| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Departments of Obstetrics and Gynecology, and Pathology, University Hospital, Zurich, Switzerland, Beckman Laser Institute and Laser Microbeam and Medical Program, and Department of Medicine, Division of Hematology/Oncology, University of California, Irvine, Irvine, California.
Address reprint requests to: Vickie J. LaMorte, PhD, Beckman Laser Institute, 1002 Health Sciences Road East, Irvine, CA 92612; E-mail: lamorte{at}uci.edu.
 |
ABSTRACT |
|---|
TOP
 ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: Fluorescence-based immunodetection of the promyelocyte protein and the small ubiquitin-like modifier protein was performed on paraffin-embedded and histopathologically graded human uterine cervical tissues. Quantitative measurements of the size and number of the promyelocyte protein-containing nuclear bodies were made and statistically analyzed.
RESULTS: We found that promyelocyte protein-containing nuclear bodies exhibit changes in both size and number throughout the continuum of cervical intraepithelial neoplasia (CIN) and cervical squamous cell carcinoma. An increase in number and size of the bodies occurs with progression from normal to CIN I/CIN II. In CIN III, two new subcategories of nuclear body are present with distinctly different promyelocyte protein patterns, with the type B CIN III losing the small ubiquitin-like modifier protein partnership. In squamous cell carcinoma, we see the loss of this colocalization in both well and poorly differentiated tumors, with a distinctly different promyelocyte protein pattern. Well-differentiated tumors have bigger nuclear bodies that are more numerous than those of the poorly differentiated tumors.
CONCLUSION: These data support the use of promyelocyte and small ubiquitin-like modifier proteins as a cytodiagnostic marker that parallels cervical cancer progression.
Cervical squamous cell carcinomas originate from a multilayered cervical epithelium and develop progressively over the course of years. The tissue goes through a spectrum of hyperplastic changes classified clinically as cervical intraepithelial neoplasia (CIN), ranging from CIN I (mild dysplasia), followed by CIN II (moderate dysplasia), to CIN III (severe dysplasia), where it may eventually result in invasive cervical cancer. Although human papillomavirus (HPV) is considered to be related to tumor induction,1,2 the molecular mechanisms that regulate the induction and progression of cervical squamous cell carcinoma remain an unsolved problem of great clinical relevance. In this study, we sought to examine the association between the dynamics of the promyelocyte protein-containing nuclear body as a molecular marker and the progression to squamous cell carcinoma by histopathologic staging.
The promyelocyte protein (PML) (see reviews3,4) was identified first in the pathogenesis of acute promyelocytic leukemia. This type of leukemia is categorically characterized by a chromosomal translocation t(15;17) that fuses the promyelocyte gene to the retinoic acid receptor 
gene. In normal cells, PML is found in discrete nuclear structures known as nuclear bodies or promyelocyte oncogenic domains. In acute promyelocytic leukemia cells, PML is displaced and presents in a microspeckled pattern. This altered pattern is thought to disrupt the normal PML function. Covalent modification to a 101-amino-acid protein, the small ubiquitin-like modifier protein (SUMO-1) (see review5), is thought to be a prerequisite for PML to maintain the nuclear body and for subsequent localization of other protein components of the body.6 In contrast, it has been shown that SUMO-1 modification of PML is not necessary for the PML to target the nuclear body.7 Although many studies have been done, a defined function has yet to be assigned to PML and its corresponding nuclear body.
Immunocytochemical studies that assist in the diagnosis of the progression or extent of the disease are prevalent in the clinical setting, as is the case of breast cancer. To date, although some markers for cervical neoplasia have been reported, their expression is not a footprint for a particular stage of the disease.8–11 Immunohistochemical detection of PML expression as a marker for disease progression has focused mainly on hematopoietic malignancies, particularly acute promyelocytic leukemia. Although reports on PML in other solid tumor cancers have been documented,12,13 systematic mapping of the changes in PML throughout the progression of the cancer from CIN to squamous cell carcinoma has not been investigated, nor has its colocalization with SUMO-1. Those studies also relied on immunohistochemical findings that are graded on the presence or absence of a colorimetric substrate in the nucleus. No spatial or quantitative information at the molecular level of the PML-containing nuclear bodies can be determined from those data. Consequently, we developed fluorescence-based methodologies for detecting the PML-containing nuclear bodies. Here, we report the dynamic changes in the PML-containing nuclear bodies, specifically their number and size and the change in PML partnership with SUMO-1, to mark the different stages of cervical neoplasia to squamous cell carcinoma.
