The zyxwvutsrqponm Laryngoscope Lippincott-Raven Publishers, Philadelphia zyxwvutsrqp 0 1998 The American Laryn ological Rhinological and Otological iociety, Ihc. Differentiation and Cathepsin D Expression in Human Oral Tumors Miriam M. Brysk, PhD; Gang Lei, MD; Karen Adler-Storthz,PhD; Zhuo Chen, PhD; Toshio Horikoshi, PhD; Henry Brysk, PhD; Stephen K. Tyring, MD, PhD; Istvan Arany, PhD ObjectiveslHypothesis: This study aimed to ascer- tain whether cathepsin D expression could be related to the stage of differentiation of oral tumors. Study Design: Human oral biopsies of 10 zyxwvut squamous cell car- cinomas and of the corresponding perilesional normal tissues were used. The tumors had all been clinically graded as advanced stage but nonmetastatic; five were classified histopathologically as poorly differen- tiated. Methods: The gene expression of cathepsin D and keratin K13 in the biopsies was measured by re- verse transcription polymerase chain reaction. Ratios of tumor-to-control readings helped compensate for sample variability. Results: Keratin K13, as a suprabasal cell marker, tended to confirm the histo- logical grading of the tumors (but was not otherwise useful in distinguishing tumors from normal tissue). Substantial overexpression of cathepsin D was found in the poorly differentiated tumors. Conclusions: Cathepsin D overexpression is considered a prognos- tic indicator of metastasis. In this sample, it was also associated with dedifferentiation. Cathepsin D might serve as a valuable gauge in clinical exploration of the connection between dedifferentiation and metastasis. Key Words: Oral tumors, squamous cell carcinoma, dif- ferentiation, cathepsin D, keratin K13. Laryngoscope, 108: 1234-1237,1998 INTRODUCTION Cathepsin D is an intracellular aspartic protease, ubiquitously expressed in mammalian cells and normally localized to the lysosomes.1 The biosynthesis and protein structure of cathepsin D have been extensively studied.192 Human cathepsin D is synthesized as a 52-kD proenzyme that undergoes subsequent proteolytic processing to pro- From the Departments of Dermatology (M.M.B., G.L., T.H., H.B., s.K.T.), Microbiology and Immunology (M.M.B., zyxwvutsr G.L., s.K.T., LA.), and Human Biolog- ical Chemistry and Genetics (M.M.B.), University of Texas Medical Branch, Galveston, and the Department of Basic Sciences, University of Texas-Houston Dental Branch (K.A.-s., z.c.), Houston, Texas; and the Cos- metics Laboratory, Kanebo Ltd. (T.H.), Odawara, Kanagawa, 250 Japan. Supported in part by National Institutes of Health grants R55-DE- OD10846 (K.A.-s.), R03-DE11869 (z.c.), and R01-DE08477 (M.M.B.). Editor’s Note: This Manuscript was accepted for publication April 17, 1998. Send Correspondence to Miriam M. Brysk, PhD, Department of Der- matology, University of Texas Medical Branch, Galveston, TX 77555-0783, U.S.A. duce the activated 48-kD intermediate form; the latter is then cleaved into the mature 34-kD form and a 14-kD fragment .34 Aside from its role in protein catabolism via the degradation of endocytosed proteins, cathepsin D has at- tracted clinical attention because of its overexpression in a variety of diseases.2 In particular, high levels of cathep- sin D in tumors have been associated with a poor progno- sis of survival. Most such studies have dealt with breast cancer and have concluded that cathepsin D is an inde- pendent prognostic indicator, especially useful for predic- tion of metastasis.5-7 A recent clinical study of laryngeal squamous cell carcinomas (SCCs) reached a similar con- clusion.8 Histological identification of poor differentiation has traditionally been considered a bad omen for survival in oral cancer, although this prognosis is only reliable with more detailed morphological examination.9 We report here that cathepsin D gene expression in head and neck SCCs is related to tumor dedifferentiation. MATERIALS zyxw AND WTHODS Tissue Samples For a previous study of epidermal growth factor receptor in oral cancers,lO biopsies from squamous cell carcinomas had been collected from 10 patients, together with biopsies from adjacent (histologically normal-appearing) areas of the same anatomic sites. The tumors were from a variety of head and neck locations, as enumerated in Table I. The clinical stage of all the tumors was classified as advanced (T3 or T4) with little or no node involve- ment (NO or N1) and with no evidence of metastasis at the time (MO). Histopathology staged half the tumors as poorly differenti- ated, the others as moderately or well differentiated. They thus represent a test set with near-uniform TNM staging but with variation in differentiation grading. Quantitative Analysis of mRNA Levels Total RNAs were extracted from tumor and normal perile- sional biopsies using RNA STAT-60 isolation reagent, following the manufacturer’s protocol (Tel-Test, Friendswood, TX). To con- currently determine different gene transcripts, one pg of RNA was subjected to cDNA synthesis at 42°C for 1 h, using Super- script RNase H- reverse transcriptase (Gibco BRL, Grand Is- land, NY) and random hexamer priming (Promega, Madison, z Laryngoscope zyxwvutsrq 108: August 1998 1234 Brysk et al.: Tumor Differentiation and Cathepsin D