Printer Friendly

M30 Expression Demonstrates Apoptotic Cells, Correlates With in Situ End-Labeling, and Is Associated With Ki-67 Expression in Large Intestinal Neoplasms.

Apoptosis, or programmed cell death, is the death of individual cells by a genetically controlled mechanism.[1-3] In general, apoptosis performs a beneficial function for the organism and occurs in many physiological circumstances, including embryological morphogenesis, the regulation of cell numbers in adult tissues (such as the removal of senescent cells in the gastrointestinal tract and the cyclic loss of endometrial cells), and the elimination of autoreactive lymphocytes. It also occurs in a wide variety of pathologic conditions. In the case of neoplastic disease, the role of apoptosis in the growth and development of tumors has been a subject of great interest during the past decade.[1-3] As part of this interest, the counting of apoptotic cells in tissue sections of tumors has been important. It is possible to identify apoptotic cells in sections prepared with routine stains, for example, hematoxylineosin, but this requires care and expertise on the part of the observer.[2] Therefore, workers have attempted to find techniques that will demonstrate apoptotic cells in tissue sections clearly, allowing easy and unambiguous identification by light microscopy in paraffin sections.

The techniques most widely used to demonstrate apoptotic cells in paraffin sections are terminal deoxyribonucleotidyl transferase-mediated deoxyuridine triphosphate-biotin nick end-labeling (TUNEL) and in situ end-labeling (ISEL).[34] Both methods detect the DNA strand breaks characteristic of apoptosis by adding nucleotides, including deoxyuridine triphosphate labeled with biotin, to the ends of the broken DNA.[5-8] The only essential difference between the two is that TUNEL uses the enzyme terminal deoxyribonucleotidyl transferase,[6] whereas ISEL uses DNA polymerase I or its Klenow fragment.[7-9] Therefore, TUNEL can label DNA fragments with blunt ends or with ends that are either 3' recessed or 5' recessed, whereas ISEL only detects 3' recessed ends.[10] Although these techniques (particularly TUNEL) have been used extensively in the study of neoplastic and nonneoplastic disease, they are not without their disadvantages. They both mark necrotic and autolytic cells in addition to apoptotic cells.[2,3,11] They require pretreatment steps that need careful optimization, and the results depend on how these steps are performed.[5,12,13] Some false-positive reactions appear to be produced by endogenous endonucleases released during proteinase digestion; endonuclease inhibition can abolish this effect.[14] Studies into the effects of fixation and prefixation times have yielded variable results.[15-17] Prolonged fixation times can reduce the number of detectable apoptotic cells,[15] while delayed fixation can increase their number[16]; on the other hand, prolonged archival storage of paraffin blocks appears to have a negligible effect.[17] There is good evidence that the physical act of cutting the tissue to produce sections creates TUNEL reactivity.[18] Furthermore, interassay variability in TUNEL has been demonstrated.[19] Some authors believe that morphology is superior to TUNEL or ISEL for the detection of apoptosis.[1,19]

Another means of identifying apoptotic cells has been described recently.[20] It is a monoclonal antibody called M30, which recognizes a neoepitope of cytokeratin 18 (CK18) in paraffin-embedded tissue. This neoepitope is produced by caspase cleavage of CK18 during apoptosis and is not present in nonapoptotic cells.[20] An advantage over TUNEL and ISEL is the lack of M30 expression in necrotic cells. Since CK18 has wide distribution, being present in virtually all simple, ductal, and pseudostratified epithelia, the demonstration of M30 immunoreactivity could be immensely useful in the investigation of apoptosis in tissue sections.

This study used adenomas and adenocarcinomas of the large intestine with 2 aims: to perform a direct comparison between M30 and ISEL to determine the relationship between them, and to compare the results with Ki-67 expression.

MATERIALS AND METHODS

One hundred cases of sporadic adenoma and adenocarcinoma of the large intestine from 100 different patients were selected from the archives of the Royal Air Force Institute of Pathology (Halton, Buckinghamshire, United Kingdom), the Royal Hospital Haslar (Gosport, Hampshire, United Kingdom), and St Mark's Hospital (Harrow, Middlesex, United Kingdom). Five-micrometer-thick sections were cut from paraffin blocks of the formalin-fixed tissue.