|
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Resection specimens of 12 patients with normal tissue, 15 patients with CIN I/CINII, 18 patients with CIN III, 11 patients with well-differentiated cervical squamous cell carcinoma, and nine patients with poorly differentiated cervical squamous cell carcinoma, and 10 pairs of primary tumors with corresponding metastasis were analyzed. All specimens were fixed by standard methods in phosphate-buffered 4% formaldehyde solution and embedded in paraffin for histologic and immunofluorescence detection.
Sections were cut (5-µm thick), floated on poly-L-lysine–coated glass microscope slides, and air dried overnight at room temperature. Sections from paraffin blocks were dewaxed in Histo-Clear (National Diagnostics, Atlanta, GA) and rehydrated through graded alcohols to deionized water. Standard hematoxylin and eosin staining was performed. For immunodetection, slides were rinsed in phosphate-buffered saline, and antigen retrieval (BioGenex, San Ramon, CA) was performed by microwave pressure cooking (30 minutes). Nonspecific binding was blocked by 20% normal swine serum phosphate-buffered saline. Sections were incubated with an affinity-purified rabbit polyclonal IgG (a gift of Dr. J. Dyck) raised against amino acids 1–14 of the PML protein for 2 hours at 37C.14 Rabbit IgG was detected with a fluorescent secondary Cy3-conjugated anti-rabbit IgG (Molecular Probes, Eugene, OR). Of note, as a positive control, vascular cells display the same intensity of PML staining as basal cells of the normal epithelium. SUMO-1 was detected with a mouse monoclonal antibody (Zymed Laboratories, San Francisco, CA) and a secondary antibody conjugated to Alexa 488 fluorophore (Molecular Probes). Negative controls consisted of replacement of primary antibody with normal serum. Coverslips were mounted with vectashield mounting medium (Vector Laboratories, Burlingame, CA). Slides were stained and analyzed in triplicate.
Fluorescently labeled samples were examined by using an inverted Zeiss laser scanning microscope (LSM 410; Carl Zeiss, Jena, Germany). The objective used was an oil immersion 100x magnification Plan-Neofluar Phase 3, NA 1.3 (Carl Zeiss, Jena, Germany). Stacks of thin optical sections were obtained for each sample. The 488-nm line of an Argon laser was used for simultaneous excitation of both fluorophores (Cy-3 and Alexa 488). Simultaneously detected red and green emissions were isolated by a long-pass 610-nm filter and by a narrow band-pass (530-nm band center) filter, respectively; blue channel was used for nonconfocal phase-contrast image acquisition. Fluorescent images were pseudocolored green and red and overlaid with the phase-contrast image. The distance between two consecutive optical sections was 0.5 µm on the Z axis. Overall depth of acquisition ranged from 0 to 5 µm, covering the whole thickness of a sample. The true-focus images were generated from stacks of stored images using the original LSM 410 software.
Hematoxylin and eosin–stained images were obtained with an Olympus DP10 microscope digital camera system (Olympus, Melville, NY) and an S Plan40xPL/ 0.70 objective. Scale bars (20 µm) were added to both the phase-contrast and fluorescence image for comparison.
True-focus images of nuclear body-associated fluorescence, obtained from a confocal scanning microscope were post-processed using scientific imaging software, IPLabs (version 3.5.5, Scanalytics, Inc, Fairfax, VA). Because of the nature of sectioning of the tissue blocks, only whole cells were selected for analysis. Images were segmented, ie, target pixels were separated from background pixels based on their values. Segmentation was based on the intensity of the fluorescence signal. Segments were quantified by an area (in µm2) and by number per selected whole cell. Single pixels with high numeric values (background noise) were eliminated from counting by introducing a limiting criterion (minimal area). Random cells were selected from the appropriate hematoxylin and eosin–stained area and were analyzed quantitatively for the average number of bodies and the average area of the bodies. In all categories, an average of 54 cells per patient was analyzed with a total of 25,648 bodies measured.
