ANATOMIC PATHOLOGY Original Article
Peripheral Airway Cell Marker Expression in Non-Small Cell Lung Carcinoma Association with Distinct Clinicopathologic Features R. ILONA LINNOILA, M.D.,1 SANDRA M. JENSEN, M.A., C.T.,1 SETH M. STEINBERG, PH.D., 2 JAMES L. MULSHINE, M.D.,1 JOSEPH C. EGGLESTON, M.D.,3* AND ADI F. GAZDAR, M.D."
For reasons not fully understood, during the past 20 years there has been a relative and absolute increase in the incidence of adenocarcinomas in the United States.4"6 Pulmonary adenocarcinoma is a heterogenous group of tumors that, according to the World Health Organization, can be divided further into four subtypes identified by their growth patterns, including acinar (gland forming), solid with mucin (poorly differentiated), bronchioloalveolar, and papillary.7 In our experience, bronchioloalveolar and papillary carcinomas constitute about one half of all pulmonary adenocarcinomas. 8 Bronchioloalveolar From the'National Cancer Institute-Navy Medical Oncology Branch, tumors, especially in an invasive or metastatic mode, Naval Hospital, Bethesda, 2Biostatistics and Data Management Section. cannot always be differentiated from papillary carcinoNational Cancer Institute, National Institutes of Health, Bethesda, and mas.9 The cellular origins of both of these subtypes overlap the * Department of Pathology, The Johns Hopkins Hospital, Baltimore, Maryland. and include mucin-secreting cells, Clara cells, and Type II pneumocytes.9"12 For these reasons, we prefer to regard Received March 7, 1991; received revised manuscript and accepted for publication June 25, 1991. bronchioloalveolar and papillary tumors as a single entity The opinions or assertions contained herein represent the private views and refer to them as papillolepidic tumors. In a previous of the authors and are not to be construed as official or as reflecting the study, we demonstrated that cell lines derived from papviews of the Department of the Navy or the Department of Defense. Address reprint requests to Dr. Linnoila: National Cancer Instituteillolepidic tumors frequently express ultrastructural, bioNavy Medical Oncology Branch, Naval Hospital Building 8, Room 5101, chemical, and molecular markers characteristic of peBethesda, Maryland 20889-5105. ripheral airway cells (PAC).'' * Deceased. Lung cancer in the United States is the leading cause of cancer deaths in both men and women. The incidence is increasing, with more than 160,000 new cases diagnosed every year, despite the fact that the national per capita consumption of cigarette tobacco, the main known risk factor of lung cancer, has decreased during the last 30 years.1"3 The major histologic types are small cell lung cancer (25%) and non-small cell lung cancer (NSCLC; 75%). The latter encompasses adenocarcinoma, squamous cell carcinoma, and large cell carcinoma subtypes.
233
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
organs and their tumors were negative, with few exceptions. Nonsmall cell lung carcinomas positive for peripheral airway cell markers were associated with younger age and less-intense smoking, and surfactant-associated protein reactivity was more common in women than in men. Peripheral airway cell markers were independent prognostic factors for survival and delayed development of metastases in patients with less-advanced disease. It is concluded that surfactant-associated protein and 10-kD Clara cell protein are specific markers for non-small cell lung carcinoma and peripheral airway cell differentiation and provide useful tools to study the pathogenesis, biology, and prognosis of non-small cell lung carcinoma. (Key words: Lung cancer; Adenocarcinoma; Papillolepidic; Clara cell; Type II cell; Surfactant; Smoking) Am J Clin Pathol 1992;97:233-243
