The epithelial alteration in interstitial pneumoniasis one of the repair processes at the sites of dis-ease activity. Regenerative epithelial cells mayparticipate in remodeling of the lung. To determinethe phenotype of regenerative epithelial cells inusual interstitial pneumonia (UIP) and nonspecificinterstitial pneumonia (NSIP), the expression ofClara cell 10KD protein (CC10), cytokeratin (CK) 14and 17, surfactant apoprotein (SP)-A, KL-6/MUC1,transforming growth factor (TGF) ββ2 were examinedin 25 patients with UIP, 9 patients with NSIP andnormal lung tissues from 10 patients with lung can-cer.

In honeycomb lesions of UIP, non-ciliated co-lumnar cells mainly expressed CC10, cuboidalcells expressed CC10, CK17, CK14 and SP-A indescending order.

Fibroblastic foci are covered by CK17, CK14,CC10, and a few SP-A positive flattened orcuboidal cells. Regenerative epithelium in NSIPmainly comprised cuboidal cells expressing SP-A,CC10 and CK17. KL-6 was more remarkablyexpressed in cuboidal and non-ciliated columnarcells both in UIP and NSIP. Expression of TGFββ2

was observed in cuboidal and flattened epithelium.In severe fibrotic areas, CC10 expressing cells

were more prominent, while SP-A positive cellswere more prominent in less fibrotic areas.

Regenerative epithelial cells in remodeling areain UIP may be derived from bronchiolar basalcells and Clara cells, while most of those in NSIPmay be derived from type II pneumocytes. The dif-ferent origin of regenerative epithelium mayreflect the severity and extent of the injury and thedegree of consequent fibrosis in UIP and NSIP.

Key words: Regenerative epithelium, phenotype,bronchiolar basal cell, UIP, NSIP

Introduction

Among the idiopathic interstitial pneumonias (IIPs),idiopathic pulmonary fibrosis (IPF) is the most frequentchronic fibrosing interstitial pneumonia, which histo-logically shows a usual interstitial pneumonia (UIP) pat-tern. The histologic hallmarks and chief diagnostic cri-teria of UIP are a heterogeneous appearance of thefibrotic areas with alternating areas of normal lung,scattered fibroblastic foci, and honeycomb change.1-5

Nonspecific interstitial pneumonia (NSIP) has a broadspectrum of histologic findings comprising cellular andfibrosing patterns and has a significantly better prog-nosis than that of IPF.1,5 The pathological diagnosis ofUIP and NSIP is primarily based on distribution andtemporal appearance of inflammation and fibrosis,and remodeling of alveoli.1-3,5 The histological alter-ations of interstitial pneumonias consist of the injuryand repair at the disease activity, and it is important to

Original Article

Phenotype of regenerative epithelium in idiopathic interstitial pneumonias

Nae Hinata1,2, Tamiko Takemura2, Soichiro Ikushima3, Takashi Yanagawa3, Tsunehiro Ando3,Junko Okada3, Masaru Oritsu3, and Morio Koike1

1) Department of Human Pathology, School of Medicine Tokyo Medical and Dental University2) Department of Pathology, Japanese Red Cross Medical Center, Tokyo3) Department of Respiratory Medicine, Japanese Red Cross Medical Center, Tokyo

J Med Dent Sci 2003; 50: 213–224

Corresponding Author: Koike Morio, M.D.Department of Human Pathology, School of Medicine TokyoMedical and Dental University5-45, Yushima 1-chome, Bunkyo-ku Tokyo, 113-8519 JapanPhone: 03-5803-5172Fax: 03-5803-0123E-mail address: koike.pth1@tmd.ac.jpReceived March 20; Accepted June 13, 2003

know whether the lesions are ongoing or steady fortreatment and prognosis.

The epithelial alteration in the interstitial pneumoniasis one of the repair processes at the sites of diseaseactivity. Type II pneumocytes and regenerative epithe-lial cells produce various chemoattractants for fibrob-lasts and may participate in the fibrosing process ofinterstitial pneumonias.6,7

Areas of honeycomb change are composed of cysticremodeling air spaces that are frequently lined withmetaplastic bronchiolar epithelia or hyperplastic alveo-lar pneumocytes, and occasionally metaplastic squa-mous epithelium.1-4 The phenotype of regenerativeepithelium is considered to be changeable, dependingupon the new insults during the progression orrestoration of the disease. As for the morphologicalphenotype of epithelial cells of fibrotic lung disorders,Kawanami et al. reported that type II cells proliferatemainly in the areas with less severe degrees of fibrosis,cuboidal cells are heterogeneous, comprising bronchi-olar basal cells, and respiratory bronchioles becomethe main source of epithelial renewal in areas of verysevere lung damage.8 Regarding the immunohisto-chemical phenotype of UIP, Iyonaga et al. reported var-ious cytokeratin (CK) expression in the regenerativeepithelium of IPF.9 They suggested that the regenera-tive epithelial cells are altered population and may playan important role in the progression of chronic inflam-mation in IPF.

