Nuclear Morphometric Analysis (NMA): Screening of Senescence, Apoptosis and Nuclear Irregularities

1The nucleus corresponds to approximately 10% of the cellular volume and, due to its nuclear envelope, presents a round shape and a well-defined and regular surface under normal conditions in vitro. Alteration in nuclear morphology occurs in physiologic situations, like during mitosis, and in several processes associated to cell death. These modifications include nuclear condensation and fragmentation observed in apoptosis, increase in nuclear size observed in senescence, and increase in nuclear irregularity observed in several conditions, such as chemical or physical stresses, defective activation or inactivation of cell cycle checkpoint signaling processes, or exogenous agents that affect microtubule dynamics or chromatin remodeling,.

2Analysis of nuclear irregularities, such as mitotic catastrophe (MC), is limited almost exclusively to visualization of nuclei stained with DAPI (4′, 6-diamino-2-phenylindole), followed by subjective counting of nuclei with signs of MC. Recently, a tool was proposed for the detection of MC based on videomicroscopy of cells expressing markers of chromatin and centrosomes. However, this tool requires specific equipment and is rather difficult to set up and analyze.

3Both replicative senescence and oncogene-induced senescence are important anti-cancer mechanisms and reactivation of senescence by therapy may represent a good anticancer strategy,. Senescence is analyzed by the level of β-galactosidase activity measured at pH 6.0 (known as senescent-associated β-galactosidase – SABG), the formation of senescence-associated heterochromatin foci (SAHF) and the increase in the expression of proteins associated with senescence induction such as p21, p27, INK4A and Arf. Senescence in culture cells produces a drastic increase in cellular and nuclear size, therefore a measure of nuclear size, associated with the reduction in mitotic cells, maybe a good indicator of the senescent state at a given cell culture population.

4Induction of MC by small molecules or specific inhibition of DNA damage-activated signaling may increase cytotoxicity in cancer cells,. In these cases, co-occurrence of senescence, MC and apoptosis is clinically desirable when compared to the occurrence of only one of these processes. Therefore, a tool to easily evaluate morphological signs of apoptosis, MC and senescence is an important part in anti-cancer therapeutic studies.

5Several cellular processes can be inferred by the analysis of nuclear morphometric features and a method without bias that evaluates nuclear phenotypes would widely improve our understanding concerning the dynamics of cell growth and death. Here, we developed an objective tool named Nuclear Morphometric Analysis (NMA) which is able to indicate the proportion of cells in senescence, apoptosis or with nuclear irregularities in a cell population in vitro based on nuclear morphology.

6In order to develop the tool we used nuclei from the glioblastoma cell line U87 treated with temozolomide (TMZ), which causes DNA damage and G2 cell cycle arrest, in combination with resveratrol, which blocks this arrest. This combination induced cells to undergo mitosis with increased expression of cyclin B1 and decreased pCDC2(Y15) levels, despite wide DNA damage, as indicated by γH2AX and comet assay (data not shown), thus generating a multitude of nuclear morphology. Nuclei were stained with DAPI and pictures were taken on a fluorescent inverted microscope. Images were analyzed with the Image Pro Plus 6.0 (IPP6) software and 46 parameters of nuclear size, shape and marking were produced for a large population of cells. These parameters were analyzed with principal component analysis (PCA) and the features Aspect (Asp), Area box (Arbx), Radius ratio (Rr) and Roundness (Rou) were combined in an index called Nuclear Irregularity Index (NII = Asp−Arbx+Rr+Rou). Details on the development of the tool are given in the Text S1 file.

7In order to permit a wider use of the NMA tool, we developed an Image J plugin (for Image J 1.45 free version) that allows the quantification of the five measurements required for NMA (i.e. those for calculation of NII plus area measurement). A detailed explanation of the plugin features and its use is provided in the Text S2 file. A comparative analysis of data generated with the IPP6 software and the Image J plugin that accompanies the NMA tool showed very similar percentages of nuclei for each population (Text S1), validating the plugin developed for the NMA analysis.

8Data generated with the IPP6 or Image J software were then analyzed in a spreadsheet (NMA.xls or NMA.ods). Data from normal, untreated cells (in which small, large and irregular nuclei are excluded) are used to set the parameters of the normal population. The user can define the number of standard deviations (SD) from the mean that will be considered an acceptable deviation. Next, a large set of data from different treatments are used to set the thresholds of the different populations. These settings (number of SD and the thresholds) are used to analyze the nuclei from different treatments. After entering the data from the nuclei in the data analysis sheets, a plot of area versus NII will be automatically produced ( Figure 1 ), showing the ellipse representing the normal nuclei and the different populations whose thresholds were set in the “Normal Nuclei and Settings” sheet. Detailed use of the spreadsheets is given in the Text S2 file. The Image J plugin, Microsoft Excel® and Open Office Calc files are available at http://www.ufrgs.br/labsinal/nma/.

9Analyzing the data produced with the nuclear population used to set up the tool we noticed a distribution in four separate populations: normal, large, irregular and small (Text S2). Analysis of a larger set of conditions, including cells treated with classical inducers of senescence, apoptosis and nuclear irregularities, lead to the suggestion of seven different categories of nuclei based on size and NII ( Figure 1, Table 1 and Text S1).

