Immunocytochemical assessment of mitotic activity with an antibody to phosphorylated histone H3 in the macaque and human endometrium

Robert M. Brenner1,4, Ov D. Slayden1, William H. Rodgers2, Hilary O.D. Critchley3, Rebecca Carroll1, Xiao Jing Nie1 and Kuni Mah1

1 Oregon National Primate Research Center, 505 NW 185th Ave., Beaverton, OR 97006, 2 Department of Pathology, Oregon Health and Science University, 3181 SW Sam Jackson Park Road, Portland, OR 97239, USA and 3 University of Edinburgh, 49 Little France Crescent, Old Dalkeith Road, Edinburgh EH16 4SB, UK

4 To whom correspondence should be addressed. e-mail: brennerr{at}ohsu.edu


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
BACKGROUND: Determination of the mitotic index in sections of endometrium stained with haematoxylin and eosin (H&E) is difficult and time-consuming. We assessed the value of two mitotic markers as immunocytochemical reagents for measuring mitotic rates in endometrium. METHODS: Mitotic protein monoclonal antibody 2 (MPM-2) and anti-phosphorylated histone H3 (Phospho H3) were applied to paraffin sections of rhesus macaque and human endometrium. RESULTS: In estrogen-treated macaque endometrium the mean ± SEM mitotic indices were: H&E 1.5 ± 0.25%, Phospho H3 antibody 1.02 ± 0.23% and MPM-2 antibody 0.69 ± 0.17%; these were not statistically significantly different, but the Phospho H3 antibody gave a stronger and cleaner signal than the MPM-2 antibody. Comparisons were made between a computer-determined Phospho H3 index, the H&E-determined mitotic index and the Ki-67 index in samples of human endometrium across the cycle. All revealed that the highest proliferative rate occurred during the follicular phase, but the Phospho H3 and the mitotic indices were more highly correlated (R2 = 0.89, P < 0.001) than the Ki-67 and mitotic indices (R2 = 0.74, P < 0.05). CONCLUSIONS: The exceptionally high contrast staining and the excellent correlation between the Phospho H3 and mitotic indices validates the specificity of the Phospho H3 antibody as a new tool for the assessment of endometrial mitotic activity.

Key words: endometrium/Ki-67/mitotic index/mitotic protein monoclonal antibody 2/phosphorylated histone H3


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Accurate measurement of cell proliferation in histological sections of endometrium is an important parameter in the assessment of hormone action (Darj et al., 1991Go; Brenner and Slayden, 1994Go; Darj et al., 1995Go), menstrual cycle stage (Noyes et al., 1950Go) and neoplastic growth (Mora et al., 1999Go). Several methods for assessing cell proliferation are in common use. These include immunostaining for the Ki-67 antigen (Gerdes et al., 1983Go), immunostaining for proliferating cell nuclear antigen (PCNA), and in-vivo labelling with bromodeoxyuridine (Br-dU) or [3H]thymidine to measure DNA synthesis (Ferenczy, 1983Go; Hall and Levison, 1990Go; Darj et al., 1995Go). Most cells that synthesize DNA go on to mitosis, but some are screened out by the DNA repair mechanism and are shifted to the apoptotic rather than the mitotic pathway (Zhou and Elledge, 2000Go). Because the Ki-67 antigen is expressed during several phases of the cell cycle, e.g. G1, S, G2 and M (Gerdes et al., 1984Go; Endl and Gerdes, 2000Go), counts of Ki-67 positive nuclei in paraffin sections are numerically far higher than the mitotic counts in the same sections, though the Ki-67 index is usually well correlated with the mitotic index. However, this correlation does not always hold. For example, in endometria of estrogenized macaques treated with either vehicle or antiprogestin for 2 weeks, the antiprogestin caused a dramatic suppression in the mitotic index but no suppression of the Ki-67 index (Slayden et al., 2001Go). Therefore, counts of Ki-67-stained cells failed to reveal that the antiprogestin treatment had suppressed estrogen-dependent proliferation. Presumably, the cells were blocked at a stage during which Ki-67 protein was continually expressed, but from which cells could not progress through mitosis. Other examples exist of treatments (e.g. X-irradiation) that block cells in G2 but fail to suppress Ki-67 expression. In such cases, only the mitotic index accurately reflects the number of cells that complete the cell cycle.

Direct counting of mitotic cells in haematoxylin and eosin (H&E)-stained sections is a time-consuming process that requires highly skilled observers (Hall and Levison, 1990Go). It is especially difficult to distinguish between cells in prophase and cells undergoing pycnosis or apoptosis in H&E-stained sections. Relatively high magnifications have to be used to distinguish the cycle phases and large numbers of cells must be counted to obtain statistical validity.