Paraffin sections of 247 primary and metastatic non-small cell lung carcinomas, the corresponding non-neoplastic lungs, and 75 other specimens were examined by immunohistochemical procedures using a panel of antibodies against the specific products of peripheral airway cells: the major surfactant-associated protein and 10-kD Clara cell protein. Non-small cell lung carcinoma tumors most frequently positive for either peripheral airway cell marker were adenocarcinomas (41%), especially those with papillolepidic growth pattern (56%), followed by large cell carcinomas (25%), other adenocarcinomas (22%), and squamous cell carcinomas (16%). Immunoreactivity was mainly focal and the expression of the two peripheral airway cell markers was discordant. The incidence of marker expression was similar in metastatic and primary non-small cell lung carcinoma. Other
234
ANATOMIC PATHOLOGY Original Article
MATERIALS AND METHODS
ments of Pathology at the NCI, Bethesda Naval Hospital, and Johns Hopkins Hospital. Only well-fixed, surgically removed material with ample amount of tumor to obtain multiple serial sections for immunohistochemical and histochemical studies was included. Five-micron-thick sections from routine formalin-fixed, paraffin-embedded blocks were cut and mounted on gelatin-coated slides. Hematoxylin-and-eosin-stained sections of each tumor were reviewed independently by three pathologists (R.I.L., J.C.E., and A.F.G.) without previous knowledge of the clinical data or immunohistochemical results. Consensus opinion was reached and tumors were classified according to the World Health Organization criteria.7 The various histologic types of lung cancers are shown in Table 1. Antibodies and Immunohistochemical Histochemical Staining
and
Two rabbit antibodies against the major human surfactant-associated protein SP-A were used: (1) a rabbit antibody (SAM) raised against a 36,000 molecular-weight glycoprotein isolated from human amniotic fluid29 was a gift from Dr. S. N. Bhattacharyya, William Beaumont Army Medical Center, El Paso, Texas; and (2) a rabbit antibody (SAP-35), a gift from Dr. J. A. Whitsett, University of Cincinnati, Cincinnati, Ohio, was generated against purified human alveolar proteinosis protein, as described elsewhere.30 A rabbit antibody (CLR-10) specific for a 10,000 molecular-weight human Clara cell protein, a gift from Dr. G. Singh, University of Pittsburgh and the VA Medical Center, Pittsburgh, Pennsylvania, was raised against the postalbumin fraction of the soluble proteins from human bronchioloalveolar lavage.17
Case Material For the following reasons we performed the study on three groups of cases: (1) to study clinicopathologic correlations, we included 122 consecutive patients in whom material and detailed clinical information was available that were entered onto a therapeutic protocol of all stages of NSCLC at the National Cancer Institute (NCI) dating from March 1984 to January 1989; (2) to confirm the incidence of tumors with PAC differentiation in an independent set of patients from another institution, 102 consecutive NSCLC patients with lung resection at Johns Hopkins Hospital, Baltimore, Maryland, dating from 1984 to 1987, were included; (3) finally, 23 representative NSCLC cases from surgical pathology files were chosen to establish optimal staining conditions and to be used as controls. A total of 247 primary and metastatic NSCLC were obtained from the surgical pathology files of the Depart-
TABLE 1. WHO HISTOLOGIC CLASSIFICATION OF THE NON-SMALL CELL LUNG CANCERS USED IN THE CURRENT STUDY Number of Cases
Adenocarcinoma Papillolepidicf Acinar/solidJ Squamous cell carcinoma Large cell carcinoma Other Total
(%)
Johns Hopkins Hospital
NCI
All Cases*
58 (56) 39 (38) 19(18) 29 (28) 13(13) 3(3)
75(61) 36 (30) 39 (32) 14(11) 23(19) 10(8)
154(62) 92 (37) 62 (25) 43(17) 36(15) 14(6)
103(100)
122(100)
247(100)
* A l l cases including the 23 staining controls. t Papillolepidic category includes bronchioloalveolar carcinomas, adenocarcinomas with marked bronchioloalveolar component, and papillary adenocarcinomas. $ Acinar/solid category includes acinar adenocarcinomas and solid adenocarcinomas with mucin formation. N C I = National Cancer Institute.