Average size and area of the PML-containing nuclear bodies for each patient were calculated as described above. Eight patient groups—normal basal (n = 12), normal upper (n = 12), CINI/CIN II basal (n = 15), CIN I/CIN II upper (n = 15), type A CIN III (n = 10), type B CIN III (n = 8), well-differentiated squamous cell carcinoma (n = 11), and poorly-differentiated squamous cell carcinoma (n = 9)—were statistically compared by analysis of variance and pairwise t tests with a Bonferroni correction for multiple comparisons. Of 28 possible comparisons, six comparisons were eliminated because the comparison would not be relevant. Using the Bonferroni method of correction for multiple comparisons, a single comparison must be significant with P < .05/22 comparison or P < .002 in order to maintain an overall type 1 error rate of 5%. A separate variance estimate is used in the t test whenever indicated by significance of the Levene test for equality of variances; otherwise, a pooled estimate is used for the variance.
|
RESULTS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
–3 are representative true-focus fluorescent micrographs of PML (green) and SUMO-1 (red) with the corresponding hematoxylin and eosin–stained image (inset). Figure 4 shows the bar graph summaries of the average number (a), size (b), and percentage of SUMO-1 colocalization (c). Tables 1 and 2 summarize the results of the analysis of variance.
|
|
|
|
|
). Here, basal cells show a consistent PML pattern with an average of five round nuclear bodies per nucleus and an area of 0.25 µm2 (Figure 4 a, b). This is consistent with the average reported size of a PML body of 0.5 to 1 µm in diameter. The suprabasal and intermediate layers display less staining for PML in that there are fewer cells that are positive for PML, and those that are positive have statistically fewer bodies (three nuclear bodies per nucleus) (P < .001). These bodies are, however, comparable in size (0.22 µm2) to those in the basal layer. The superficial layer did not contain any PML. Promyelocyte protein is also present throughout the stroma. Cells of the connective tissue exhibit less intense fluorescence. All of the PML-containing nuclear bodies colocalize with SUMO-1 (Figure 1b; Figure 4c). Because SUMO-1 modifies other proteins, it is present in other regions of the nucleus in addition to the PML-containing nuclear bodies.
Morphologically a CIN I/CIN II lesion is close to normal epithelium, in that histologically the basal layer can be distinguished from the upper layers. Consequently, we analyzed the PML and SUMO-1 staining for the basal and the upper layers separately for comparison with normal epithelium. In CIN I/CIN II, PML is restricted to the lower half of the epithelium, and no longer are the nuclear bodies always perfectly round (Figure 1c). The number and size of nuclear bodies are increased (doubled) compared with normal basal and upper layers (P < .001). An average of nine bodies per nucleus (0.52 µm2) is in the nuclei of the basal layer and ten bodies per nucleus (0.54 µm2) in the nuclei in upper layers (Figure 4a). The basal and upper areas of CIN I/CIN II are not statistically different from each other. SUMO-1 appears to be further recruited to the nuclear bodies as a result of their increase in size (Figure 1d), with 95% colocalization (Figure 4c).
In grade III lesions, PML positivity extended to the full thickness of the epithelium. After analysis of the 18 patient samples, the data segregated into two subgroups, which we classified as type A (ten cases) and type B (eight cases) CIN III. Type A CIN III was statistically similar to CIN I/CIN II with an average of nine PML-containing nuclear bodies that were 0.55 µm2 in area (Figure 2a, Figure 4 a, b). The nuclear bodies remained significantly larger and more frequent than those of normal epithelium (P < .001). The size and shape showed high irregularity, with an increase in the appearance of track-like structures. These structures were not resolvable by confocal microscopy because of the limitations of pixel size. SUMO-1 continued to colocalize with the PML-containing nuclear bodies in 93% of the cells (Figure 2b, Figure 4c).
|
). The number of nuclear bodies was, on average, five per nucleus with an average size of 0.33 µm2 (Figure 4 a,b). Both the area and number are comparable to those of normal epithelium (P value not significant, see Tables 1 and 2). The number of nuclear bodies is significantly different from CIN I/CIN II and type A CIN III (P < .01 and P < .001, respectively). The decreased area is statistically significantly different from CIN I/CIN II upper (P < .05) and very close to significance with P values of .005 and .004, respectively, compared with CIN I/CIN II basal and type A CIN III. In type B CIN III, as PML body size and number were decreased, SUMO-1 colocalization decreased to 49% of the cells, clearly differentiating it from type A CIN III.