It has been reported that Clara cells may aid bron-chiolar repopulation following the injury.10-12 Secretoryprotein specific for Clara cells is identified as Clara cell10KD protein (CC10).13 Surfactant apoprotein (SP)-A ishydrophilic glycoprotein, and secreted by type IIpneumocytes.14 KL-6 is a high molecular weight glyco-protein and classified as “cluster 9 (MUC1)” of lungtumor,15 and useful indicator of disease activity inpatients with interstitial pneumonias.16 It has beenreported that KL-6 was expressed in regenerativetype II pneumocytes of interstitial pneumonias.16 As forfibrosis in the lung, transforming growth factor (TGF)βcan regulate cell growth and differentiation as well asstimulation of fibroblasts to synthesize collagen,fibronectin, proteoglycans, and other proteins of theextracellular matrix, and participates in the fibroticprocess in interstitial lung disease.17,18

This study intends to 1) clarify the phenotype ofregenerative epithelial cells in UIP and NSIP at the sitesof disease activity using the immunohistochemicalanalysis, and 2) compare the pathological status of UIPand NSIP, referring the correlation between regenera-

tive epithelial cells and the degree of fibrosis as theactivity of the disease.

Material and methods

We examined 25 cases of IPF/UIP (22 males and 3females) and 9 cases of NSIP (6 males and 3females) by video-assisted thoracoscopic lung biopsyand surgical resection at the Japanese Red CrossMedical Center in Tokyo between 1990 and 2002. Thetwenty-five cases of IPF/UIP consisted of 9 cases ofIPF/UIP without lung cancer and 16 cases of UIP asso-ciated with lung cancer. The average age of patientswith IPF/UIP was 63±9 years, NSIP was 64±10years, and the average smoking index (cigarettes/day, years) for IPF/UIP was 1254±963, and NSIP was1486±953. The diagnosis of UIP and NSIP wasmade according to ATS/ERS international multidisci-plinary consensus classification of the idiopathic inter-stitial pneumonias, 2002.1 The diagnosis of IPF/UIPwas confirmed based on clinical symptoms such asdyspnea on exertion and dry cough, reduced pul-monary function, reticulonodular shadow on chest X-rayand CT, high levels of serum SP-A, SP-D and KL-6,and histopathology of UIP.3-5 However, as the data forSP-D and KL-6 were available only in the severalcases, we do not compare these serum data withhistopathology in this study. NSIP was confirmedbased on typical bilateral pulmonary interstitial infiltra-tion showing as an opaque ground glass shadow onchest X-ray and CT, other clinical symptoms similar tothose for IPF/UIP, and histopathology of diffuse, tem-porally homogeneous interstitial infiltrates and fibro-sis.5,19 In this study, 7 cases of NSIP are classified asgroup II and 2 cases as group III, according toKatzenstein.19 Any cases of NSIP are not associatedwith collagen vascular disease from serological data.

We also examined 10 cases of normal lung tissuewithout IPF or other chronic lung diseases, whichwere taken well away from the primary lung cancer andexamined as controls. They were 6 males and 4females with a mean age of 65±9 years and an aver-age smoking index of 1000±910.

All the lungs were fixed in 10% buffered formalin andembedded in paraffin. All paraffin sections were rou-tinely stained with hematoxylin-eosin, elastica vanGieson, Masson-trichrome and periodic acid-schiff-alcian blue.

For immunohistochemical analysis of regenerativeepithelial cells, we used polyclonal antibodies against