In situ end-labeling was performed according to a protocol based on the techniques of Ansari et al[7] and Wijsman et al.[9,12] After being dewaxed in xylene and hydrated in ethanol and water, endogenous peroxidase activity was abolished by placing the slides in a 1% solution of hydrogen peroxide in distilled water for 10 minutes. The slides were rinsed and incubated at room temperature in a 0.5% solution of pepsin. This solution was prepared by dissolving lyophilized pepsin 1:60 000 (2500-3500 units/mg protein) (Sigma, Poole, Dorset, United Kingdom) in 0.1 mol/L hydrochloric acid. Following a thorough rinsing, the slides were allowed to dry (about 20 minutes). The solution for the nucleotide-polymerase reaction was 50 mmol/L Tris-HCl, pH 7.4 (Sigma); 5 mmol/L magnesium chloride; 10 mmol/L 2-mercaptoethanol; 0.005% bovine serum albumin; 0.02 mmol/L deoxyadenosine triphosphate, deoxycytosine triphosphate, and deoxyguanosine triphosphate (Sigma); 0.004 mmol/L bio-21-dUTP (Clontech, Palo Alto, Calif); and 30 units/mL of DNA polymerase I (Sigma) in distilled water. The sections were covered with this solution and incubated for 90 minutes at 37 [degrees] C. The biotin-labeled nucleotide was detected with streptavidin-biotin (Dakopatts, Glostrup, Denmark) and 3,3'-diaminobenzidine-hydrogen peroxide. The sections were then counterstained with hematoxylin, dehydrated, and mounted. Each batch of slides included a positive control (a colonic adenocarcinoma) and a negative control (the same carcinoma treated in the same way, except that the polymerase was omitted from the nucleotide solution). The sections were examined microscopically at high power. At least 2000 cells per case were counted; positive cells were defined as cells with nuclei, whole or fragmented, showing morphology consistent with apoptosis.[12] Zones of necrosis and apoptotic fragments in glandular lumina were not counted. The number of positive cells was expressed as a percentage of the total number of epithelial cells counted.

M30 immunostaining was performed as follows. Sections were dewaxed in xylene and hydrated in ethanol. Antigen retrieval was performed by placing the slides in plastic Coplin jars containing target retrieval solution (Dakopatts) and heating in a microwave oven (Energy Beam Sciences model H2800, The Laboratory Microwave Company, Agawam, Mass). The temperature of the retrieval solution was raised to 98 [degrees] C, and then this temperature was maintained for 15 minutes. The sections were allowed to cool in this solution for about 15 minutes and then washed. Hydrogen peroxide solution from the EnVision kit (Dakopatts) was applied for 7 minutes, then the slides were washed and incubated with serum-free protein block (Dakopatts). The primary antibody solution was a 1:150 dilution of M30 (Roche Diagnostics, Lewes, Sussex, United Kingdom) in Tris-buffered saline (Dakopatts) with 0.05% Tween 20; it was applied for 1 hour at room temperature. After rinsing, labeled EnVision polymer (Dakopatts) was added for 30 minutes, and the presence of bound antibody was demonstrated with the substrate-chromogen solution (hydrogen peroxide and 3,3'-diaminobenzidine) from the EnVision kit. The slides were counterstained with hematoxylin, dehydrated, and mounted. Each batch of slides included a positive control (a colonic adenocarcinoma). The negative control was the same carcinoma treated identically, except that the primary antibody was omitted. In an additional experiment, the primary antibody was replaced with isotypic mouse immunoglobulin IgG2b (Sigma) at the same concentration. The sections were examined at high power, with at least 1000 cells being counted in each case. Cells exhibiting immunoreactivity for M30 were expressed as a percentage of the total number of epithelial cells counted.

Immunohistochemistry for Ki-67 followed standard procedures. Tissue sections were dewaxed and hydrated as described for M30. Antigen retrieval was performed by placing the slides in a citrate buffer (0.01 mol/L citric acid with 80 g of sodium hydroxide per liter, adjusted to pH 6.0) and heating in a microwave for 15 minutes at 95 [degrees] C. After cooling and washing, a 3% solution of hydrogen peroxide in distilled water was applied, followed by washing and incubation with the blocking reagent from the LSAB kit (Dakopatts) for 5 minutes. The primary antibody solution was a 1:25 dilution of Ki-67 monoclonal antibody (Dakopatts) in phosphate-buffered saline; slides were incubated in antibody solution for 3 hours at room temperature. The slides were rinsed, incubated with the link antibody, and then incubated with the streptavidin peroxidase solution from the LSAB kit. The presence of bound antibody was demonstrated with 3,3'-diaminobenzidine-hydrogen peroxide solution. The slides were counterstained with hematoxylin, dehydrated, and mounted. Each batch of slides included a positive control and a negative control, as for M30 and ISEL. At least 1000 cells were counted in each case. Cells with nuclei showing immunoreactivity for Ki-67 were expressed as a percentage of the total number of epithelial cells counted.