The goal of the current work was to assess the value of two antibodies, known to be markers of mitosis, for determining the mitotic index by immunocytochemistry in paraffin sections of human and macaque endometrium, and to determine whether such counts reflected the cyclic and hormonally induced changes that occur in these tissues. These antibodies were: mitotic protein monoclonal 2 (MPM-2) and anti-phosphoylated histone H3 (Phospho H3), a rabbit polyclonal antibody. MPM-2 recognizes several proteins that become phosphorylated during mitosis (Davis et al., 1983Go), including cdc25 (Kuang et al., 1994Go), microtubule-associated proteins (Vandre et al., 1991Go) and an M phase-specific H1 kinase (Kuang et al., 1991Go). The Phospho H3 antibody recognizes histone H3 after it becomes phosphorylated on serine 10 when the chromosomes condense during prophase (Chadee et al., 1999Go; Strahl and Allis, 2000Go; Cunningham, 2002Go). Histone H3 remains phosphorylated until telophase, at which time it becomes dephosphorylated by specific phosphatases. The Phospho H3 antibody has been used to identify chromosomes during mitosis in cultured cells or whole mounts, but to our knowledge, has not been used to identify mitotic figures in formalin-fixed, paraffin-embedded tissue sections of endometrium. The MPM-2 antibody has been used in paraffin-embedded sections, but a comparison with Phospho H3 has not been made, and the suitability of either marker for computer-assisted counting in sections of endometrium under different hormonal conditions or during the human menstrual cycle has not been evaluated.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Antibodies
Mouse monoclonal MPM-2 antibody was purchased from Dako Corp. (USA; Cat. 05-368), rabbit polyclonal Phospho H3 from Upstate Biotechnology (USA; Cat. 06-570) and mouse monoclonal MIB-1 (anti Ki-67 antigen) from Zymed Laboratories (USA; Cat. 08-0156). All other reagents and materials were purchased from Sigma Chemical Co. (USA) unless otherwise indicated.

Animals
Archived endometrial tissues in paraffin blocks were available from rhesus macaques (Macaca mulatta) used in studies at the Oregon National Primate Research Center (ONPRC). These animals had been ovariectomized and either treated with Silastic implants of estradiol (E2) for 28 days to create an extended proliferative phase (n = 3), or sequentially treated with E2 for 14 days followed by E2 + progesterone to create an artificial secretory phase (Rudolph-Owen et al., 1998Go). Secretory phase samples were collected on day 3, 7 and 14 of E2 + progesterone treatment (n = 3/day). Animal care was provided by the veterinary staff of the Division of Animal Resources in accordance with the NIH Guide for the Care and Use of Laboratory Animals.

Human samples
Endometrial tissue was collected from 21 adult women (age range: ~38–48 years) undergoing hysterectomy. Written informed consent was provided by all subjects and ethical approval for tissue collection granted by the Lothian Research Ethics Committee. All women reported regular menstrual cycles (25–35 days) and had not received exogenous hormones or used an intrauterine device in the 3 months prior to inclusion in the study. After the uterus had been removed, a wedge of tissue from the lumen to the muscular myometrial layer that included superficial and basal endometrium as well as myometrium was taken. These tissues were fixed overnight at 4°C in 4% paraformaldehyde (PFA), rinsed and stored in 70% ethanol before routine processing into paraffin wax. The stage of the menstrual cycle was consistent with the patient’s reported last menstrual period and histological dating using the criteria of Noyes et al., (1950Go). Cases with severe uterine pathology, e.g. polyps or fibroids, were excluded. Serum samples taken at the time of hysterectomy were used for determination of circulating progesterone and estradiol levels by radioimmunoassay. All were found to be consistent with the designated cycle stage based on morphological criteria and circulating progesterone concentrations were significantly lower in the late secretory phase compared with the early and mid-secretory phases. Proliferative and secretory endometrial samples were classified as early, mid or late. For statistical analysis, data from the early and mid-proliferative phases were pooled as ‘early’ and compared with all other cycle stages. Paraffin sections of human endometrial adenocarcinoma were obtained from previously approved studies at Oregon Health and Sciences University and had been archived as formaldehyde-fixed, paraffin-embedded blocks.

Antigen retrieval
All samples compared in this study were fixed with 4% buffered paraformaldehyde and embedded in paraffin by standard methods. Consecutive 5 µm sections were cut and mounted on Superfrost/Plus (Fisher Scientific, USA) or poly-L-lysine-coated slides. The slides were deparaffinized in xylene and then rehydrated stepwise in ethanol and rinsed briefly in deionized H2O. Antigen retrieval was performed by heating sections in citrate buffer (BioGenex, USA) for 10 min (as recommended by Biogenex) in a household-type pressure cooker. The antigen-retrieved slides were allowed to cool to room temperature, rinsed once in PBS and then immediately used for immunocytochemistry.