N. HINATA et al. J Med Dent Sci214

CC10 and TGFβ2, and monoclonal antibodies againstCK17, CK14, SP-A and KL-6. The antibodies andsources are given in Table 1. Paraffin-embedded,4µm-thick sections were deparaffinized in xylene andrehydrated. They were washed in phosphate-bufferedsaline (PBS) with 0.01% Tween 20, and incubated for30 minutes in 0.3% H2O2 in methanol to inactivateendogenous peroxidase. For antigen retrieval, thesections were digested with 0.05% trypsin for 30 min-utes at room temperature or 0.1% hyaluronidase at37°C for 30 minutes, or heated in a microwave oven at500w for 15 minutes in a 10mM sodium citrate buffer atPH6.0. Then the sections were incubated at 4°Covernight with the primary antibodies, then the sec-ondary antibody, biotinylated goat antimouse IgG orantirabbit IgG (DAKO Japan, Kyoto, Japan) wasapplied for 30 minutes at room temperature. Afterwashing three times in PBS, they were reacted withperoxidase-conjugated streptavidin (0397, DAKOJapan, Kyoto, Japan). Finally, the reaction productswere visualized with 0.02% 3,3’ diaminobenzidine(DAB) in a 0.1 mol/L Tris-HCl buffer, PH7.6, in0.005% H2O2 for 5 minutes, and lightly counter-stained with hematoxylin. As for CC10 and CK17,CC10 and SP-A, CC10 and KL-6 double staining fol-lowing visualization by DAB of the first stain, the sec-tions were heated in a microwave oven at 500w for 5minutes. Then the sections were incubated with the pri-mary antibody at 4°C overnight, then with secondaryantibody, alkaline phosphatase-conjugated goat anti-mouse IgG for 30 minutes at room temperature. Afterwashing in PBS, finally, the reaction products werevisualized with Vector Blue Alkaline phosphatase sub-strate kitIII, and lightly counter-stained with nuclear fastred. Negative controls consisted of a non-immune, iso-type-matched monoclonal antibody. The sameimmunohistochemical study was performed in normal

controls. To evaluate the heterogeneity of regenerative

epithelial cells we counted the positive cells for CC10,CK17, CK14, SP-A, KL-6 and TGFβ2 in the unit area(mm2) in UIP and NSIP cases, using an Olympusmicrometer with a ×20 objective lens in four repre-sentative areas with fibrotic remodeling in each case.The whole epithelial cells in the unit area (mm2) corre-spond to the parent population of the cells. The per-centage of positive cells was counted in each sectionand presented as mean±standard deviation. Amongthe regenerative epithelial cells, we focused oncuboidal cells and counted them in three types,because they frequently appear in fibrotic process ininterstitial pneumonias.

To determine the relationship between regenerativeepithelial phenotype and fibrosis, we counted the pos-itive cells for CC10, CK17, SP-A, KL-6 and TGFβ2 inthe determined areas of fibrosis using the grading ofAshcroft,20 and expressed them as percentage.Criteria for grading lung fibrosis by Ashcroft: histologi-cal features of fibrosis score of grade 0 was normallung, grade 1 to 2 was minimal fibrous thickening ofalveolar or bronchiolar walls, grade 3 to 4 was moder-ate thickening of walls without obvious damage to lungarchitecture, grade 5 to 6 was increased fibrosis withdefinite damage to lung structure and formation offibrous bands or small fibrous masses, grade 7 to 8was severe distortion of structure and large fibrousareas and honeycomb lesion is placed in this category.The statistical analysis was performed by the MannWhitney test, and p values less than 0.05 were con-sidered significant.

215EPITHELIAL PHENOTYPE IN PULMONARY FIBROSIS

Table 1. Antibodies used

Results

1. Immunohistochemical analysisNormal lung tissues1) Expression of CC10, CK17, CK14 and SP-A.In the bronchiolar epithelium, CC10 was intensely

expressed in the cytoplasm of Clara cells in a uniformstaining pattern and some of the basal cells [Fig. 1a].

CK17 was constantly expressed in basal cells,while CK14 expressed in some of the basal cells ofbronchiole [Fig. 1b, c]. SP-A was strongly granular-pos-itive in type II pneumocytes [Fig. 1d].

2) Expression of KL-6 and TGFβ2.KL-6 was strongly expressed in the apical portion

and uniformly expressed in the cytoplasm of Clara cellsand some of the basal cells of bronchiole [Fig. 2a], andalso in the cytoplasmic membrane and cytoplasm oftype II pneumocytes [Fig. 2b]. Occasionally, KL-6 wasco-expressed in CC10 positive non-ciliated columnarcells. TGFβ2 was expressed in some of the non-ciliatedcolumnar cells, basal cells of the bronchiole [Fig. 2c]and alveolar macrophages whereas type II pneumo-cytes were negative to TGFβ2.

