM citrate buffer (pH, 6.0) for 10 min, sections were immunostained for p53 protein by the labeled streptavidin biotin (LSAB) method. Briefly, the slides were reacted with the primary anti-p53 monoclonal antibody, DO7 (Novocastra Laboratories, Ltd, Newcastle, UK; diluted ×50) for 2 h at room temperature. After incubation with the secondary antibody for 20 min, sections were treated with alkaline-phosphatase-conjugated streptavidin-biotin complex and visualized by colorimetric reaction using chromogen. The nuclei were counterstained with methyl green. The primary antibody was replaced by phosphate-buffered saline or control mouse immunoglobulin to monitor background staining. We used lung cancer sections previously known to have strong or no expression for p53 to serve as positive and negative controls, respectively (17). Samples were scored as positive for p53 expression when more than 10% of nuclei in each lesion had clear immunostaining, irrespective of cytoplasmic staining, and p53 expression in the specimens was referred to our previously published results (11), except that tumors from pneumoconiosis for p53 IHC were further extended to a total of 79 specimens in the present study.

Statistical Analysis

Statistical analyses were performed using the chi-square or Fisher's test, or the Mann-Whitney U test, as appropriate, to compare subgroups classified by various clinicopathologic and biologic settings, excluding a few patients, if any, for their lack of medical records and examinations in each setting.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinicopathologic Characteristics of Pneumoconiosis-related Lung Cancers

We analyzed 563 patients with nonasbestos pneumoconiosis and occurrence of lung cancer and DIF, who had been followed at Iwamizawa Rosai Hospital and eventually died during the period from 1985 through 1996 (Table 1). Throughout this period, a total of 107 (19%) of 563 patients were found to have developed lung cancers (Table 1A). The histology of the 107 pneumoconiosis-related lung cancers showed 60 SCCs (57%), 36 adenocarcinomas (33%), five small cell carcinomas (5%), two large cell carcinomas (2%), and four of unknown histology (Table 1B). SCCs arising from patients with pneumoconiosis had a significantly higher frequency of the peripheral type than SCCs from patients without pneumoconiosis (46 [78%] of 59 versus 44 [46%] of 96, p < 0.001). Moreover, peripheral-type SCCs accounted for 46 (43%) of the 107 pneumoconiosis-related lung cancers, which indicated an absolutely increased number of peripheral-type SCCs, as compared with those from patients without pneumoconiosis (46 [43%] of 107 versus 44 [18%] of 247, p < 0.001) (Table 1B). We next divided pneumoconiosis-related lung cancers into upper (including lingula), middle, and lower lobes with respect to the sites of origin, and analyzed them statistically in association with their histologic types. Peripheral-type SCCs developed significantly more often in lower lobes of the lung as compared with upper lobes than central-type SCCs (lower, 30; upper, 13 in peripheral SCCs versus lower, 3; upper, 9 in central SCCs, p = 0.01) (Table 1C).

Clinicopathologic Characteristics of Pneumoconiosis with DIF, with Special Reference to the Concurrence of Lung Cancer

DIF was evaluated clinically by both chest radiography and chest CT, as described in METHODS. Fifty-five (10%) of 563 pneumoconioses possessed DIF. Patients who were DIF-positive were all confirmed to simultaneously harbor upper-zone predominant profusion of small opacities or massive fibrosis in chest radiography and CT, which is consistent with the criteria of pneumoconiosis (see SUBJECTS in METHODS), and may be distinguishable from idiopathic pulmonary fibrosis and other interstitial lung diseases (data not shown). Of 55 patients with DIF-positive pneumoconiosis, autopsied specimens were available for 17, and it was confirmed that all 17 definitely had the characteristics of DIF not only on chest radiography and CT but also on pathologic examination, as previously shown (10, 12). Namely, lower-zone predominant diffuse fibrosis with honeycombing similar to that in interstitial pneumonia was also observed in pathologic examination for radiographically proven DIF-positive cases. The areas with DIF often harbored collections of dust-laden macrophages, mixed dust fibrous nodules, or silicotic nodules (Figure 1). Furthermore, asbestos bodies were carefully examined at more than 10 sites per case in hematoxylin-eosin stain and at upper and lower lobes per case in Berlin blue stain for the above 17 DIF-positive specimens to exclude the possibility of asbestosis. As a result, none of them was shown to possess asbestos bodies, again confirming that DIF-positive pneumoconioses were not confounded by asbestosis at all in the present study (data not shown). Pneumoconiosis with DIF had an exceedingly high occurrence of lung cancers, compared with pneumoconiosis without DIF (29 [53%] of 55 versus 78 [15%] of 508, p < 0.001) (Table 2A). Peripheral-type SCCs also arose significantly more often in DIF-positive cases than in DIF-negative ones (13 [24%] of 55 versus 33 [6%] of 508, p < 0.001) (Table 2A, 2B). SCCs arising from pneumoconiosis with DIF constituted an exceptionally high incidence of peripheral-type ones, as compared with those from pneumoconiosis without DIF (13 [100%] of 13 versus 33 [72%] of 46, p = 0.03) (Table 2B). This significance also applied to peripheral-type pneumoconiosis-related lung cancers covering all histologic types (27 [96%] of 28 versus 60 [80%] of 75, p = 0.03). Moreover, lung cancers from pneumoconiosis with DIF proved to occur significantly more often in lower lobes as compared with upper lobes than lung cancers from pneumoconiosis without DIF (lower, 18; upper, 5 in pneumoconiosis with DIF versus lower, 36; upper, 38 in pneumoconiosis without DIF, p = 0.02) (Table 2B). Finally, lung cancers appeared to arise frequently from the area of DIF in pneumoconiosis with DIF (20 [74%] of 27). The incidence of lung cancer, however, was not associated with the grade of DIF (21 [60%] of 35 "limited DIF" versus eight [40%] of 20 "extensive DIF").