Eleven cases of well-differentiated tumors were analyzed. Here, PML body size and number were increased (Figure 3a). There were on average nine nuclear bodies of 0.73 µm2 in area per nucleus (Figure 4 a,b), and there was greater variation in number and size of the nuclear bodies. Some cells exhibited few but large nuclear bodies. The average number of bodies was significantly different from that of normal epithelium (P < .05) and nearly significantly different from that of type B CIN III (P = .003). The average size of the nuclear body was significantly larger than normal (P < .001), CIN I/CIN II (P < .05), and type B CIN III (P < .01). SUMO-1 colocalization with the PML-containing nuclear body was lost, with only 18% of the cells exhibiting SUMO-1 colocalization (Figure 3b, Figure 4c).
Nine cases of poorly differentiated carcinomas were analyzed. As in type B CIN III, and in contrast to a well-differentiated tumor, PML body number and size remained lower, with an average of five nuclear bodies per nucleus of 0.37 µm2 in area (Figure 3c, Figure 4 a, b). As expected, the size and number of nuclear bodies were not statistically significantly different from type B CIN III if a progression from type B CIN III to squamous cell carcinoma was predicted. In contrast, they were statistically significantly different in number of nuclear bodies and area from the well-differentiated tumors (P < .05, P < .01) as well as type A CIN III (P < .001, P < .05). Poorly-differentiated carcinomas had significantly fewer nuclear bodies than CIN I/CIN II (P < .001) and the area was almost significantly less. In agreement with progressive neoplasia, SUMO-1 colocalization with the PML-containing nuclear body was lost, with only 10% of the cells exhibiting SUMO-1 colocalization (Figure 3d, Figure 4c).
In addition, ten cases of primary and secondary tumors were examined for differences in PML pattern. In some pairs, immunohistochemistry showed differences in PML expression, ie, size and number of bodies, between primary tumors and their metastases; however, in some pairs, the pattern was the same. Differences in pattern could not be attributed to location of the metastases. In addition, for these cases, no correlation was found between the tumor stage (Ib, II, III), age of patient, or presence of HPV infection as previously assessed diagnostically by polymerase chain reaction (data not shown).
|
DISCUSSION |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We have shown that PML distribution reflects a differential phenotype in cervical tumors, which corresponds with the tumor grading.15 One even more striking observation is the irregularity of the shapes of the PML-containing nuclear bodies, which varies from cell to cell in poorly differentiated tumors. Abnormal track-like structures containing PML were observed, consistent with those seen in viral infection.4 We suggest that the decreased size of the PML-containing nuclear bodies and this track-like distribution reflect an aggressive type of tumor and are distinctively different from the well-differentiated squamous cell carcinoma.
It is remarkable to note that some of the distant metastatic sites display different PML expression patterns than their primary tumors. This implies that a tumor does not have a characteristic PML pattern, which is in accordance with the notion that a tumor has different phenotypes and that metastatic cells have abilities different than those of primary tumors, supporting the idea that the ability to metastasize requires additional mutations or epigenetic changes.
As our findings suggest, well-differentiated tumors might follow a different disease progression than poorly differentiated tumors.15 This difference might be why we observed two distinct staining patterns in CIN III. These data suggest that well-differentiated tumors progress from what we termed type A CIN III, which is characterized by a sustained increase in PML body number and size. In contrast, we speculate that what is termed type B CIN III may progress to a more aggressive, poorly-differentiated tumor.