Fibrotic lung tissues1) UIP cases.(1) Expression of CC10, CK17, CK14 and SP-A.Fibrous areas showed a patchy distribution at the

subpleural and periphery of the lobules. Honeycomblesions were characterized by irregularly dilated air-

N. HINATA et al. J Med Dent Sci216

Fig. 1. Expression of CC10, CK17, CK14 and SP-A in normal lung tissue.(a) CC10 was intensely expressed in the cytoplasm of Clara cells in uniform staining pattern and some of the basal cells.(b) CK17 was constantly expressed in basal cells of bronchiole.(c) CK14 was expressed in some of the basal cells of bronchiole.(d) SP-A was expressed in type II pneumocytes in granular staining pattern.

spaces that were lined by bronchiolar columnar,cuboidal, flattened and occasionally squamous meta-plastic epithelium [Fig. 3a]. Honeycomb lesions weredetected in 22 of 25 cases (88%) of UIP, while micro-scopic small honeycomb lesion was seen in 1 of 9cases (11.1%) of NSIP.

In honeycomb lesion, CC10 was intenselyexpressed in the cytoplasm of non-ciliated columnarepithelial cells [Fig. 3b]. CC10-positive non-ciliatedcolumnar epithelium occupied an average of 50% ofthe lining epithelium in the honeycomb lesion (Table 2).The cuboidal cells of the lining epithelium in the hon-eycomb lesion were divided into three phenotypes;CC10-positive, CK-17 and CK-14-positive and SP-A-positive cells [Fig. 3c] in descending order. But eachincidence was less than 25% (Table 2). A few of thenon-ciliated columnar epithelial cells co-expressedCC10 and SP-A in honeycomb areas [Fig. 3b]. Some ofthe basal cells of bronchiolar metaplastic epitheliumexpressed CC10, CK14 and CK17 [Fig. 3d]. Flattenedepithelium were occasionally observed in the honey-comb lesion, which were positive for CC10, CK14,CK17 and SP-A, in random distribution. Squamousmetaplastic epithelial cells intensely expressed CK14and CK17 (Table 2).

Fibroblastic foci was detected in 80% of the UIPcases. They were sparsely distributed in the margin ofdense fibrosis of the periphery of the lobule and hon-eycomb lesion. They were usually lined by flattened orcuboidal cells [Fig. 3a]. Flattened cells usuallyexpressed CK17, CK14 and CC10, while cuboidalcells expressed CC10 and CK17. CK17 and CC10-pos-itive cells each occupied an average of 50% and 20% of the lining epithelium in the fibroblastic foci.Occasionally, CC10 positive cuboidal cells were over-lying CK17 positive cells [Fig. 3e]. Rarely, SP-A wasexpressed in cuboidal cells in the fibroblastic foci.

(2) Expression of KL-6 and TGFβ2.In honeycomb lesion, KL-6 was strongly expressed in

the cytoplasm and apical portion of non-ciliatedcolumnar and cuboidal cells [Fig. 4a]. KL-6 positivecells occupied an average of 75% of the lining epitheli-um (Table 2). KL-6 was expressed in flattened epithelialcells and some of the basal cells in honeycomb lesion[Fig. 4a]. Some of the non-ciliated and cuboidal cellsco-expressed CC10 and KL-6. KL-6 was also ex-pressed flattened and cuboidal epithelial cells infibroblastic foci [Fig. 4b].

In honeycomb lesion, TGFβ2 was expressed mainlyin cuboidal and flattened epithelial cells, occasionally insome of the non-ciliated columnar epithelial cells,

217EPITHELIAL PHENOTYPE IN PULMONARY FIBROSIS

Fig. 2. Expression of KL-6 and TGFβ2 in normal lung tissue.(a) KL-6 was strongly expressed in the apical portion and uniformlyexpressed in the cytoplasm of Clara cells and some of the basal cellsof bronchiole.(b) KL-6 was strongly expressed in the cytoplasmic membrane andcytoplasm of type II pneumocytes.(c) TGFβ2 was expressed in some of the non-ciliated columnar cellsand basal cells of the bronchiole.

squamous metaplastic epithelial cells and some of themetaplastic bronchiolar basal cells [Fig. 4c]. In fibrob-lastic foci, flattened and cuboidal cells also expressedTGFβ2 [Fig. 4d].

2) NSIP cases(1) Expression of CC10, CK17, CK14 and SP-A.NSIP revealed relatively less structural remodeling

with proliferation of cuboidal epithelial cells and flat-tened epithelial cells [Fig. 5a]. Cuboidal epithelial cellswas divided into three phenotypes; i.e., SP-A-positivecells were the most prominent [Fig. 5b], and CC10-pos-itive, CK14 and CK17-positive cells were less than25%. CC10-positive cuboidal cells were sparse andusually appeared close to the bronchiole [Fig. 5c].Some of the flattened epithelial cells in the muralincorporation expressed CC10, CK17, and SP-A.(2)Expression of KL-6 and TGFβ2.