View larger version (101K): [in this window] [in a new window]   Figure 1.   (a-c ) The different areas with DIF from a DIF-positive pneumoconiosis, showing diffuse interstitial fibrosis with honeycombing and concomitant silicotic nodules composed of whorled hyalinized collagen (arrows). Hematoxylin-eosin stain. Original magnification: a and b, ×5; c, ×25.

Clinical Characteristics of Pneumoconiosis in Relation to Concurrence of Lung Cancer and DIF

Lung cancers frequently occurred with simple pneumoconiosis (Category 1 to 3) compared with complicated pneumoconiosis with massive fibrosis (Category 4), a difference that was highly significant (60 [33%] of 183 versus 45 [12%] of 375, p < 0.001) (Table 3). DIF also appeared significantly more often with simple pneumoconiosis than with complicated pneumoconiosis (29 [16%] of 183 versus 26 [7%] of 375, p = 0.002) (Table 3). Neither concurrent lung cancers nor DIF was associated with occupation (i.e., coal miners, metal miners, or others), duration of dust exposure (working years) or other clinical parameters of pneumoconiosis (Table 3).

The consumption of tobacco was investigated in terms of pack-years. Mean tobacco consumption was significantly greater in pneumoconiosis with lung cancer than in that without lung cancer (40.7 ± 19.1 versus 30.9 ± 23.4, p < 0.001) (Table 3). Pneumoconiosis with DIF also had a significantly greater pack-years of smoking than did its counterpart without DIF (40.0 ± 26.3 versus 32.0 ± 22.5, p = 0.02) (Table 3).

Anatomic Sites of Dysplasia from Pneumoconiosis with and without DIF

Autopsied specimens were collected from pneumoconiotic patients with and without DIF, and anatomic sites of dysplasia (dysplasias in bronchi or bronchioles) were examined in relation to DIF status, which was evaluated by both chest radiography and chest CT, as described in METHODS. We chose the autopsied lung specimens prepared with more than eight hematoxylin-eosin staining sites covering all bilateral lobes over proximal and peripheral airways per case to ensure unbiased comparison. Interestingly, dysplasias from pneumoconiosis with DIF were significantly more frequently observed in peripheral bronchioli than in dysplasias from pneumoconiosis without DIF (11 [69%] of 16 versus 20 [30%] of 66, p = 0.01), but not in proximal bronchi (2 [13%] of 16 versus 19 [29%] of 66).

p53 Expression in Lung Cancers and Dysplasias from Pneumoconiosis with DIF

Of 18 lung cancers available for analysis from pneumoconiosis with DIF, p53 expression was observed in 10 tumors (56%), which was comparable to the rate of p53 expression in lung cancers associated with pneumoconiosis without DIF (37 [61%] of 61) (Table 4 and Figure 2). This frequency was not significantly different from that of lung cancers without pneumoconiosis (86 [57%] of 151). p53 expression evaluated by IHC was frequently observed in dysplasias from pneumoconiosis with DIF, although its frequency was not significantly different from that in dysplasias from pneumoconiosis without DIF (5 [50%] of 10 versus 12 [38%] of 32) and that in dysplasias from nonpneumoconiosis (10 [23%] of 44) (Table 4).

View larger version (108K): [in this window] [in a new window]   Figure 2.   (a) p53 expression in lung cancers, showing clear positive nuclear immunostaining in SCC and (b) adenocarcinoma, both from patients with DIF-type pneumoconiosis. (c ) SCC from a patient with DIF-type pneumoconiosis without p53 expression. LSAB method with methyl green counterstaining. Original magnification: ×200.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study, we demonstrated an increased incidence of peripheral-type SCCs of the lung associated with pneumoconiosis. It was also shown that peripheral-type SCCs arose preferentially from lower lobes of the lung in pneumoconiosis. Surprisingly, we found that cases of pneumoconiosis with DIF had lung cancers much more frequently than did those from pneumoconiosis without DIF. Moreover, SCCs from pneumoconiosis with DIF appeared to be composed of all peripheral types and preferentially arise from the lower lobes in the area of DIF. These results indicate that DIF-type pneumoconiosis is highly predisposed to lung cancer, especially peripheral-type SCCs of the lung.