The data were analyzed using Stat-100 (Biosoft, Cambridge, United Kingdom). P values are quoted to 2 decimal places; P [is less than or equal to] .05 was considered significant.

RESULTS

Nineteen of the 100 cases (12 adenomas, 7 carcinomas) failed to give satisfactory results with ISEL, either because of nonspecific reactions in normal nuclei or lack of reaction despite the presence of apoptotic cells morphologically. These cases were excluded from the study, leaving 81 cases for analysis. The mean age of the 81 patients was 62 years. Thirty-one cases involved adenomas; the mean age of patients in this group was 60 years, and the male-female ratio was 18:13. Sixteen adenomas were from the right side of the large intestine, and 15 were from the left. Mild dysplasia was present in 17 adenomas, moderate dysplasia was noted in 11, and severe dysplasia was identified in 3. Fifty cases represented carcinomas; the mean age of patients with carcinomas was 63 years, and the male-female ratio was 34:16. Twenty-three carcinomas were from the right side, and 27 were from the left. Five cases were Dukes stage A, 19 were Dukes stage B, and 21 were Dukes stage C. In 5 patients the Dukes stage was unknown. Histologically, all the carcinomas were adenocarcinomas.

The results of the ISEL and M30 experiments are given in the Table and a scatterplot of the counts is shown in Figure 1. Simple linear regression showed a strong positive correlation between ISEL and M30 counts (t = 7.44, P = .00). The residuals around the regression line were normally distributed. The distributions of the counts were skewed markedly to the right, and this skewness persisted even after logarithmic transformation. Therefore, nonparametric tests of correlation were used. The Spearman rank correlation coefficient (p) was 0.80 (2-sided probability P = .00). The Kendall correlation coefficient ([Tau]), adjusted for ties and continuity corrected, was 0.60 (P = .00).

[Figure 1 ILLUSTRATION OMITTED]
Counts of In Situ End Labeling (ISEL) and M30

 All Cases Adenomas

 ISEL M30 ISEL M30

n 81 81 31 31
Mean 0.98 1.30 0.64 0.98
95% CI for mean 0.47-1.49 0.36-2.24 0.34-0.94 0.06-1.91

 Carcinomas

 ISEL M30

n 50 50
Mean 1.19 1.50
95% CI for mean 0.28-2.01 0.06-2.90


Immunoexpression of M30 was generally easier to interpret than ISEL, since cells giving ambiguous signals were rare with M30, but were common with ISEL. The background tended to be "clean" with M30 (Figure 2), although some nonspecific reactivity in mucin caused slight problems in mucinous adenocarcinomas. The range of appearances was similar to that described by Leers et al.[20] Cells with condensed chromatin and cell fragments with pyknotic nuclear material showed granular or diffuse cytoplasmic staining. Some apoptotic bodies apparently at an advanced stage of degradation were negative. Granular reactivity was seen in the cytoplasm of some cells with normal nuclei; these cells were assumed to be at an early stage of apoptosis and were counted as positive.

[Figure 2 ILLUSTRATION OMITTED]

The negative control slides showed no cytoplasmic staining, but a weak nonspecific nuclear signal was observed in 1% to 2% of cells. This signal did not prove troublesome, because only cytoplasmic reactivity is significant in this context. The slides treated with mouse IgG2b showed identical results to the negative control slides.

Despite the significant correlation between M30 expression and ISEL, inspection of Figure 1 shows a wide scatter around the regression line. There is a particularly prominent outlier; it was a signet ring adenocarcinoma from a 73-year-old woman.