KL-6 was strongly expressed in cuboidal and flat-tened epithelial cells [Fig. 5d]. CC10 and KL-6 were co-expressed in some cuboidal epithelial cells.Occasionally, TGFβ2 was expressed in cuboidal andflattened epithelial cells [Fig. 5e]. The phenotypes ofregenerative epithelial cells in fibrotic lung tissuesfrom patients with UIP and NSIP are summarized inTable 2.

2. Correlation between regenerative epithelialphenotype and fibrosis in UIP and NSIP pattern.

The grading of fibrosis in each of the lung section

was scored on a scale from 0 to 8, according toAshcroft. Fibrosis scores of grade 5 to 6 and grade 7 to8 were recognized in the 25 cases of UIP, grade 3 to 4was recognized in the 25 cases comprising 16 UIP and9 NSIP.

Most of the cases with UIP had fibrosis scores ofgrade 7 to 8, with NSIP scores of grade 3 to 4. The inci-dence of various phenotypes of regenerative epithelialcells and grade of fibrosis are shown in Figure 6. Thepercentage of positive cells for CC10 was significantlyhigher in the fibrosis score of grade 7 to 8 (p<0.005)and 5 to 6 (p<0.05) areas, than in the fibrosis score ofgrade 3 to 4 areas. The percentage of positive cells forSP-A was significantly higher in the fibrosis score ofgrade 3 to 4 areas, than in the fibrosis score of grade 5to 6 (p<0.005) and 7 to 8 (p<0.0001) areas. The per-centage of positive cells for KL-6 was significantly high-er in the 2 groups with fibrosis score of grade 3 to 6areas, than in the fibrosis score of grade 7 to 8 areas(p<0.0001).

Positive cells for CC10 and CK17 proliferated main-ly in the more fibrotic areas, while SP-A and KL-6-pos-itive cells proliferated mainly in the relatively lessfibrotic areas. Further, the percentage of TGFβ2 positivecells was significantly higher in fibrosis score of grade7 to 8 than in fibrosis score of grade 3 to 4(p<0.0001).

In fibrosis score of grade 3 to 4, the percentage ofpositive cells to CC10, CK17, SP-A, KL-6 and TGFβ2

N. HINATA et al. J Med Dent Sci218

Table 2. Immunohistochemical phenotype of regenerative epithelial cells in normal, UIP and NSIP lung tissue

219EPITHELIAL PHENOTYPE IN PULMONARY FIBROSIS

Fig. 3. Expression of CC10, CK17 and SP-A in UIP.(a) Honeycomb lesion was lined by bronchiolar columnar, cuboidal and flattened epithelium, and fibroblastic focus (arrow) lined by cuboidal andflattened epithelium. (HE)(b) CC10 (brown) was strongly expressed in non-ciliated columnar epithelial cells, and SP-A (blue) was expressed in some of the non- ciliat-ed columnar cells, also a few of the non-ciliated columnar epithelial cells co-expressed CC10 and SP-A in the honeycomb lesion.(c) SP-A (brown) was expressed in some of the cuboidal cells in the honeycomb lesion. (d) CC10 (brown) was strongly expressed in non-ciliated columnar cells, and CK17 (blue) was strongly expressed in basal cells of the bron-chiolar metaplastic epithelium in the honeycomb lesion. (e) CC10 (brown) positive cuboidal cells were overlying CK17 (blue) positive cells of lining epithelium of the fibroblastic focus.

was not significantly different between UIP and NSIP.

Discussion

The present immunohistochemical study demon-strated more heterogeneous appearance of pheno-types of regenerative epithelial cells in the fibroticareas of UIP than in those of NSIP. Although the fibrousareas in UIP show temporal heterogeneity, this studydemonstrated heterogeneity of phenotypes of regen-erative epithelial cells in the remodeling areas. Theregenerative epithelium in the honeycomb lesion and

fibroblastic foci consisted of CC10-positive cells,CK17, CK14-positive cells and a few SP-A-positivecells. In the honeycomb lesion, CC10-positive cellswere the most prominent. CC10-positive cells com-prised non-ciliated columnar, cuboidal and occasional-ly flattened cells in the honeycomb lesion and fibrob-lastic foci of UIP. CC10 is specific for Clara cells13 whichhas been recognized as progenitor cells of the bron-chiolar epithelium,10 and contribute to repopulationafter injury.10-12 CC10 has been reported as a candidatefor controlling inflammatory events in the lung,11-13,21-23

and inhibits platelet-derived growth factor (PDGF)-induced chemotaxis of lung fibroblasts in a dose-

N. HINATA et al. J Med Dent Sci220

Fig. 4. Expression of KL-6 and TGFβ2 in UIP.(a) KL-6 was strongly expressed in the apical portion and cytoplasm of non-ciliated columnar cells and some of the basal cells of bronchiolarmetaplastic epithelium in the honeycomb lesion.(b) KL-6 was expressed in cuboidal and flattened cells of lining epithelium of the fibroblastic foci.(c) TGFβ2 was expressed in cuboidal and non-ciliated epithelial cells, and some of the basal cells of bronchiolar metaplastic epithelium in thehoneycomb lesion.(d) TGFβ2 was expressed in the lining epithelium in fibroblastic foci.