Immunocytochemistry
The slides bearing antigen-retrieved sections were incubated in 3% hydrogen peroxide in methanol for 30 min to block endogenous peroxidases, rinsed once in deionized H2O and then three times in phosphate-buffered saline (PBS). Unless otherwise indicated all the rinsing steps were three times for 3 min each. The sections were incubated with the appropriate species’ serum (Vector Laboratories Inc., USA) for 20 min to block non-specific binding. As standard in our laboratory, sections were incubated overnight with primary antibody at room temperature in a moist chamber; shorter times at moderately elevated temperatures would undoubtedly be appropriate but these were not tested. Primary antibodies were diluted from stock in 1% BSA:PBS and a variety of dilutions was assessed in preliminary studies as follows: 0.05, 0.1 and 1 µg/ml of Phospho H3, and 5.2, 2.6, 1.3, 0.65, 0.325 and 0.16 µg/ml of MPM-2. The best compromise between signal to noise (specific staining versus background) for Phospho H3 was 0.1 µg/ml and for MPM-2 was 2.6 µg/ml. All subsequent staining was done with these concentrations; antibodies from other vendors were not tested. For Ki-67 staining, mouse MIB-1 primary antibody was used in prediluted form as provided by Zymed (200 µg/ml) and incubated similarly overnight; subsequent steps were similar for all antibodies. After overnight incubation with primary antibody the slides were rinsed twice in PBS (10 min each). Sections were incubated again with normal serum (20 min) and then with biotinylated secondary antibody for 30 min. The slides were rinsed in PBS, incubated in ABC solution (Vector Laboratories) for 60 min and then re-rinsed in PBS followed by Tris buffer (38 mmol/l Tris, pH 7.60). Slides were incubated in 0.025% diaminobenzidine (Dojindo Inc.) in Tris buffer, for 10–15 min, and rinsed again in Tris buffer followed by deionized H2O. Slides were then placed in 0.026% osmium tetroxide for 60 s, and rinsed with deionized H2O. Sections were lightly counterstained with Meyer’s haematoxylin, dehydrated and mounted with Permount. A set of consecutive sections was strongly stained with H&E to facilitate identification of mitotic cells, dehydrated and mounted. Digital images were captured with an Optronics DEI-750 CCD camera (Optronics, USA) and analysed with Image Pro-plus (Media Cybernetics Inc., USA). Minor adjustments to contrast and sharpness were made with Adobe Photoshop (Adobe Systems Inc., USA) before printing.

Direct manual counting
Technique
Mitotic figures in sections stained only with H&E and immunopositive cells stained with either MPM-2, Phospho H3 or MIB-1 were counted through the microscope at a magnification of x400 by a trained technician with a mechanical tabulator. Non-overlapping fields were selected with the aid of an ocular grid. Negative (unstained or non-mitotic) background cells (1200–5000 cells per animal) were also counted and the percentage of positive cells [(X positive/Y total count)x100] was calculated for each stain. The abundance of positive or mitotic cells stained by the methods above was assessed by analysis of variance and mean values compared by Fisher’s protected least significant difference test (Petersen, 1985Go).

Observer variation
Others have suggested that the major source of variation in direct counting of stained nuclei in sections is inter-observer variation (Darj et al., 1995). In order to assess whether counting error varied with the different nuclear stains, four different observers counted the same section on four separate days for H&E, MPM-2, Phospho H3 and MIB-1. The counts made by each observer on each day were summed and averaged, and both the within-observer coefficient of variation (CV) and between-observer CV were calculated. Also, counts from all observers for all days were summed and averaged and the CV calculated as a measure of overall precision.

Computer-assisted counting
The specimens stained with the Phospho H3 antibody, but not those stained for MPM-2 (or MIB-1) were of sufficient homogeneity, high contrast and low background to allow computer-assisted thresholding and subsequent computer-assisted counting. For this analysis, digital images were captured at ~x25 original magnification (x2.5 objective). The images were imported into Image Pro-plus, the endometrial glands were traced and defined as regions of interest (ROI), their area was measured, and the number of Phospho H3 positive cells per unit area was determined and expressed as positive cells per 10 000 µm2. To assess the precision of computer-assisted counting, one section was counted six times and the CV calculated.

To assess the degree of correlation between the computer-assisted counting of Phospho H3-labelled mitotic figures and the direct, manual counting of H&E-stained mitotic figures, we plotted the results of direct counting versus computer-assisted counting of the functional layer from the same animals as a scatter plot, and analysed the points by least squares linear regression (Origin Ver 5.0; OriginLab Corp., USA). The macaque data included counts from animals treated with E2 for 14 and 28 days and with E2 + progesterone for 3, 7 and 14 days, which provided a range of high and low values for the correlation analysis. A similar correlation analysis was performed on data from the human endometrium samples taken throughout the menstrual cycle.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Immunostaining with MPM-2 and Phospho H3
The pattern of MPM-2 and Phospho H3 staining, along with standard H&E staining, of macaque endometrium from animals treated for 28 days with E2, is shown in Figure 1a–c. Both MPM-2 and Phospho H3 antibodies produced specific nuclear staining with low background, but the Phospho H3 staining was much stronger, yielding a high signal-to-noise ratio that produced very high contrast between stained and unstained cells. MPM-2 staining was much weaker and lower in contrast.



View larger version (54K):
[in this window]
[in a new window]
 
Figure 1. Mitotic cells in macaque epithelium. Immunostaining with mitotic protein monoclonal antibody 2 (MPM-2) (a), anti-phosphorylated histone H3 (Phospho H3) (b) and haematoxylin and eosin (c). The strongest and cleanest staining was with Phospho H3. Arrows indicate mitotic cells.

 
For these reasons, our efforts focused on the Phospho H3 antibody. Figure 2 illustrates that prophase, metaphase and anaphase were all clearly revealed by Phospho H3 staining in the macaque endometrium. Figure 2d (arrow) shows a telophase figure that is not stained by Phospho H3 antibody due to the dephosphorylation of histone H3 that occurs during telophase. Apoptotic figures were also negative for Phospho H3 staining (Figure 2b).