Four cases did not have satisfactory Ki-67 immunoreactivity (failure of internal positive controls to react) and were excluded. The mean Ki-67 counts of the remaining 77 cases were 26.8 for adenomas, 25.0 for carcinomas, and 25.7 for all cases combined. Since the set of Ki-67 results had a different number of items than the ISEL and M30 results, the Mann-Whiney U test was used to assess the correlation between Ki-67 counts and the apoptotic indices. There was a strong positive correlation in each case (P = .01 for Ki-67 vs ISEL and for Ki-67 vs M30).

No relationship between M30 count and age, sex, site (left vs right colon), Dukes stage, degree of dysplasia (adenomas), or degree of differentiation (carcinomas) was found. Cases in which the M30 count was more than twice the ISEL count or less than half the ISEL count were considered separately, but there were no correlations with any of the variables studied.

COMMENT

The results show that ISEL and M30 immunoexpression are correlated in adenomas and adenocarcinomas of the large intestine. However, there is a considerable scatter around the regression line (Figure 1), and it is clear that in any individual case the ISEL count does not necessarily predict the M30 count with any degree of accuracy. There are 2 possible explanations for differences between M30 and ISEL values in a given case. First, the techniques recognize different aspects of the apoptotic pathway. Second, the results of both tests depend on technique. In situ end-labeling in particular has been shown to be sensitive to the pretreatment step, which requires optimization for best results.[12] The overall counts are consistent with previous reports in the literature, which describe mean values of around 1% to 2% for adenomas and adenocarcinomas of the large intestine.[21-25] Some studies have shown higher[25-27] or lower[28,29] values, at least for some types of lesion; this phenomenon might be explained by differences in case selection, technical procedures, or interpretation of the histologic appearances.

In their report on M30, Leers et al[20] compared M30 expression with TUNEL and found that M30 expression was an earlier event in the apoptotic pathway than TUNEL positivity, and that TUNEL positivity persisted in late apoptosis with complete nuclear disintegration, although M30 became negative. However, a quantitative comparison between TUNEL counts and M30 counts was not presented in their article. My results are consistent with the onset of M30 expression preceding ISEL positivity, the latter remaining positive after M30 expression is no longer detectable.

Ki-67 is expressed in proliferating cells and is widely used to determine the growth fraction in tissue samples.[30] Although Ki-67 is superior to other proliferation markers in that it is not expressed during excision-repair of DNA,[31] it can be expressed in cells even when DNA synthesis is blocked, and can thus overestimate the proportion of actively cycling cells.[32] Studies have not shown any relationship between Ki-67 expression and prognosis in colorectal carcinoma.[30,33-35] Like other studies,[23,24,29] my results showed a correlation between the expression of Ki-67 and the apoptotic count, using both ISEL and M30.

In conclusion, M30 immunoexpression and ISEL positivity correlate strongly. Since M30 immunohistochemistry is technically simpler and easier to interpret than ISEL, it has the potential to become the method of choice for demonstrating apoptosis in formalin-fixed tissue for cells containing cytokeratin 18.

This study was supported by the Education and Research Fund of the Royal Hospital Haslar (Gosport, Hampshire, United Kingdom). The author is an Honorary Research Fellow at St Mark's Hospital (Harrow, Middlesex, United Kingdom) and thanks Ian Talbot, MD, FRCPath, for his comments and support.

References

[1.] Butler LM, Hewett PJ, Fitridge RA, Cowled PA. Deregulation of apoptosis in colorectal carcinoma: theoretical and therapeutic implications. Aust N Z J Surg. 1999;69:88-94.

[2.] Cummings MC, Winterford CM, Walker NI. Apoptosis. Am J Surg Pathol. 1997;21:88-101.

[3.] Staunton MJ, Gaffney EF. Apoptosis: basic concepts and potential significance in human cancer. Arch Pathol Lab Med. 1998;122:310-319.

[4.] Gorczyca W, Bruno S, Darzynkiewicz RJ, et al. DNA strand breaks during apoptosis: their early in situ detection by the terminal deoxynucleotidyl transferase and nick translation systems and prevention by serine protease inhibitors. Int J Oncol. 1992;1:639-648.

[5.] Mainwaring PN, Ellis PA, Detre S, Smith IE, Dowsett M. Comparison of in-situ methods to assess DNA cleavage in apoptotic cells in patients with breast cancer. J Clin Pathol. 1998;51:34-37.

[6.] Gavrieli Y, Sherman Y, Ben-Sasson SA. Identification of programmed cell death in situ via specific labeling of DNA fragmentation. J Cell Biol. 1992;119: 493-501.