221EPITHELIAL PHENOTYPE IN PULMONARY FIBROSIS

Fig. 5. Expression of SP-A, CC10, KL-6 and TGFβ2 in NSIP.(a) Proliferation of cuboidal epithelial cells in the alveolar walls. (HE) (b) SP-A was strongly expressed in cuboidal epithelial cells in a granular staining pattern.(c) CC10 was expressed in some of the cuboidal cells, was appeared close to the bronchiole.(d) KL-6 was strongly expressed in cuboidal epithelial cells remarkably at the apical portion.(e) TGFβ2 was expressed in cuboidal epithelial cells.

dependent manner,21-23 and a decreased availability ofCC10 may facilitate the recruitment of fibroblasts infibrosing lung disorders.21-23 Subpopulation of CC10-expressing cells may importantly contribute to regen-eration after various injury.10-12 Cuboidal cells arecharacteristic pattern of regenerative epithelial cells,which usually appear in interstitial lung disorders.Electron microscopic study by Kawanami revealedtwo types of cuboidal cells, one derived from bronchio-lar basal cells and the other from cuboidal cells in therespiratory bronchioles.8

In UIP, cuboidal cells in honeycomb lesion may be divided into CC10(+)/SP-A(-), CK14(+)/SP-A(-),CK17(+)/SP-A(-), and SP-A(+) cells. In fibroblasticfoci, CK17(+)/SP-A(-), CK14(+)/SP-A(-), CC10(+)/SP-A(-) and CK14(-) CK17(-)/SP-A(+) cuboidal and flat-tened cells are observed. In NSIP, cuboidal cells maybe divided into SP-A(+)/CK17(-) or CK14(-) hyper-plastic type II pneumocytes, CC10(+)/SP-A(-), CK17(+)/SP-A(-) and CK14(+)/SP-A(-) cells. The latter two

revealed phenotype of bronchiolar basal cells, whichmay be derived from bronchiolar basal cells as report-ed by Iyonaga.9

Fibroblastic foci in UIP are considered to formyoung reparative lesions due to new insults occurring inthe remodeling fibrotic area in UIP.1-4 CC10-positivecuboidal cells overlying CK17-positive cells in fibrob-lastic foci suggest that CK17-positive cells appear in theearly phase and then CC10-positive cells cover them.CK17-positive bronchiolar basal cells may contribute tothe initial repair. CK17 or CK14 and CC10-positive flat-tened cells may be the stem cell population in theremodeling areas of chronic interstitial lung disordersand some of them may develop to SP-A expressingcells.

In UIP, regenerative epithelial cells reveal bronchiolarphenotype, while in NSIP, most of those reveal pheno-type of type II pneumocyte. The bronchiolar phenotypeor type II pneumocyte of regenerative epithelial cellsmay reflect the severity and extent of the injury of the

N. HINATA et al. J Med Dent Sci222

Fig. 6. Correlation between regenerative epithelial phenotype and grade of fibrosis in UIP and NSIP.* p<0.005 CC10-positive cells(%) in the fibrosis score of grade 7 to 8 vs grade 3 to 4. ** p<0.05 CC10-positive cells(%) in the fibrosis score of grade 5 to 6 vs grade 3 to 4. + p<0.0001 SP-A-positive cells(%) in the fibrosis score of grade 7 to 8 vs grade 3 to 4. ++ p<0.005 SP-A-positive cells(%) in the fibrosis score of grade 5 to 6 vs grade 3 to 4. @ p<0.0001 KL-6-positive cells(%) in the fibrosis score of grade 3 to 4 vs grade 7 to 8.@@ p<0.0001 KL-6-positive cells(%) in the fibrosis score of grade 5 to 6 vs grade 7 to 8.# p<0.0001 TGFβ2-positive cells(%) in the fibrosis score of grade 7 to 8 vs grade 3 to 4.

Grade of fibrosis

alveoli, which also may be associated with the degreeof fibrosis.