View larger version (105K):
[in this window]
[in a new window]
 
Figure 2. Mitotic cells in macaque epithelium. Illustrations of the phases of the mitotic cycle stained with anti-phosphorylated histone H3 (Phospho H3), specifically, prophase (a), metaphase (b) and anaphase (c). In (b), Phospho H3 antibody did not stain any apoptotic (Ap) cells. Staining is lost during telophase (d) due to the dephosphorylation of histone H3. P = prophase; M = metaphase; A = anaphase; T = telophase cells. In (c) a mitotic cell is evident in the stroma.

 
The Phospho H3 antibody also clearly revealed mitotic figures, but not apoptotic figures, in paraffin-embedded sections of proliferative human endometrium (Figure 3a,b). Additionally, in the human (but not the macaque) MPM-2 stained a cytoplasmic component in the apical region of the epithelial cells, indicating that this antibody lacked specificity for human nuclear components in the fixed, sectioned material that we used (see Figure 3c).



View larger version (129K):
[in this window]
[in a new window]
 
Figure 3. Mitotic cells in human tissues. (a, b) Human proliferative phase endometrium stained with anti-phosphorylated histone H3 (Phospho H3): P = prophase; A = anaphase; P' = prophase in stromal cell; M = metaphase; Ap = apoptosis. (c) Human proliferative endometrium stained with mitotic protein monoclonal antibody 2 (MPM-2). ‘Mitotic’ arrow points to a positive nucleus in the glandular epithelium. Arrowheads point to non-specific staining in the apical regions of all glandular epithelial cells. This non-specific staining only occurred in the human tissues. (d, e) Closely spaced serial sections of human endometrial adenocarcinoma stained with haematoxylin and eosin (d) and Phospho H3 (e). In (d), short arrow labelled P points to a mitotic figure (prophase), long arrows point to equivocal figures (apoptotic or pycnotic). In (e), long arrows point to definitive mitotic figures, and arrow labelled G points to a grazing section of positively stained chromosomes.

 
In a paraffin section of human endometrial adenocarcinoma stained with H&E (Figure 3d) some mitotic figures are evident, but other nuclei cannot be distinctly defined as apoptotic, pycnotic or mitotic. In Figure 3e, which is a close serial section of Figure 3d stained with Phospho H3, several mitotic figures are evident, including grazing sections of chromosomes, that are not evident in the H&E-stained section. Because of the specificity of Phospho H3 for mitosis, no apoptotic or pycnotic figures are stained.

Direct manual counting
The direct counts of mitotic figures in endometrium of macaques that had been treated with E2 for 28 days were as follows: H&E (1.5 ± 0.25%), Phospho H3 antibody (1.02 ± 0.23%), and MPM-2 antibody (0.69 ± 0.17%). The MIB-1 stained cells were, as expected, the most numerous (19.7%). Aside from the MIB-1 data, which was significantly higher than any of the other endpoints, there were no statistically significant differences among the H+E, Phospho H3 and MPM-2 indices; however, the MPM-2 positive cells were numerically lowest. Consequently, the Phospho H3 index is a valid equivalent to the mitotic index. The differences in direct counting due to observer variation are shown in Table I. Among the direct stains of mitotic figures (H+E, MPM-2 and Phospho H3) the coefficient of variation was lowest for the Phospho H3 index.


View this table:
[in this window]
[in a new window]
 
Table I. Coefficient of variation (CV, %) for counts of positive cells in sections of macaque endometrium stained with haematoxylin and eosin (H&E), and anti-phosphorylated Histone H3 (Phospho H3), mitotic protein monoclonal antibody 2 (MPM-2) and mouse monoclonal MIB-1
 
Computer-assisted counting
Figure 4 illustrates the computer-assisted counting method. Figure 4a shows a low power field (x2.5 objective) with arrows pointing to many of the Phospho H3-stained mitotic figures after thresholding by the software. Once the threshold levels were set, the areas of the glands were traced to provide regions of interest (ROI; Figure 4b). The software then labelled each ROI alphanumerically (A1–A14 in Figure 4b), numbered the thresholded cells in each ROI (1–23 in Figure 4b), determined the total area of the ROI, and calculated the number of positive cells/10 000 µm2. For example, the mean number of thresholded mitotic figures in endometria treated for 14 days with E2 was 2.74 ± 0.96 cells/10 000 µm2. Computer-assisted counting was more rapid than manual counting and had high reproducibility (CV = 4.2%). Time required to make these counts was ~15 min/section which included calculating the area of the ROI, compared to ~60 min by direct, manual counting.



View larger version (91K):
[in this window]
[in a new window]
 
Figure 4. Visualization of computer-assisted counting of anti-phosphorylated histone H3 (Phospho H3)-stained mitotic figures. (a) Digital images at low magnification (x2.5 objective) were imported into Image-Pro Plus, and the positive stained cells identified by thresholding show as intensely black cells; inset shows how the unthresholded cells appear at higher magnification (x25 objective). (b) The glands were traced as regions of interest (ROI) and identified alphanumerically by the software (A1–A14). Thresholded cells were then automatically counted and labelled in each ROI from 1–23 by the software. The software finally determined the number of positive cells/10 000 µm2, which we define as the Phospho H3 index.