[7.] Ansari B, Coates PJ, Greenstein BD, Hall PA. In situ end-labeling detects DNA strand breaks in apoptosis and other physiological and pathological states. J Pathol. 1993;170:1-8.

[8.] Gold R, Schmied M, Rothe G, et al. Detection of DNA fragmentation in apoptosis: application of in situ nick translation to cell culture systems and tissue sections. J Histochem Cytochem. 1993;41:1023-1030.

[9.] Wijsman JH, Jonker RR, Keijzer R, et al. A new method to detect apoptosis in paraffin sections: in situ end labeling of fragmented DNA. J Histochem Cytochem. 1993;41:7-12.

[10.] Mundle SD, Gao XZ, Khan S, et al. Two in situ labeling techniques reveal different patterns of DNA fragmentation during spontaneous apoptosis in vivo and induced apoptosis in vitro. Anticancer Res. 1995;15:1895-1904.

[11.] Hayashi R, Ito Y, Matsumoto K, et al. Quantitative differentiation of both free 3'-OH and 5'-OH DNA ends between heat-induced apoptosis and necrosis. J Histochem Cytochem. 1998;46:1051-1059.

[12.] Carr NJ, Talbot IC. In situ end labeling: effect of proteolytic enzyme pretreatment and hydrochloric acid. Mol Pathol. 1997;50:160-163.

[13.] Gold R, Schmied M, Giegerich G, et al. Differentiation between cellular apoptosis and necrosis by the combined use of in situ tailing and nick translation techniques. Lab Invest. 1994;71:219-225.

[14.] Stahelin BJ, Marti U, Solioz M, et al. False positive staining in the TUNEL assay to detect apoptosis in liver and intestine is caused by endogenous nucleases and inhibited by diethyl pyrocarbonate. Mol Pathol. 1998;51:204-208.

[15.] Davison FD, Groves M, Scaravilli F. The effects of formalin fixation on the detection of apoptosis in human brain by in situ end-labeling of DNA. Histochem J. 1995;27:983-988.

[16.] Tateyama H, Tada T, Hattori H, Murase T, Li WX, Eimoto T. Effects of prefixation and fixation times on apoptosis detection by in situ end-labeling of fragmented DNA. Arch Pathol Lab Med. 1998;122:252-255.

[17.] Bardales RH, Xie SS, Hsu SM. In situ DNA fragmentation assay for detection of apoptosis in paraffin-embedded tissue sections: technical considerations. Am J Clin Pathol. 1997;107:332-336.

[18.] Sloop GD, Roa JC, Delgado AG, Balart JT, Hines MO 3rd, Hill JM. Histologic sectioning produces TUNEL reactivity: a potential cause of false-positive staining. Arch Pathol Lab Med. 1999;123:529-532.

[19.] Hawkins NJ, Lees J, Ward RL. Detection of apoptosis in colorectal carcinoma by light microscopy and in situ end labeling. Anal Quant Cytol Histol. 1997;19:227-232.

[20.] Leers MPG, Kolgen W, Bjorklund V, et al. Immunocytochemical detection and mapping of a cytokeratin 18 neo-epitope exposed during early apoptosis. J Pathol. 1999;187:567-572.

[21.] Moss SF, Scholes JV, Holt PR. Abnormalities of epithelial apoptosis in multistep colorectal neoplasia demonstrated by terminal deoxyuridine nick end labeling. Dig Dis Sci. 1996;41:2238-2247.

[22.] Ikenaga M, Takano Y, Saegusa M, et al. Apoptosis of colon cancers assessed by in situ DNA nick end-labeling method. Pathol Int. 1996;46:33-37.

[23.] Hao X, Du M, Bishop AE, Talbot IC. Imbalance between proliferation and apoptosis in the development of colorectal carcinoma. Virchows Arch. 1998;433: 523-527.

[24.] Koike M. Significance of spontaneous apoptosis during colorectal tumorigenesis. J Surg Oncol. 1996;62:97-108.

[25.] Ikenaga M, Takano Y, Ohtani Y, et al. Low levels of apoptosis and proliferative activity in colorectal villous tumors: comparison with tubular adenomas. Pathol Int. 1998;48:453-459.