Taken together, these data suggest that PML is closely involved in the tumorigenesis of cervical carcinoma. Changes in PML pattern may not be restricted to up-regulation and down-regulation of the nuclear protein, but rather to its distribution, ie, size, shape, number, and its ability to associate with SUMO-1. Although the PML’s loss of modification by SUMO-1 is not possible to examine biochemically in patient biopsies because of the heterogeneity of the samples, future work may require the use of cervical cell lines as a model. We suggest that perhaps the PML-containing nuclear body is a downstream target of environmental and intracellular factors that are involved in determining and changing the cellular properties of the tumor. Furthermore, these findings may be adaptable to clinical diagnosis and grading of cervical biopsies. Future studies are needed to extend these findings to Papanicolaou-stained cells, together with an approach to correlate the HPV status of the patients at these earlier stages.
|
Footnotes |
|---|
doi:10.1016/S0029-7844(03)00845-7
Received May 5, 2003. Received in revised form June 27, 2003. Accepted July 30, 2003.
|
REFERENCES |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Munoz N, Bosch FX, deSanjose S, Herrero R, Castellsague X, Shah KV, et al. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N Engl J Med 2003;348:489–90.
3. Lin RJ, Egan DA, Evans RM. Molecular genetics of acute promyelocytic leukemia. Trends Genet 1999;15:179–83.[Medline]
4. Ruggero D, Wang ZG, Pandolfi PP. The puzzling multiple lives of PML and its role in the genesis of cancer. Bioessays 2000;22:827–35.[Medline]
5. Muller S, Hoege C, Pyrowolakis G, Jentsch S. SUMO, ubiquitin’s mysterious cousin. Nat Rev Mol Cell Biol 2001;2:202–10.[Medline]
6. Zhong S, Mueller S, Ronchetti S, Freemont PS, Dejean A, Pandolfi PP. Role of SUMO-1-modified PML in nuclear body formation. Blood 2000;95:2748–52.
7. Ishov AM, Sotnikov AG, Negorev D, Vladimirova OV, Neff N, Kamitani T, et al. PML is critical for ND10 formation and recruits the PML-interaction Protein Daxx to this nuclear structure when modified by SUMO-1. J Cell Biol 1999;147:221–33.
8. Litvinov SV, van Driel W, van Rhijn CM, Bakker HA, van Krieken H, Fleuren GJ, et al. Expression of Ep-CAM in cervical squamous epithelia correlates with an increased proliferation and the disappearance of markers for terminal differentiation. Am J Pathol 1996;148:865–75.[Abstract]
9. Skyldberg B, Salo S, Eriksson E, Aspenblad U, Moberger B, Tryggvason K, et al. Laminin-5 as a marker of invasiveness in cervical lesions. J Natl Cancer Inst 1999;91:1882–7.
10. Vassallo J, Derchain SF, Pinto GA, Martinez EZ, Syrjänen KJ, Andrade LA. High risk HPV and p53 protein expression in cervical intraepithelial neoplasia. Int J Gynaecol Obstet 2000;71:45–8.[Medline]
11. Klaes R, Friedrich T, Spitkovsky D, Ridder R, Rudy W, Petry U, et al. Overexpression of p16(INK4A) as a specific marker for dysplastic and neoplastic epithelial cells of the cervix uteri. Int J Cancer 2001;92:276–84.[Medline]
12. Koken MH, Linares-Cruz G, Quignon F, Viron A, Chelbi-Alix MK, Sobczak-Thépot J, et al. The PML growth-suppressor has an altered expression in human oncogenesis. Oncogene 1995;10:1315–24.[Medline]
13. Terris B, Baldin V, Dubois S, Degott C, Flejou JF, Hénin D, et al. PML nuclear bodies are general targets for inflammation and cell proliferation. Cancer Res 1995;55:1590–7.
14. Dyck JA, Maul GG, Miller WH, Chen JD, Kakizuka A, Evans RM. A novel macromolecular structure is a target of the promelocyte-retinoic acid receptor oncoprotein. Cell 1994;76:333–43.[Medline]
15. Kristensen GB, Abeler VM, Risberg B, Trope C, Bryne M. Tumor size, depth of invasion, and grading of the invasive tumor front are the main prognostic factors in early squamous cell cervical carcinoma. Gynecol Oncol 1999;74:245–51.[Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