 
The least squares regression analysis between the computer-assisted counting of Phospho H3-labelled mitotic figures and direct counting of H&E-stained mitotic figures is shown graphically in Figure 5. The correlation between the two methods was very high (R2 = 0.90, P < 0.001), further indicating that the computer-determined Phospho H3 index is a valid estimate of the abundance of mitotic cells in the macaque endometrium.



View larger version (17K):
[in this window]
[in a new window]
 
Figure 5. Least squares analysis between direct, manual counting of mitotic figures in haematoxylin and eosin (H&E)-stained sections and computer-assisted counting of anti-phosphorylated histone H3 (Phospho H3)-stained cells in the functionalis zone of the rhesus macaque endometrium. The analysis included data from all phases of the artificial cycle. The analysis revealed a strong correlation (R2 = 0.90) between computer-assisted Phospho H3 index determinations and direct, manual mitotic count in H&E-stained sections, indicating that the computer-assisted Phospho H3 index can be used to quantitatively determine mitotic cell abundance in histological specimens.

 
To further validate the computer-assisted Phospho H3 index, we compared the Phospho H3 proliferative index of the functionalis and basalis zones in the macaque endometrium during the artificial proliferative phase (14 days of E2) and on days 3, 7 and 14 of the artificial luteal phase (E2 + progesterone) of the cycle. These results are shown in Figure 6. Phospho H3 positive cells were abundant in the functionalis after E2 treatment. After 3, 7 and 14 days of E2 + progesterone treatment the counts in the functionalis sharply decreased, while mitotic activity in the basalis zone, which was minimal during E2 treatment, significantly increased on days 3 and 7 of progesterone treatment. These results are consistent with previous reports that progesterone stimulates proliferation in the basalis zone (Padykula et al., 1989Go) while suppressing mitotic activity in the functionalis of the macaque endometrium (Brenner and Slayden, 1994Go).



View larger version (16K):
[in this window]
[in a new window]
 
Figure 6. Anti-phosphorylated histone H3 (Phospho H3) index determinations (mean ± SEM) in the functionalis and basalis zones of the macaque endometrium at various time points of the artificial menstrual cycle. Means within each zone with different superscripts are statistically different.

 
Data from the endometrium across the human menstrual cycle are presented in Figure 7. The computer-assisted Phospho H3 index showed a clear pattern of elevated mitotic activity in the proliferative phase that significantly decreased in the secretory phase; this pattern was highly correlated with the mitotic index as determined by H&E staining (R2 = 0.89, P < 0.001). A high, though lesser, correlation was also evident between the Phospho H3 index and the Ki-67 index (R2 = 0.74, P < 0.05) and between the H&E mitotic index and the Ki-67 index (R2 = 0.74, P < 0.05). The exceptionally high correlation between the Phospho H3 index and the mitotic index validates the Phospho H3 antibody as a useful, highly specific mitotic marker for the human endometrium.



View larger version (25K):
[in this window]
[in a new window]
 
Figure 7. Changes in anti-phosphorylated histone H3 (Phospho H3), mitotic and Ki-67 indices in the human endometrium during the natural menstrual cycle. Bars with different letters are significantly different (within each index). The number of patients sampled at each cycle stage is indicated in the bars.

 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
General significance
In this work, our goal was to develop and validate a new method for detecting and counting mitotic figures in paraffin sections of the human endometrium. Toward this end, we compared two known markers of phosphorylated proteins associated with mitosis, MPM-2 and Phospho H3, which to our knowledge had not been previously evaluated in paraffin-embedded sections of the endometrium. Of the two, only Phospho H3 staining was intense enough and had low enough background to allow thresholding of positive nuclei to facilitate computerized analysis. The MPM-2 antibody is also known to recognize phosphorylated RNA polymerase II during interphase (Albert et al., 1999Go), and it stained cytoplasmic components in the human, though not the macaque endometrium, so it is less useful for the current purpose than the Phospho H3 marker. In sum we found that a computer-assisted Phospho H3 index was highly correlated with the mitotic index, that it could be used to assess hormonally induced changes in mitotic activity in the macaque endometrium, and that it provided a new tool to assess cyclic changes in endometrial proliferation during the cycle or in histopathological states. As such it complements other methods of assessing proliferation, such as the Ki-67 index, by providing a specific count of cells that actually divide.

There were a few initial concerns about the method. First, because histone H3 is dephosphorylated during telophase, one might expect direct counts of mitoses in Phospho H3-stained sections to be lower than direct counts of mitotic figures in H&E-stained sections, because a microscopist doing H&E counts would count telophases. But the microscopist generally misses many prophases, as these are difficult to discern in H&E-stained sections, whereas prophases are easy to detect with the Phospho H3 stain. In any event, when direct counts of Phospho H3-stained cells were compared with direct counts of H&E-stained mitotic figures there was no statistically significant difference.