A.J.C.P. • February 1992
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
Clara cells and Type II pneumocytes are the progenitor cells of PACs.13 Clara cells, located in the bronchiolar epithelium, are metabolically active in mixed-function oxidation and may participate in water and salt transportation. 1415 They are characterized by dense granules in the apical cytoplasm and produce a specific 10-kD protein (CLR-10) whose function is unknown.16"18 Antibodies to CLR-10 stain Clara cells in humans, dogs, and rats and provide a tool to study the pathobiology of these cells in humans and in experimental models. 1719 Type II pneumocytes are located in the alveoli, composing 16% of the parenchymal cells in the human lung.20 They produce surfactant, which exerts a detergent-like action essential for maintaining the patency of the alveoli.21 By ultrastructural examination, Type II cells are characterized by multilamellar bodies that represent the storage site of surfactant.22 Although phospholipids comprise the major component of surfactant, several functionally important surfactant-associated proteins have been identified. The major protein is called SP-A (previously referred to as SAP-35), which is a molecular-weight 28,000-36,000 glycoprotein, also present in alveolar lavage.23 Specific antibodies against SP-A are available.24"30 In this report, we examined the specificity and patterns of expression of the PAC markers SP-A and CLR-10 in a large panel of NSCLC as well as other tissues, and their potential clinicopathologic correlates and applications. In particular we were interested in the associations of PAC differentiation markers with histologic type and stage of tumor, patient's age, sex, smoking history, as well as survival.
LINNOILA ET AL. Peripheral Airway Cell Marker Expresion in Non-Small Cell Lung Carcinoma
Clinical Information and Statistical
Methods
Clinicopathologic correlation was performed using the information on 122 consecutive patients whose specimens were available for analysis and who were entered to a prospective clinical trial at NCI, described in detail elsewhere 32. Briefly, patients with any stage or histologic results of pathologically confirmed NSCLC with an Eastern Cooperative Oncology Group performance status of 3 or better were entered. For the purpose of analysis, patients were placed in one of two treatment groups: potentially curable (60 patients) or suitable only for palliative (62 patients) therapy.32 Curable group patients had tumor confined to the thorax, including Stages I, II, and IIIA patients, excluding what is designated as Stage IIIB.33 The curable group patients underwent either surgical resection (with or without postoperative involved-field radiotherapy) or definitive radiotherapy using standard indica-
tions.34 On relapse, any patient with good performance status and disease not requiring palliative radiotherapy was offered chemotherapy. Patients presenting with regionally advanced intrathoracic disease or supraclavicular nodes (Stage IIIB) and patients with distant metastases (Stage IV) were considered the palliative group. Patients with good performance status and measurable or evaluable disease not requiring palliative irradiation were offered immediate chemotherapy. Patients received regular follow-up examinations and were restaged at relapse. Standard chemotherapy response criteria, applied to both measurable and evaluable disease, were applied as previously described.35 We performed univariate analyses for the relationship between staining characteristics or survival and the variables of age (grouped 0-49, 50-59, 60-69, and 70 and older), sex, performance status, initial serum lactic dehydrogenase (LDH, divided in quartiles), smoking history (pack-years of cigarette use grouped 0, 1-29, 30-39, 4049, 50-69, 70-99, and 100 and over, or 0-49 and 50 and over, as indicated), presence of either supraclavicular nodes or distant metastases, stage, treatment group (curable or palliative), previous treatment with either radiation or chemotherapy, and histologic findings (Table 1). These analyses were performed using Fisher's exact test for 2 X 2 contingency tables and the Mantel test for trend for contingency tables.36 Survival times were calculated from the date of protocol entry to the date of death or last known date alive. Death from any cause was treated as the main outcome event in all such analyses. Survival probabilities were calculated and displayed using the method of Kaplan and Meier37 and the curves were compared statistically to one another using the Mantel-Haenszel test.38 Cox proportional hazards modeling39 was used to evaluate the importance of several factors simultaneously to predict survival. The resulting model parameters (bi) were converted to relative risks by computing exp(b;) where exp(a) = 2.71828a.40 The 95% confidence interval for the relative risk was computed as [exp(biL), exp (biH)] where b iL = b; -1.96 [estimated standard error (b;)] and b iH = bj + 1.96 [estimated standard error(bj)]. The relative risk indicates the risk associated with dying while being in a greater risk category compared with that of being in a lower risk category. All probability values in this report correspond to two-sided significance tests. RESULTS Non-Neoplastic Lung and Other Organs General staining patterns with both antibodies against SP-A were similar in non-neoplastic lungs, revealing a
Vol. 97 No. 2
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
Immunohistochemical staining was performed by the avidin-biotinylated peroxidase-complex technique using Vectastain avidin-biotinylated peroxidase-complex staining kits (Vector Laboratories, Burlingame, CA) according to vendor's instructions and modified as previously reported.31 Incubations with primary antisera were performed at 4 °C for 18 hours. Each primary antibody was titered specifically for the present study using paraffin sections of the control tumors. Furthermore, the dilution was considered optimal when the antibodies SAM and SAP-35 produced strong staining in Type II pneumocytes and the CLR-10 antibody stained intensely Clara cells in the bronchioles of non-neoplastic lung, whereas other pulmonary cells remained unstained. Positive control sections for each antibody from appropriate tumors were included in each assay. Controls for specificity of staining consisted of the incubation of serial lung tumor sections with (1) nonimmune (normal) rabbit serum or (2) phosphate-buffered saline. Results of the immunostaining and mucicarmine histochemical analysis were reviewed by two of the authors (R.I.L. and S.M.J.) who independently scored both for the intensity of the staining (0 = negative; 1 = weak; 2 = moderate; 3 = strong reaction) and the number of positive cells (distribution score: 0 = no positive cells; 1 for 10% of tumor cells positive). However, the presence of rare, scattered cells immunoreactive for SPA or CLR-10 was a common finding in many more tumors.