The frequency of phenotype of regenerative epithe-lium is different between less fibrous areas and fibroticareas in UIP and NSIP. In the more fibrotic areas in UIPwith grade 7 to 8 of Ashcroft’s fibrosis score, CC10-pos-itive columnar cells are significantly predominant.CK17-positive cells are also detected in these areas. Inthese more fibrous areas, inflammatory infiltration isvery mild. These fibrous areas in the periphery of thelobule in UIP with CC10 and CK17 positive regenera-tive cells may be reconstructed to the epithelial andstromal microenvironment of bronchiolar type,although the alterations of microvessels have notbeen studied.

Recently, KL-6, SP-A and SP-D have been reportedto be sensitive, useful serum markers for the diagnosisand monitoring of patients with interstitial pneumo-nias.16,24-28 KL-6 has a property as a chemotactic factorfor fibroblasts.29 The elevation of serum KL-6 reflectssevere epithelial damage as well as alveolar wall cap-illary damage in acute and subacute interstitial lung dis-orders.16,25-28 This study clearly demonstrated that KL-6was strongly positive to the regenerative cuboidalcells and non-ciliated columnar cells, while in the nor-mal lung, KL-6 was positive to type II pneumocytes,Clara cells and basal cells of the bronchiole. SP-A andKL-6-positive cuboidal cells are more prominent in theless fibrous areas with grade 3 to 6, especially KL-6-positive cuboidal cells. These results are consistentwith several reports on serum markers of an activephase of interstitial pneumonias with ongoing epithelialand capillary damage.16,24-28

Type II pneumocytes and regenerative epitheliumalso produce chemoattractants for fibroblasts such asfibronectin,6,7 TGFβ,17,18 and PDGF.23 It has beendemonstrated that TGFβ2 was ubiquitously expressedin pulmonary epithelial cells in normal and fibroticlungs.30 In this study, TGFβ2 was expressed mainly inthe regenerative cuboidal cells and flattened cells in theareas than in non-ciliated columnar cells, indicating thatthese cells appear in the sites of active fibroblastic pro-liferation in UIP and NSIP. The TGFβ family has beenimplicated in the fibrosis in interstitial lung disor-ders.17,18 TGFβ family is a chemoattractant for fibrob-lasts, so that at sites of injury the presence of TGFβcould expand the population of fibroblasts, which arethe main source of conective tissue proteins.17,18,31

Our results suggest that TGFβ2 in regenerativeepithelium may participate in a more active fibroblasticproliferation process.

In conclusion, regenerative epithelial cells inremodeling area in UIP may be derived from bronchio-lar basal cells and Clara cells, while most of those inNSIP may be derived from type II pneumocytes. Thedifferent origin of regenerative epithelium may reflectthe severity and extent of the injury and the degree ofconsequent fibrosis in UIP and NSIP.

References1. American thoracic society. American thoracic society /

European respiratory society international multidisciplinaryconsensus classification of the idiopathic interstitial pneumo-nias. Am J Respir Crit Care Med 2002;165:277-304.

2. American thoracic society. Idiopathic pulmonary fibrosis:diagnosis and treatment. International consensus statement.Am J Respir Crit Care Med 2000;161:646-64.

3. Katzenstein ALA, Myers JL. Idiopathic pulmonary fibrosis.Clinical relevance of pathologic classification. Am J Respir CritCare Med 1998;157:1301-15.

4. Liebow AA. Definition and classification of interstitial pneu-monias in human pathology. Prog Respir Res 1975;8:1-33.

5. Travis WD, Matsui K, Moss J, et al. Idiopathic nonspecificinterstitial pneumonia: prognostic significance of cellular andfibrosing patterns. Survival comparison with usual interstitialpneumonia and desquamative interstitial pneumonia. Am JSurg Pathol 2000;24:19-33.

6. Rannels SR, Fisher CS, Heuser LJ, et al. Culture of type IIpneumocytes on a type II cell-derived fibronectin-rich matrix.Am J Physiol 1987;253:759-65.

7. Dunsmore SE, Martinez-Williams C, Goodman RA, et al.Turnover of fibronectin and laminin by alveolar epithelialcells. Am J Physiol 1995;269:766-75.

8. Kawanami O, Ferrans VJ, Crystal RG. Structure of alveolarepithelial cells in patients with fibrotic lung disorders. LabInvest 1982;46:39-53.

9. Iyonaga K, Miyajima M, Suga M, et al. Alterations in cytoker-atin expression by the alveolar lining epithelial cells in lung tis-sues from patients with idiopathic pulmonary fibrosis. JPathol 1997;182:217-24.