[26.] Arai T, Kino I. Role of apoptosis in modulation of the growth of human colorectal tubular and villous adenomas. J Pathol. 1995;176:37-44.

[27.] Tsujitani S, Shirai H, Tatebe S, et al. Apoptotic cell death and its relationship to carcinogenesis in colorectal carcinoma. Cancer. 1996;77:1711-1716.

[28.] Tanimoto T, Tanaka S, Haruma K, et al. MUC1 expression in intramucosal colorectal neoplasms. Oncology 1999;56:223-231.

[29.] Takano Y, Saegusa M, Ikenaga M, Mitomi H, Okayasu I. Apoptosis of colon cancer: comparison with Ki-67 proliferative activity and expression of p53. J Cancer Res Clin Oncol. 1996; 122:166-170.

[30.] Kubota T, Petras RE, Easley KA, et al. Ki-67-determined growth fraction versus standard staging and grading parameters in colorectal carcinoma: a multivariate analysis. Cancer. 1992;70:2602-2609.

[31.] McCormick D, Chong H, Hobbs C, et al. Detection of the Ki-67 antigen in fixed and wax-embedded sections with the monoclonal antibody MIB1. Histopathology. 1993;22:355-360.

[32.] van Oijen MGCT, Medema RH, Slootweg PJ, et al. Positivity of the proliferation marker Ki-67 in non-cycling cells. Am J Clin Pathol. 1998;110:24-31.

[33.] Handa K, Yamakawa M, Takeda H, Kimura S, Takahashi T. Expression of cell cycle markers in colorectal carcinoma: superiority of cyclin as an indicator of poor prognosis. Int J Cancer. 1999;84:225-233.

[34.] Kyzer S, Gordon PH. Determination of proliferative activity in colorectal carcinoma using monoclonal antibody Ki67. Dis Colon Rectum. 1997;40:322-325.

[35.] Jansson A, Sun XF. Ki-67 expression in relation to clinicopathological variables and prognosis in colorectal adenocarcinomas. APMIS. 1997;105:730-734.

Accepted for publication July 14, 2000.

From the Department of Pathology, Royal Hospital Haslar, Gosport, Hampshire, United Kingdom.

Reprints: Norman John Carr, FRCPath, Department of Cellular Pathology, Level E, South Block, Southampton General Hospital, Tremona Road, Southampton, Hampshire SO16 6YD, United Kingdom.
COPYRIGHT 2000 College of American Pathologists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2000 Gale, Cengage Learning. All rights reserved.

 Reader Opinion

Title:

Comment:



 

Article Details
Printer friendly Cite/link Email Feedback
Author:Carr, Norman John
Publication:Archives of Pathology & Laboratory Medicine
Geographic Code:4EUUK
Date:Dec 1, 2000
Words:3680
Previous Article:Loss of High-Molecular-Weight Cytokeratin Antigenicity in Prostate Tissue Obtained by Transurethral Resections.
Next Article:Characteristics of Salivary Diffuse Infiltrative Lymphocytosis Syndrome in West Africa.
Topics:


Related Articles
Anaplastic Large Cell Lymphoma Arising in Bone.
CD13-Positive Anaplastic Large Cell Lymphoma of T-Cell Origin--A Diagnostic and Histogenetic Problem: A Case Report and Review of the Literature.
Hoechst 33342-Induced Apoptosis Is Associated With Intracellular Accumulation of E2F-1 Protein in BC3H-1 Myocytes and HL-60 Cells.
Aberrant Localization of the Neuronal Class III [Beta]-Tubulin in Astrocytomas.
Correlation of Ki-67 and p53 With the New World Health Organization/International Society of Urological Pathology Classification System for...
An Immunostaining Panel for Diagnosis of Malignancy in Mucinous Tumors of the Pancreas.
Glial cell-derived neurotrophic factor expression in normal human lung and congenital cystic adenomatoid malformation.
Increased apoptosis in infiltrating mononuclear cells of colorectal cancer: a mechanism for tumor escape.
Phosphorylated histone H3, Ki-67, p21, fatty acid synthase, and cleaved caspase-3 expression in benign and atypical granular cell tumors.
Basal/myoepithelial cells in chronic sinusitis, respiratory epithelial adenomatoid hamartoma, inverted papilloma, and intestinal-type and...

Terms of use | Copyright © 2014 Farlex, Inc. | Feedback | For webmasters