Vol. 97 • No. 2
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
Initial studies indicated that with SAP-35 antibody there was a variable amount of nonspecific stromal connective tissue reactivity. Furthermore, the tumor-vmi«-background resolution with this antibody was frequently less optimal than that achieved using antibody SAM. When we compared the staining patterns of various histologic types of lung tumors with the two SP-A antibodies, a higher number of nonadenocarcinomas were positive after staining with SAP-35 antibody than with SAM antibody (Fig. 3). Consequently, in further screening studies of lung tumors, we chose to use antibody SAM to detect SP-A immunoreactivity. The immunoreactivity for SP-A and CLR-10 in tumors was mostly focal and often less intense than in non-neoplastic reactive cells in the same sections that stained positively for these markers. A tumor was scored positive if more than 10% tumor cells demonstrated immunoreactivity for any given marker. Immunostaining was cytoplasmic and the nuclei remained negative, except for nucleolar immunoreactivity, which was seen in 19 of 48 (40%) of SP-A and nuclear immunoprecipitate in 6 of 38
238
ANATOMIC PATHOLOGY Original Article
" -• r':-
'''MM^'^'^L'
V-*
*"
^
*'*« 2 = 0.10)
21/79(27) 9/34 (26) (^2= 1.0)
7/41(17) 1/40(3) (P2 = 0.06) 6/25 (24) 3/14(21) (^2 = 1.0)
13/41 (31) 8/38(21) (P2 = 0.32) 9/20 (45) 0/14(0) (P2 = 0.004)
Data
TABLE 4. SUMMARY OF BIOGRAPHIC DATA SP-A and/or CLR-10 Immunoreactivity in Tumors
Age (years) 0-49 50-59 60-69 70+
Data
Positive*
Negative
Sex Male Female
Number of tumors Median age, years Male sex, percentage Smoking history, pack-years Curable patients, percentaget
44 (36%) 54 61 38 43
78 (64%) 58 71 56 53
Pack Years Male 0-49 (light) 50+ (heavy)
• £10% tumor cells positive for SP-A and/or CLR-10 by immunohistochemistry. t Patients receiving therapy with curative intent who had tumor confined to the thorax, including Stage 1, II, and I1A patients who underwent either surgical resection or definitive radiotherapy using standard indications.
Female 0-49 (light) 50+ (heavy)
* Mantel test for trend. For all other statistical analyses Fisher's exact test was used.
A.J.C.P. • February 1992
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
NCI = National Cancer Institute. * Positive = ^ 10% of tumor cells stain, t Staining data using antibody SAM. } SP-A/CLR-10 = SP-A and/or CLR-10-positive tumors. § Papillolepidic catefory includes bronchioloalveolar carcinomas, adenocarcinomas with marked bronchioloalveolar component, and papillary adenocarcinomas.