Second, during computer-assisted analysis of Phospho H3-stained samples, image thresholding would identify cells during late anaphase as two mitotic figures, because the chromosomes are cleanly separated, while a microscopist performing direct counts would count them as one cell in anaphase. The computer-assisted Phospho H3 index could therefore be higher than the directly counted mitotic index. However, this difference is probably balanced by the lower number of telophases that are included in the Phospho H3 index. The tendency to overcount anaphases is also counterbalanced by the fact that late anaphase is a rapid component of the mitotic cycle and tends to be rare in sections. In any event, there was a very high correlation (R2 = 0.90; Figure 6) between the computer-assisted, Phospho H3 index and the direct, mitotic count (expressed as a percentage) in H&E-stained samples of macaque endometrium. Consequently, the Phospho H3 index is a valid equivalent of the mitotic index, whether determined by direct or computer-assisted counting.

Because of the high contrast of the Phospho H3 stain, the computerized technique can sample large fields at low magnifications (see Figure 5), while the higher magnifications that are required to precisely identify mitotic figures in H&E-stained specimens necessarily limits the size of the microscopic field. The larger sample size obtained during computer-assisted counting reduces experimental error.

In addition to marking mitosis, phosphorylation of histone H3 on serine 10 may occur during chromatin modifications associated with cellular differentiation (Juan et al., 1999Go; Salvador et al., 2001Go) and oncogene-induced transformation (Chadee et al., 1999Go). For example, phosphorylation of histone H3 at serine 10 is involved in FSH-stimulated differentiation of rat ovarian granulosa cells and is mediated by protein kinase A (Salvador et al., 2001Go). This phenomenon was observed in cell culture under serum-free conditions in the absence of proliferation. Whether this reaction would interfere with the use of a Phospho H3 antibody to evaluate mitotic figures in the primate ovary remains to be determined.

Additionally, the nuclei of human monocytes have relatively high levels of phosphorylated histone H3 during interphase, and differentiation of HL-60 (human myelogenous leukaemia) cells along the monocytic lineage is accompanied by a rise in H3 phosphorylation (Juan et al., 1999Go). The phosphorylated sites are distributed in a speckled, punctate manner throughout the monocyte nucleus. This punctate pattern is strikingly different from the more homogeneous pattern seen throughout the chromosomes in mitotic nuclei after Phospho H3 staining. Although monocytes and macrophages are numerous in the primate endometrium, no speckled nuclear staining was observed in these cells with our current Phospho H3 staining protocol. Therefore, interphase phosphorylation of histone H3 in monocytes should not interfere with the use of the Phospho H3 antibody to evaluate mitotic indices in the glandular epithelium of the primate endometrium.

During the early, pioneering work on the endometrium, the proliferative state of endometrial specimens was assessed by direct microscopic counting of mitotic cells in H&E-stained sections (Noyes et al., 1950Go). More recently, detection of the Ki-67 antigen (Gerdes et al., 1983Go), which is expressed in all phases of the cell cycle (G1, S, G2 and M) has been extensively used as a measure of proliferative activity in various tissues (Scholzen and Gerdes, 2000Go; Brown and Gatter, 2002Go) including endometrial biopsies (Jurgensen et al., 1996Go). However, the time cells spend in G1 is highly variable, and may be affected by the hormonal or neoplastic state of the tissues (Hall and Levison, 1990Go). Also, staining for proteins that are expressed throughout several phases of the cell cycle cannot provide a precise measure of the rate of proliferation at the time of sampling (Hall and Levison, 1990Go; Hall and Woods, 1990Go). For a staining technique to give the most accurate estimate of proliferative rate, the period of time during which the cells are positive should be short. Because mitosis is one of the shortest and least variable phases of the cell cycle, the mitotic index provides the most precise estimate of the proliferative rate.

Consequently, although the Ki-67 antigen is a highly useful marker to detect proliferating populations, a count of mitotic figures adds information on the number of cells that actually complete the cell cycle, thus enhancing and complementing the information obtained from the Ki-67 marker. In sum, the Phospho H3 antibody adds a powerful new tool to our armamentarium for the analysis of cell proliferation in archived paraffin-embedded sections.

Potential clinical relevance
There are various clinical situations in which analysis of the endometrial Phospho H3 index would be valuable. Endometrial biopsies are commonly performed in pre- and post-menopausal women with abnormal uterine bleeding to exclude neoplastic endometrial disease. One important component of the pathological diagnosis of such biopsies is a determination of the proliferative state of the tissue, typically done by identifying mitotic figures in H&E-stained tissue sections. However, in pathological endometrial tissues stained with H&E, necrotic and apoptotic cells frequently resemble prophase nuclei, complicating assessment of the mitotic index. The Phospho H3 stain, because of its chromosomal specificity and high contrast, greatly facilitates such determinations and also aids in identifying abnormal mitotic figures—a pathological feature associated with neoplastic disease.

Several common women’s reproductive health disorders, including endometriosis and polycystic ovary syndrome (PCOS), are estrogen-dependent conditions. In PCOS there may be overexpression of endometrial estrogen receptor co-activators which would enhance estrogen sensitivity of this tissue (Gregory et al., 2002Go). Androgen receptors are elevated in PCOS endometria (Apparao et al., 2002Go) and these may act to counteract such estrogenic effects. Women with PCOS are therefore exposed to counteracting influences that could affect endometrial proliferative rates.

