ABNORMAL EXPRESSION PATTERN OF CYCLIN E IN TUMOUR CELLS Fredrik ERLANDSSON 1 , Carolina W¨ AHLBY 2 , Susanna EKHOLM-REED 1 , Ann-Cathrin HELLSTR ¨ OM 3 , Ewert BENGTSSON 2 and Anders ZETTERBERG 1 * 1 Department of Oncology-Pathology, Karolinska Institutet, Stockholm, Sweden 2 Centre for Image Analysis, Uppsala University, Uppsala, Sweden 3 Department of Gynecological Oncology, Radiumhemmet, Karolinska Hospital, Stockholm, Sweden The expression pattern of cyclin E during the cell cycle was studied in normal and tumour cells in culture and in tumour biopsies. This pattern was found to be abnormal in tumour cells. A triple immunostaining protocol, digital microscopy and image analysis were used to find the position of the individual cells in the cell cycle and to measure the nuclear cyclin E levels. In normal cells, the number of cyclin E-posi- tive cells decreased rapidly when the cells entered the S- phase. In the tumour cell lines, cyclin E was not downregu- lated in early S-phase, as in normal cells. Instead the number of cyclin E-positive cells remained high throughout S-phase, and the cyclin E staining intensity per cell often increased during S-phase. In about half of the analysed tumour cell lines, many cells stained positive for cyclin E even in the G 2 -phase. This abnormal expression over the cell cycle of cyclin E was also found in tumour biopsies from cervical, breast and prostatic carcinomas, even though it varied greatly between individual tumours. In some tumours, the expression pattern of cyclin E was similar to that of normal cells in culture, whereas in others high cyclin E levels could be seen in S-phase cells, as in the transformed cell lines. A high percentage of cells expressing cyclin E during S- or G 2 -phase was found to be related to poor outcome (p < 0.025) in a small group of cervical carcinoma patients (n  12). © 2003 Wiley-Liss, Inc. Key words: cyclin E; cyclin A; transformation; cell cycle; immuno- fluorescence staining Eukaryotic cells are driven through the cell cycle by successive activation and inactivation of cyclin-dependent kinases (CDKs). The CDKs are activated through association with their regulatory subunits, the cyclins. The first cyclins were identified in marine invertebrates as proteins oscillating during the cell cycle. 1 Human cyclin E was identified through screening of human cDNA librar- ies for genes that could complement the mutated cyclin in a strain of S. cereviciae. 2,3 The different cyclins have temporally distinct and highly regulated patterns of expression, i.e., they are synthe- sized and degraded at specific stages of the cell cycle. Cyclin E appears in late G 1 after passage through the restriction point. 4 The level of cyclin E peaks in late G 1 and disappears again in early S. 2,4,5 Cyclin A appears in the nucleus precisely at the G 1 /S border, 6 accumulates throughout the rest of interphase and disap- pears at the beginning of mitosis. 7,8 Both cyclin E and cyclin A control the progression through the cell cycle primarily by acti- vating CDK2. Cyclin E and cyclin A both exist in 2 isoforms with high homology, designated cyclin E1 and E2, and cyclin A1 and A2, respectively. 9 –12 No major differences in expression or function between cyclin E1 and E2 have been found, and their expression have been assumed to be governed by the same molecular cir- cuitry. 13 Increased expression of cyclin E has been found in several types of tumours. 14 –26 However, these studies have only shown either an increased amount of cyclin E in tumour tissue using Western blot techniques, or an increased percentage of cyclin E-positive cells using immunohistochemistry or flow cytometry. Only measuring the total amount of cyclin E or the number of cells containing cyclin E in a tissue does only give limited information about true abnormalities in expression pattern during the cell cycle. Elevated amounts of cyclin E, or increased numbers of cells staining posi- tive for cyclin E, could simply be due to increased proliferation in the analysed tissue samples. However, cyclin E might also be abnormally expressed over the cell cycle, i.e., expressed during the wrong phase of the cell cycle. The cyclin E expression pattern over the cell cycle has to our knowledge never been investigated in solid tumours in vivo. Only Western blot studies on synchronized cells in culture 5 and flow cytometric analysis of cell lines and blood samples from patients with leukaemia 24,25 have been carried out. We therefore developed a technique that enabled a high- resolution analysis of the cyclin E expression over the cell cycle. The technique was first validated on cell lines and then applied to biopsies from cervical, prostatic and breast carcinomas. We found that by using a triple immunofluorescence staining procedure, in which cyclin E, cyclin A and incorporated BrdU are detected simultaneously in individual cells, it is possible to accu- rately determine the number of cells staining positive for cyclin E in each phase of the cell cycle. This sensitive single-cell technique permits the detection of true disturbances in the cyclin E expres- sion pattern over the cell cycle in unperturbed cell populations and is independent of intercellular variations in cell cycle progression. We utilized this technique to study 5 different normal cell cultures and 9 different tumour cell lines. In normal cells cyclin E was rapidly degraded in early S-phase. In the tumour cell lines, cyclin E levels remained high throughout S- and sometimes also G 2 - phase. Of great importance was the finding that the abnormal pattern of cyclin E expression was also found in tumour biopsies from human cervical, prostatic and breast carcinomas, indicating the generality of this abnormality in tumours. Furthermore, a high percentage of cells expressing cyclin E during S- or G 2 -phase was found to be related to poor outcome in cervical carcinomas. Also, high levels of cyclin E during S-phase could be seen in highly malignant prostatic and breast carcinomas. However, studies on large patient groups are required to prove the prognostic impor- tance of this cell cycle abnormality. Such studies are underway. MATERIALS AND METHODS Some of the experimental procedures used, such as handling of cell lines, immunofluorescence staining, controls and image acqui- sition, have been previously described in detail. 6 Grant sponsor: Swedish Cancer Society; Grant number: 0046-B01- 33XAC; Grant sponsor: Cancerf¨ oreningen in Stockholm; Grant sponsor: Karolinska Institutet; Grant sponsor: Svenska L¨ akares¨ allskapet; Grant sponsor: Foundation for Strategic Research through the Visual Information Technology program. *Correspondence to: Department of Oncology-Pathology, Karolinska Institutet, CCK R8:04, Karolinska Hospital, 171 76 Stockholm, Sweden. Fax: +46-8-321047. E-mail: anders.zetterberg@onkpat.ki.se Received 8 April 2002; Revised 5 July 2002; Accepted 25 September 2002 DOI 10.1002/ijc.10949 Int. J. Cancer: 104, 369 –375 (2003) © 2003 Wiley-Liss, Inc. Publication of the International Union Against Cancer