LINNOILA ET AL. Peripheral Airway Cell Marker Expression in Non-Small Cell Lung Carcinoma
241
TABLE 6. COX PROPORTIONAL HAZARDS MODEL Variable
P*
-0.72 0.53 0.67 -0.60 0.42 2.37
0.0048 0.0399 0.0253 0.0375 0.0063
Peripheral Airway Cell Marker Expression in Non-Small Cell Lung Carcinoma Association with Distinct Clinicopathologic Features R. ILONA LINNOILA, M.D.,1 SANDRA M. JENSEN, M.A., C.T.,1 SETH M. STEINBERG, PH.D., 2 JAMES L. MULSHINE, M.D.,1 JOSEPH C. EGGLESTON, M.D.,3* AND ADI F. GAZDAR, M.D."
For reasons not fully understood, during the past 20 years there has been a relative and absolute increase in the incidence of adenocarcinomas in the United States.4"6 Pulmonary adenocarcinoma is a heterogenous group of tumors that, according to the World Health Organization, can be divided further into four subtypes identified by their growth patterns, including acinar (gland forming), solid with mucin (poorly differentiated), bronchioloalveolar, and papillary.7 In our experience, bronchioloalveolar and papillary carcinomas constitute about one half of all pulmonary adenocarcinomas. 8 Bronchioloalveolar From the'National Cancer Institute-Navy Medical Oncology Branch, tumors, especially in an invasive or metastatic mode, Naval Hospital, Bethesda, 2Biostatistics and Data Management Section. cannot always be differentiated from papillary carcinoNational Cancer Institute, National Institutes of Health, Bethesda, and mas.9 The cellular origins of both of these subtypes overlap the * Department of Pathology, The Johns Hopkins Hospital, Baltimore, Maryland. and include mucin-secreting cells, Clara cells, and Type II pneumocytes.9"12 For these reasons, we prefer to regard Received March 7, 1991; received revised manuscript and accepted for publication June 25, 1991. bronchioloalveolar and papillary tumors as a single entity The opinions or assertions contained herein represent the private views and refer to them as papillolepidic tumors. In a previous of the authors and are not to be construed as official or as reflecting the study, we demonstrated that cell lines derived from papviews of the Department of the Navy or the Department of Defense. Address reprint requests to Dr. Linnoila: National Cancer Instituteillolepidic tumors frequently express ultrastructural, bioNavy Medical Oncology Branch, Naval Hospital Building 8, Room 5101, chemical, and molecular markers characteristic of peBethesda, Maryland 20889-5105. ripheral airway cells (PAC).'' * Deceased. Lung cancer in the United States is the leading cause of cancer deaths in both men and women. The incidence is increasing, with more than 160,000 new cases diagnosed every year, despite the fact that the national per capita consumption of cigarette tobacco, the main known risk factor of lung cancer, has decreased during the last 30 years.1"3 The major histologic types are small cell lung cancer (25%) and non-small cell lung cancer (NSCLC; 75%). The latter encompasses adenocarcinoma, squamous cell carcinoma, and large cell carcinoma subtypes.
233
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
organs and their tumors were negative, with few exceptions. Nonsmall cell lung carcinomas positive for peripheral airway cell markers were associated with younger age and less-intense smoking, and surfactant-associated protein reactivity was more common in women than in men. Peripheral airway cell markers were independent prognostic factors for survival and delayed development of metastases in patients with less-advanced disease. It is concluded that surfactant-associated protein and 10-kD Clara cell protein are specific markers for non-small cell lung carcinoma and peripheral airway cell differentiation and provide useful tools to study the pathogenesis, biology, and prognosis of non-small cell lung carcinoma. (Key words: Lung cancer; Adenocarcinoma; Papillolepidic; Clara cell; Type II cell; Surfactant; Smoking) Am J Clin Pathol 1992;97:233-243
Paraffin sections of 247 primary and metastatic non-small cell lung carcinomas, the corresponding non-neoplastic lungs, and 75 other specimens were examined by immunohistochemical procedures using a panel of antibodies against the specific products of peripheral airway cells: the major surfactant-associated protein and 10-kD Clara cell protein. Non-small cell lung carcinoma tumors most frequently positive for either peripheral airway cell marker were adenocarcinomas (41%), especially those with papillolepidic growth pattern (56%), followed by large cell carcinomas (25%), other adenocarcinomas (22%), and squamous cell carcinomas (16%). Immunoreactivity was mainly focal and the expression of the two peripheral airway cell markers was discordant. The incidence of marker expression was similar in metastatic and primary non-small cell lung carcinoma. Other
234
ANATOMIC PATHOLOGY Original Article
MATERIALS AND METHODS