The management of women with PCOS with amenorrhoea or oligomenorrhoea includes administration of cyclic progestin therapy to induce endometrial withdrawal bleeding in order to prevent endometrial hyperplasia (Balen, 2001Go). In addition, clinical management of such cases includes a transvaginal ultrasound scan to assess endometrial thickness as an aid to predicting endometrial hyperplasia (Cheung, 2001Go). However, those subjects with an increased endometrial thickness on ultrasound require an endometrial biopsy to provide a definitive assessment of the histological and proliferative aspects of the tissues, because fluid content can affect the thickness measurement obtained by ultrasound. Phospho H3 provides an excellent additional marker to detect excessive mitotic activity in such samples. Computer-assisted determinations of mitotic activity in such non-neoplastic endometrial disease conditions should permit a more precise assessment of the effect of steroid hormone therapy on endometrial proliferation.

Because of the clinical importance of endometrial hyperplasia and adenocarcinoma, this paper has focused on the endometrium, but the method should be applicable to quantification of mitotic activity in clinical specimens from formalin-fixed, paraffin-embedded samples of other tissue types. Also, the Phospho H3 antibody works well in cryosections (data not shown) and should work well in double-labelling studies with any antibody that requires frozen sections. In sum, immunocytochemical detection of the phosphorylated histone H3 marker should facilitate analysis of how steroid hormones, steroid antagonists and antineoplastic drugs affect endometrial mitotic activity. Phospho H3 is a valuable tool that should improve our understanding of the mechanisms involved in the pathogenesis of female reproductive disorders.


    Acknowledgements
 
We thank Dr Alistair Williams of the Department of Pathology, University of Edinburgh, for providing tissue samples and for helpful and encouraging discussions. This work was supported by grants HD RR00163 (R.M.B.), H.D. 19182 (R.M.B. and O.S.), HD 18185 (R.M.B.) and HD 39363 (O.S.).


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Albert, A., Lavoie, S. and Vincent, M. (1999) A hyperphosphorylated form of RNA polymerase II is the major interphase antigen of the phosphoprotein antibody MPM-2 and interacts with the peptidyl-prolyl isomerase Pin1. J. Cell Sci., 112 (Part 15), 2493–2500.[Abstract/Free Full Text]

Apparao, K.B., Lovely, L.P., Gui, Y., Lininger, R.A. and Lessey, B.A. (2002) Elevated endometrial androgen receptor expression in women with polycystic ovarian syndrome. Biol. Reprod., 66, 297–304.[Abstract/Free Full Text]

Balen, A. (2001) Polycystic ovary syndrome and cancer. Hum. Reprod. Update, 7, 522–525.[Abstract/Free Full Text]

Brenner, R.M. and Slayden, O.D. (1994) Cyclic changes in the primate oviduct and endometrium. In Knobil, E. and Neill, J.D. (eds), The Physiology of Reproduction. Raven Press, New York, pp. 541–569.

Brown, D.C. and Gatter, K.C. (2002) Ki67 protein: the immaculate deception? Histopathology, 40, 2–11.[CrossRef][ISI][Medline]

Chadee, D.N., Hendzel, M.J., Tylipski, C.P., Allis, C.D., Bazett-Jones, D.P., Wright, J.A. and Davie, J.R. (1999) Increased Ser-10 phosphorylation of histone H3 in mitogen-stimulated and oncogene-transformed mouse fibroblasts. J. Biol. Chem., 274, 24914–24920.[Abstract/Free Full Text]

Cheung, A.P. (2001) Ultrasound and menstrual history in predicting endometrial hyperplasia in polycystic ovary syndrome. Obstet. Gynecol., 98, 325–331.[Abstract/Free Full Text]

Cunningham, B.A. (2002) Unraveling chromatin’s secrets: increasing numbers of high-quality, modification-specific histone antibodies give new life to chromatin studies. The Scientist, 16, 47–50.

Darj, E., Nilsson, S., Axelsson, O. and Hellberg, D. (1991) Clinical and endometrial effects of oestradiol and progesterone in post-menopausal women. Maturitas, 13, 109–115.[ISI][Medline]

Darj, E., Axelsson, O., Nilsson, G., Nilsson, S. and Risberg, B. (1995) Ki-67 immunostaining of endometrial biopsies with special reference to hormone replacement therapy. Gynecol. Obstet. Invest., 39, 120–124.[ISI][Medline]

Davis, F.M., Tsao, T.Y., Fowler, S.K. and Rao, P.N. (1983) Monoclonal antibodies to mitotic cells. Proc. Natl Acad. Sci. USA, 80, 2926–2930.[Abstract]

Endl, E. and Gerdes, J. (2000) The Ki-67 protein: fascinating forms and an unknown function. Exp. Cell Res., 257, 231–237.[CrossRef][ISI][Medline]

Ferenczy, A. (1983) Cytodynamics of endometrial hyperplasia and neoplasia, Part II: In vitro DNA histoautoradiography. Hum. Pathol., 14, 77–82.[ISI][Medline]

Gerdes, J., Schwab, U., Lemke, H. and Stein, H. (1983) Production of a mouse monoclonal antibody reactive with a human nuclear antigen associated with cell proliferation. Int. J. Cancer, 31, 13–20.[ISI][Medline]