10. Otto WR. Lung epithelial stem cells. J Pathol 2002;197:527-35.

11. Giangreco A, Reynolds SD, Stripp BR. Terminal bronchiolesharbor a unique airway stem cell population that localizes tothe bronchoalveolar duct junction. Am J Pathol2002;161:173-82.

12. Evans MJ, Cabral-Anderson LJ, Freeman G. Role of the Claracell in renewal of the bronchiolar epithelium. Lab Invest1978;38:648-53.

13. Hermans C, Bernard A. Lung epithelium-specific proteins.Characteristics and potential applications as markers. Am JRespir Crit Care Med 1999;159:646-78.

14. Kuroki Y, Mason RJ, Voelker DR. Pulmonary surfactantapoprotein A structure and modulation of surfactant secretionby rat alveolar type II cells. J Biol Chem 1998;263:3388-94.

15. Kohno N, Akiyama M, Kyoizumi S, et al. Detection of solubletumor- associated antigens is sera and effusions using novelmonoclonal antibodies, KL-3 and KL-6, against lung adeno-carcinoma. Jpn J Clin Oncol 1988;18:203-16.

16. Kohno N, Kyoizumi S, Awaya Y, et al. New serum indicator ofinterstitial pneumonitis activity. Sialylated carbohydrate antigen

223EPITHELIAL PHENOTYPE IN PULMONARY FIBROSIS

KL-6 Chest 1989;96:68-73.17. Khalil N, O’Connor RN, Unruh HW, et al. Increased production

and immunohistochemical localization of transforminggrowth factor-β in idiopathic pulmonary fibrosis. Am J RespirCell Mol Biol 1991;5:155-62.

18. Ignotz RA, Massague J. Transforming growth factor-betastimulates the expression of fibronectin and collagen and theirincorporation into the extracellular matrix. J Biol Chem1986;261:4337-45.

19. Katzenstein ALA, Fiorelli RF. Nonspecific interstitial pneumo-nia/ fibrosis. Histologic features and clinical significance. Am JSurg Pathol 1994;18:136-47.

20. Ashcroft T, Simpson JM, Timbrell V. Simple method of esti-mating severity of pulmonary fibrosis on a numerical scale. JClin Pathol 1998;41:467-70.

21. Lesur O, Bernard A, Arsalane K, et al. Clara cell protein(CC16) induces a phospholipase A2-mediated inhibition offibroblast migration in vitro. Am J Respir Crit Care Med1995;152:290-7.

22. Johnston CJ, Mango GW, Finkelstein JN, et al. Altered pul-monary response to hyperoxia in Clara cell secretory proteindeficient mice. Am J Respir Cell Mol Biol 1997;17:147-55.

23. Antoniades HN, Bravo MA, Avila RE, et al. Platelet-derivedgrowth factor in idiopathic pulmonary fibrosis. J Clin Invest1990;86:1055-64.

24. Takahashi H, Fujishima T, Koba H, et al. Serum surfactant pro-tein A and D as prognostic factors in idiopathic pulmonaryfibrosis and their relationship to disease extent. Am J RespirCrit Care Med 2000;162: 1109-14.

25. Kohno N. Serum marker KL-6/MUC1 for the diagnosis andmanagement of interstitial pneumonitis. J Med Invest1999;46:151-8.

26. Ohnishi H, Yokoyama A, Kondo K, et al. Comparative study ofKL-6, surfactant protein-A, surfactant protein-D, and monocytechemoattractant protein-1 as serum markers for interstitial lungdiseases. Am J Respir Crit Care Med 2002;165:378-81.

27. Kobayashi J, Kitamura S. KL-6: A serum marker for interstitialpneumonia. Chest 1995;108:311-5.

28. Yokoyama A, Kohno N, Hamada H, et al. Circulating KL-6 pre-dicts the outcome of rapidly progressive idiopathic pul-monary fibrosis. Am J Respir Crit Care Med 1998;158:1680-4.

29. Hirasawa Y, Kohno N, Yokoyama A, et al. KL-6, a humanMUC1 mucin, is chemotactic for human fibrosis. Am J RespirCell Mol Biol 1997;17:501-7.

30. Khalil N, O’Connor RN, Flanders KC, et al. TGF-β1, but notTGF-β2 or TGFβ3, is differentially present in epithelial cells ofadvanced pulmonary fibrosis: An immunohistochemicalstudy. Am J Respir Cell Mol Biol 1996;14:131-8.

31. Postlethwaite AE, Keski-Oja J, Moses HL, Kang AH.Stimulation of the chemotactic migration of human fibroblastsby transforming growth factor-β. J Exp Med 1987;165:251-6.

N. HINATA et al. J Med Dent Sci224