ments of Pathology at the NCI, Bethesda Naval Hospital, and Johns Hopkins Hospital. Only well-fixed, surgically removed material with ample amount of tumor to obtain multiple serial sections for immunohistochemical and histochemical studies was included. Five-micron-thick sections from routine formalin-fixed, paraffin-embedded blocks were cut and mounted on gelatin-coated slides. Hematoxylin-and-eosin-stained sections of each tumor were reviewed independently by three pathologists (R.I.L., J.C.E., and A.F.G.) without previous knowledge of the clinical data or immunohistochemical results. Consensus opinion was reached and tumors were classified according to the World Health Organization criteria.7 The various histologic types of lung cancers are shown in Table 1. Antibodies and Immunohistochemical Histochemical Staining
and
Two rabbit antibodies against the major human surfactant-associated protein SP-A were used: (1) a rabbit antibody (SAM) raised against a 36,000 molecular-weight glycoprotein isolated from human amniotic fluid29 was a gift from Dr. S. N. Bhattacharyya, William Beaumont Army Medical Center, El Paso, Texas; and (2) a rabbit antibody (SAP-35), a gift from Dr. J. A. Whitsett, University of Cincinnati, Cincinnati, Ohio, was generated against purified human alveolar proteinosis protein, as described elsewhere.30 A rabbit antibody (CLR-10) specific for a 10,000 molecular-weight human Clara cell protein, a gift from Dr. G. Singh, University of Pittsburgh and the VA Medical Center, Pittsburgh, Pennsylvania, was raised against the postalbumin fraction of the soluble proteins from human bronchioloalveolar lavage.17
Case Material For the following reasons we performed the study on three groups of cases: (1) to study clinicopathologic correlations, we included 122 consecutive patients in whom material and detailed clinical information was available that were entered onto a therapeutic protocol of all stages of NSCLC at the National Cancer Institute (NCI) dating from March 1984 to January 1989; (2) to confirm the incidence of tumors with PAC differentiation in an independent set of patients from another institution, 102 consecutive NSCLC patients with lung resection at Johns Hopkins Hospital, Baltimore, Maryland, dating from 1984 to 1987, were included; (3) finally, 23 representative NSCLC cases from surgical pathology files were chosen to establish optimal staining conditions and to be used as controls. A total of 247 primary and metastatic NSCLC were obtained from the surgical pathology files of the Depart-
TABLE 1. WHO HISTOLOGIC CLASSIFICATION OF THE NON-SMALL CELL LUNG CANCERS USED IN THE CURRENT STUDY Number of Cases
Adenocarcinoma Papillolepidicf Acinar/solidJ Squamous cell carcinoma Large cell carcinoma Other Total
(%)
Johns Hopkins Hospital
NCI
All Cases*
58 (56) 39 (38) 19(18) 29 (28) 13(13) 3(3)
75(61) 36 (30) 39 (32) 14(11) 23(19) 10(8)
154(62) 92 (37) 62 (25) 43(17) 36(15) 14(6)
103(100)
122(100)
247(100)
* A l l cases including the 23 staining controls. t Papillolepidic category includes bronchioloalveolar carcinomas, adenocarcinomas with marked bronchioloalveolar component, and papillary adenocarcinomas. $ Acinar/solid category includes acinar adenocarcinomas and solid adenocarcinomas with mucin formation. N C I = National Cancer Institute.
A.J.C.P. • February 1992
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
Clara cells and Type II pneumocytes are the progenitor cells of PACs.13 Clara cells, located in the bronchiolar epithelium, are metabolically active in mixed-function oxidation and may participate in water and salt transportation. 1415 They are characterized by dense granules in the apical cytoplasm and produce a specific 10-kD protein (CLR-10) whose function is unknown.16"18 Antibodies to CLR-10 stain Clara cells in humans, dogs, and rats and provide a tool to study the pathobiology of these cells in humans and in experimental models. 1719 Type II pneumocytes are located in the alveoli, composing 16% of the parenchymal cells in the human lung.20 They produce surfactant, which exerts a detergent-like action essential for maintaining the patency of the alveoli.21 By ultrastructural examination, Type II cells are characterized by multilamellar bodies that represent the storage site of surfactant.22 Although phospholipids comprise the major component of surfactant, several functionally important surfactant-associated proteins have been identified. The major protein is called SP-A (previously referred to as SAP-35), which is a molecular-weight 28,000-36,000 glycoprotein, also present in alveolar lavage.23 Specific antibodies against SP-A are available.24"30 In this report, we examined the specificity and patterns of expression of the PAC markers SP-A and CLR-10 in a large panel of NSCLC as well as other tissues, and their potential clinicopathologic correlates and applications. In particular we were interested in the associations of PAC differentiation markers with histologic type and stage of tumor, patient's age, sex, smoking history, as well as survival.