Gerdes, J., Lemke, H., Baisch, H., Wacker, H.H., Schwab, U. and Stein, H. (1984) Cell cycle analysis of a cell proliferation-associated human nuclear antigen defined by the monoclonal antibody Ki-67. J. Immunol., 133, 1710–1715.[Abstract/Free Full Text]

Gregory, C.W., Wilson, E.M., Apparao, K.B., Lininger, R.A., Meyer, W.R., Kowalik, A., Fritz, M.A. and Lessey, B.A. (2002) Steroid receptor coactivator expression throughout the menstrual cycle in normal and abnormal endometrium. J. Clin. Endocrinol. Metab., 87, 2960–2966.[Abstract/Free Full Text]

Hall, P.A. and Levison, D.A. (1990) Review: Assessment of cell proliferation in histological material. J. Clin. Pathol., 43, 184–192.[ISI][Medline]

Hall, P.A. and Woods, A.L. (1990) Immunohistochemical markers of cellular proliferation: achievements, problems and prospects. Cell Tissue Kinet., 23, 505–522.[ISI][Medline]

Juan, G., Traganos, F. and Darzynkiewicz, Z. (1999) Histone H3 phosphorylation in human monocytes and during HL-60 cell differentiation. Exp. Cell Res., 246, 212–220.[CrossRef][ISI][Medline]

Jurgensen, A., Mettler, L., Volkov, N.I. and Parwaresch, R. (1996) Proliferative activity of the endometrium throughout the menstrual cycle in infertile women with and without endometriosis. Fertil. Steril., 66, 369–375.[ISI][Medline]

Kuang, J., Penkala, J.E., Wright, D.A., Saunders, G.F. and Rao, P.N. (1991) A novel M phase-specific H1 kinase recognized by the mitosis-specific monoclonal antibody MPM-2. Dev. Biol., 144, 54–64.[ISI][Medline]

Kuang, J., Ashorn, C.L., Gonzalez-Kuyvenhoven, M. and Penkala, J.E. (1994) cdc25 is one of the MPM-2 antigens involved in the activation of maturation-promoting factor. Mol. Biol. Cell, 5, 135–145.[Abstract]

Mora, L.B., Diaz, J.I., Cantor, A.B. and Nicosia, S.V. (1999) Differential diagnosis of endometrial hyperplasia and carcinoma by computerized image cytometry of cell proliferation, apoptosis and Bcl-2 expression. Ann. Clin. Lab. Sci., 29, 308–315.[Abstract]

Noyes, R.W., Hertig, A.T. and Rock, J. (1950) Dating the endometrial biopsy. Fertil. Steril., 1, 3–25.[ISI][Medline]

Padykula, H.A., Coles, L.G., Okulicz, W.C., Rapaport, S.I., McCracken, J.A., King, N.W., Jr, Longcope, C. and Kaiserman-Abramof, I.R. (1989) The basalis of the primate endometrium: a bifunctional germinal compartment. Biol. Reprod., 40, 681–690.[Abstract]

Petersen, R.G. (1985) Owen, D.B. (ed.), Design and Analysis of Experiments. Marcel Dekker, New York, pp. 72–111.

Rudolph-Owen, L.A., Slayden, O.D., Matrisian, L.M. and Brenner, R.M. (1998) Matrix metalloproteinase expression in Macaca mulatta endometrium: evidence for zone-specific regulatory tissue gradients. Biol. Reprod., 59, 1349–1359.[Abstract/Free Full Text]

Salvador, L.M., Park, Y., Cottom, J., Maizels, E.T., Jones, J.C., Schillace, R.V., Carr, D.W., Cheung, P., Allis, C.D., Jameson, J.L. and Hunzicker-Dunn, M. (2001) Follicle-stimulating hormone stimulates protein kinase A-mediated histone H3 phosphorylation and acetylation leading to select gene activation in ovarian granulosa cells. J. Biol. Chem., 276, 40146–40155.[Abstract/Free Full Text]

Scholzen, T. and Gerdes, J. (2000) The Ki-67 protein: from the known and the unknown. J. Cell. Physiol., 182, 311–322.[CrossRef][ISI][Medline]

Slayden, O.D., Chwalisz, K. and Brenner, R.M. (2001) Reversible suppression of menstruation with progesterone antagonists in rhesus macaques. Hum. Reprod., 16, 1562–1574.[Abstract/Free Full Text]

Strahl, B.D. and Allis, C.D. (2000) The language of covalent histone modifications. Nature, 403, 41–45.[CrossRef][ISI][Medline]

Vandre, D.D., Centonze, V.E., Peloquin, J., Tombes, R.M. and Borisy, G.G. (1991) Proteins of the mammalian mitotic spindle: phosphorylation/dephosphorylation of MAP-4 during mitosis. J. Cell Sci., 98 (Part 4), 577–588.[Abstract]

Zhou, B.B. and Elledge, S.J. (2000) The DNA damage response: putting checkpoints in perspective. Nature, 408, 433–439.[CrossRef][ISI][Medline]

Submitted on January 30, 2003; accepted on February 28, 2003.