Furthermore, in this study, dysplasias were observed predominantly in bronchiolar peripheral epithelia rather than in proximal bronchi in pneumoconiosis with DIF. Bronchiolar dysplasias were more frequently observed in pneumoconiosis with DIF than in its DIF-negative counterpart. We also observed that p53 expression was frequently manifest not only in bronchiolar dysplasias but also in lung cancers arising from pneumoconiosis with DIF. These pathologic and biologic findings may further support the etiologic linkage of pneumoconiosis-associated DIF, bronchiolar dysplasia, and lung cancer, particularly peripheral-type SCC.

Lung cancers were found to arise significantly more often from simple pneumoconiosis than from complicated pneumoconiosis with massive fibrosis, as was shown in the previous studies (5, 18). There is no evidence thus far to show that radiographic severity of pneumoconiosis correlates with the degree of pulmonary epithelial damage related to dysplastic changes. Conversely, it is possible to assume that massive fibrosis may cause involved epithelia to deplete rather than be damaged persistently, thereby diminishing the number of potential precursor lesions (5, 19), leading to the reduced incidence of lung cancer in complicated pneumoconiosis. In addition, the duration of working years neither correlated with the incidence of lung cancer nor DIF in pneumoconiosis in the present study, although cumulative silica exposure was not yet measured quantitatively. Therefore, a clear-cut dose-response relationship, at least, is unlikely to apply to the level of silica exposure and incidence of lung cancer. This may be in line with the previous reports arguing against a direct carcinogenic effect of silica (20, 21). Taking these data together, it is plausible to speculate that lung cancers might arise preferentially from certain subsets of pneumoconiosis that are highly inclined to generate diffuse fibrosis with concomitant precancerous dysplasias and further able to retain their precursor lesions long enough for them to progress to lung cancer.

In this study, pneumoconiosis with lung cancer or DIF was associated with a significantly increased amount of cigarette consumption. This finding does not seem surprising because it is generally accepted that smoking independently plays a significant role in the development of lung cancer among nonpneumoconiotic subjects (22, 23). Smoking may interact, if not synergistically, with carcinogenic particles to develop lung cancer in the asbestos- or chromate-exposed lung (24). The influence of smoking, however, remains controversial in pneumoconiosis-related lung cancer (27). Thus, it awaits further investigation whether smoking exerts a synergistic or additive effect on the development of lung cancer and/or DIF associated with pneumoconiosis.

p53 is a well-characterized tumor suppressor gene that is deeply involved in a broad array of human malignancies, including lung cancer (28). p53 expression detected by IHC is considered to be based on mutation of the p53 gene, since mutated p53 protein is stabilized to acquire a prolonged half-life to an extent sufficiently detectable by conventional IHC. More recently, wild-type p53 was found to become stabilized under various circumstances such as gene repair, apoptosis, and hypoxia, which may tend to indicate that IHC for p53 protein is a poorer surrogate than expected earlier for screening of p53 mutations (29). Nonetheless, p53 expression evaluated by IHC was fairly concordant with its genetic mutations in lung cancers, ranging from 62 ~ 90% as previously demonstrated (32). We scored samples positive for p53 expression only when more than 10% of nuclei in the lesions showed p53 immunostaining. This criterion may be reliable to exclude nonspecific or irrelevant wild-type p53 expression in lung cancer and to further distinguish precancerous atypias from benign reactive atypias of the lung (35). p53 expression was observed frequently in lung cancer and dysplasia from pneumoconiosis with DIF, though the frequency was not statistically significant compared with the DIF-negative counterpart. The frequent p53 expression in dysplasia as well as in lung cancer from pneumoconiosis with DIF may be compatible with our current clinicopathologic observations that pneumoconiosis with DIF coincided with excessive bronchiolar dysplasia and lung cancer. We have previously demonstrated that pneumoconiosis-related SCCs occur predominantly in peripheral epithelia, with concomitant frequent p53 expression in bronchiolar dysplasias from pneumoconiotic patients, which implies that precancerous and/or cancerous targets in pneumoconiosis may be distributed over a more peripheral zone of the bronchial tree than in nonpneumoconiosis (11). Nevertheless, other additional target molecules and/or mechanisms must underlie the extremely high incidence of lung cancers in the subset of pneumoconiosis with DIF since frequent p53 expression alone cannot explain this particular clinicopathologic observation, as revealed in this study. Further investigation in this respect may also be helpful to clarify the molecular mechanisms of the fibrogenesis-attributable onset of lung cancer in other pulmonary diseases such as asbestosis and idiopathic pulmonary fibrosis.

In summary, we have demonstrated that pneumoconiosis-associated DIF may be one of the important accelerators toward lung cancer development, especially for peripheral-type SCCs. It should be noted that patients with DIF-type pneumoconiosis may be strongly predisposed to develop lung cancer and therefore must be carefully followed for early detection, if any, of lung cancer. The present study has suggested a positive causal relationship between pneumoconiosis and peripheral-type SCCs of the lung, and further has indicated a pivotal role of diffuse fibrosis in the excess incidence of lung cancer, especially of the peripheral-type SCC, associated with DIF-type pneumoconiosis.