LINNOILA ET AL. Peripheral Airway Cell Marker Expresion in Non-Small Cell Lung Carcinoma
Clinical Information and Statistical
Methods
Clinicopathologic correlation was performed using the information on 122 consecutive patients whose specimens were available for analysis and who were entered to a prospective clinical trial at NCI, described in detail elsewhere 32. Briefly, patients with any stage or histologic results of pathologically confirmed NSCLC with an Eastern Cooperative Oncology Group performance status of 3 or better were entered. For the purpose of analysis, patients were placed in one of two treatment groups: potentially curable (60 patients) or suitable only for palliative (62 patients) therapy.32 Curable group patients had tumor confined to the thorax, including Stages I, II, and IIIA patients, excluding what is designated as Stage IIIB.33 The curable group patients underwent either surgical resection (with or without postoperative involved-field radiotherapy) or definitive radiotherapy using standard indica-
tions.34 On relapse, any patient with good performance status and disease not requiring palliative radiotherapy was offered chemotherapy. Patients presenting with regionally advanced intrathoracic disease or supraclavicular nodes (Stage IIIB) and patients with distant metastases (Stage IV) were considered the palliative group. Patients with good performance status and measurable or evaluable disease not requiring palliative irradiation were offered immediate chemotherapy. Patients received regular follow-up examinations and were restaged at relapse. Standard chemotherapy response criteria, applied to both measurable and evaluable disease, were applied as previously described.35 We performed univariate analyses for the relationship between staining characteristics or survival and the variables of age (grouped 0-49, 50-59, 60-69, and 70 and older), sex, performance status, initial serum lactic dehydrogenase (LDH, divided in quartiles), smoking history (pack-years of cigarette use grouped 0, 1-29, 30-39, 4049, 50-69, 70-99, and 100 and over, or 0-49 and 50 and over, as indicated), presence of either supraclavicular nodes or distant metastases, stage, treatment group (curable or palliative), previous treatment with either radiation or chemotherapy, and histologic findings (Table 1). These analyses were performed using Fisher's exact test for 2 X 2 contingency tables and the Mantel test for trend for contingency tables.36 Survival times were calculated from the date of protocol entry to the date of death or last known date alive. Death from any cause was treated as the main outcome event in all such analyses. Survival probabilities were calculated and displayed using the method of Kaplan and Meier37 and the curves were compared statistically to one another using the Mantel-Haenszel test.38 Cox proportional hazards modeling39 was used to evaluate the importance of several factors simultaneously to predict survival. The resulting model parameters (bi) were converted to relative risks by computing exp(b;) where exp(a) = 2.71828a.40 The 95% confidence interval for the relative risk was computed as [exp(biL), exp (biH)] where b iL = b; -1.96 [estimated standard error (b;)] and b iH = bj + 1.96 [estimated standard error(bj)]. The relative risk indicates the risk associated with dying while being in a greater risk category compared with that of being in a lower risk category. All probability values in this report correspond to two-sided significance tests. RESULTS Non-Neoplastic Lung and Other Organs General staining patterns with both antibodies against SP-A were similar in non-neoplastic lungs, revealing a
Vol. 97 No. 2
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
Immunohistochemical staining was performed by the avidin-biotinylated peroxidase-complex technique using Vectastain avidin-biotinylated peroxidase-complex staining kits (Vector Laboratories, Burlingame, CA) according to vendor's instructions and modified as previously reported.31 Incubations with primary antisera were performed at 4 °C for 18 hours. Each primary antibody was titered specifically for the present study using paraffin sections of the control tumors. Furthermore, the dilution was considered optimal when the antibodies SAM and SAP-35 produced strong staining in Type II pneumocytes and the CLR-10 antibody stained intensely Clara cells in the bronchioles of non-neoplastic lung, whereas other pulmonary cells remained unstained. Positive control sections for each antibody from appropriate tumors were included in each assay. Controls for specificity of staining consisted of the incubation of serial lung tumor sections with (1) nonimmune (normal) rabbit serum or (2) phosphate-buffered saline. Results of the immunostaining and mucicarmine histochemical analysis were reviewed by two of the authors (R.I.L. and S.M.J.) who independently scored both for the intensity of the staining (0 = negative; 1 = weak; 2 = moderate; 3 = strong reaction) and the number of positive cells (distribution score: 0 = no positive cells; 1 for 10% of tumor cells positive). However, the presence of rare, scattered cells immunoreactive for SPA or CLR-10 was a common finding in many more tumors.
Vol. 97 • No. 2
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
Initial studies indicated that with SAP-35 antibody there was a variable amount of nonspecific stromal connective tissue reactivity. Furthermore, the tumor-vmi«-background resolution with this antibody was frequently less optimal than that achieved using antibody SAM. When we compared the staining patterns of various histologic types of lung tumors with the two SP-A antibodies, a higher number of nonadenocarcinomas were positive after staining with SAP-35 antibody than with SAM antibody (Fig. 3). Consequently, in further screening studies of lung tumors, we chose to use antibody SAM to detect SP-A immunoreactivity. The immunoreactivity for SP-A and CLR-10 in tumors was mostly focal and often less intense than in non-neoplastic reactive cells in the same sections that stained positively for these markers. A tumor was scored positive if more than 10% tumor cells demonstrated immunoreactivity for any given marker. Immunostaining was cytoplasmic and the nuclei remained negative, except for nucleolar immunoreactivity, which was seen in 19 of 48 (40%) of SP-A and nuclear immunoprecipitate in 6 of 38
238
ANATOMIC PATHOLOGY Original Article
" -• r':-
'''MM^'^'^L'
V-*
*"
^
*'*« 2 = 0.10)
21/79(27) 9/34 (26) (^2= 1.0)
7/41(17) 1/40(3) (P2 = 0.06) 6/25 (24) 3/14(21) (^2 = 1.0)
13/41 (31) 8/38(21) (P2 = 0.32) 9/20 (45) 0/14(0) (P2 = 0.004)
Data
TABLE 4. SUMMARY OF BIOGRAPHIC DATA SP-A and/or CLR-10 Immunoreactivity in Tumors
Age (years) 0-49 50-59 60-69 70+
Data
Positive*
Negative
Sex Male Female
Number of tumors Median age, years Male sex, percentage Smoking history, pack-years Curable patients, percentaget
44 (36%) 54 61 38 43
78 (64%) 58 71 56 53
Pack Years Male 0-49 (light) 50+ (heavy)
• £10% tumor cells positive for SP-A and/or CLR-10 by immunohistochemistry. t Patients receiving therapy with curative intent who had tumor confined to the thorax, including Stage 1, II, and I1A patients who underwent either surgical resection or definitive radiotherapy using standard indications.
Female 0-49 (light) 50+ (heavy)
* Mantel test for trend. For all other statistical analyses Fisher's exact test was used.
A.J.C.P. • February 1992
Downloaded from http://ajcp.oxfordjournals.org/ by guest on June 6, 2016
NCI = National Cancer Institute. * Positive = ^ 10% of tumor cells stain, t Staining data using antibody SAM. } SP-A/CLR-10 = SP-A and/or CLR-10-positive tumors. § Papillolepidic catefory includes bronchioloalveolar carcinomas, adenocarcinomas with marked bronchioloalveolar component, and papillary adenocarcinomas.
LINNOILA ET AL. Peripheral Airway Cell Marker Expression in Non-Small Cell Lung Carcinoma
241
TABLE 6. COX PROPORTIONAL HAZARDS MODEL Variable
P*
-0.72 0.53 0.67 -0.60 0.42 2.37
0.0048 0.0399 0.0253 0.0375 0.0063