10Normal nuclei of adhered cells in culture vary in size, but have a very regular shape ( Figure 1.h,i,j), with some located at the borderline between N and I ( Figure 1.k) and analysis of these nuclei are important for the precise setting of the threshold that separates these populations. Nuclei classified as I, ( Figure 1.a,b,c,l,n,r) present high irregularity. Among the large nuclei, we found a clear separation between two groups in three experimental settings ( Figure 2, Figure 3 and Figure S1 and S2). On the lower NII range are the well-defined and regular nuclear shapes ( Figure 1.e,g), whereas at higher NII range are the nuclei presenting morphological alterations ( Figure 1.d,f), which may be undergoing MC.

11Nuclei undergoing mitosis have higher values of NII than normal nuclei because of their elliptic shapes, and a broader range of NII due to the variations in the shapes of the mitotic plate ( Figure 1.m,p). However, some small nuclei also present very high irregularity ( Figure 1.q), being thus classified as small and irregular (SI). It is also important to stress that in the NMA tool, each mitotic cell will be counted as one before anaphase ( Figure 1.o), while after the anaphase the NMA tool will count two nuclei, which are normally located close to each other in the graph ( Figure 1.m). Therefore, the NMA is not a good tool for counting mitotic cells, since at anaphase the two nuclei counted will still be part of a single cell, and will be evaluated as two nuclei. Notwithstanding, it provides a broad assessment of mitotic proportion in a cell population, as illustrated by the almost complete absence of mitotic nuclei in the senescent population analyzed in Figure 2.

12For the validation of senescence, we used the rat glioma cell line C6 treated with the combination of resveratrol and quercetin, which we have previously shown to induce high levels of senescence. About 90% (35 of 39) of the large nuclei(LR+LI) were β-galactosidase positive (blue diamonds and examples of positive cells in inserts a–g), whereas only 4 out of 36 nuclei that localized in the normal population were β-gal positive ( Figure 2 – bottom graph). Additionally, in this population, only 2 mitosis were found in around 90 cells (around 2.2%), whereas in the control population this ratio was around 14% ( Figure 2 – top graph). These data support the usefulness of the NMA tool in predicting the senescent status of a cell population, with the advantage of being much less subjective than the current standard method of β-gal staining and visual identification of the positive cells.

13To test the usefulness of the NMA tool to detect nuclear irregularities, we measured nuclei of a population of U87 glioma cells treated with vincristine (50 nM), which is described to induce several nuclear phenotypes, including MC and/or senescence,, depending on the cell type and treatment. We observed that the treatment induced an increase of irregular nuclei (I and LI), from 7% in control to 41% (12%+29%) ( Figure 3 ). Moreover, no nuclei were observed in the region were normal mitotic nuclei localized in the NMA graph.

14In order to test the applicability of the NMA tool in other cell types, we treated the colorectal carcinoma cell line HCT116 with vincristine 50 nM. As expected, vincristine increased the percentage of both LI and LR nuclei. Average values of area of the LI and LR populations also increased, as did the NII of the LI population of nuclei treated with vincristine when compared to control (Figure S1). It is important to note that induction of senescence-like phenotype in cancer cells usually is accompanied by MC. Thus, our tool allows a dynamic analysis of the development of both events over time.

15We have also performed an analysis of data from published papers of Shao et al and Vakifahmetoglu et al, which presented enough control nuclei to set the standards for the studied cell line as well as pictures of nuclei undergoing MC that could be measured. Despite the small number of images for control nuclei, NMA provided a well-defined distribution for the control nuclei and a broad distribution of the catastrophic nuclei, mainly in the LI group. All treated nuclei localized in the I or LI regions of the NMA plot. This is expected, since the illustrative pictures probably were selected to clearly show the catastrophic nature of the nuclei (Figure S2).

16Cells undergoing apoptosis suffer a high and regular condensation of the nucleus which in culture occurs prior to nuclear fragmentation and cell detachment but almost concomitantly with the externalization of phophatidylserine and formation of blebbings in the cell membrane. Due to this high condensation in a near spherical form, we hypothesized that nuclei of cells undergoing apoptosis may appear as small and regular (SR), i. e. on the lower left corner of the NMA plot. In order to validate this hypothesis, we evaluated the induction of apoptosis in U87 cells treated with cisplatin the presence of zAsp, a caspases 3/7 inhibitor. Apoptosis was confirmed by annexin V-FITC staining. Fourteen out of 29 nuclei located in the SR were annexin V positive as visualized with a fluorescence microscope and none of the annexin V-positive cell was located in any region outside SR ( Figure 4B ), whereas no cell was annexin V-positive in the vehicle-treated ( Figure 4A ).

17A side by side analysis with NMA and annexin analysis through flow cytometry presented a very good correlation ( Figure 5B ). Inhibition of apoptosis with a caspases inhibitor reduced the apoptotic population detected with NMA to 7% and annexin to 10% in comparison to 49 and 57%, respectively, in the absence of zAsp ( Figure 5A and B ).

18Additionally, NMA was also able to detect etoposide-induced apoptosis in HeLa cells, as well as differentiate apoptotic and mitotic cells (Figure S3), which normally appear in the S region of the NMA graph. Precise separation of S and SR regions may difficult the quantitative analysis of mitotic cells using NMA and additional methods should be employed to confirm mitotic cells. We also performed a kinetic analysis of Hoechst 3342-stained nuclei of HeLa cells treated with cisplatin. NMA graphs with the dynamics of area and NII, together with visible images of the cells undergoing apoptosis support the notion that nuclei located in the SR region of the NMA graph are apoptotic independent of the size and shape of the nuclei before apoptosis (Figure S4).

19We present a new and objective tool based on measurements of morphometric data that allows the unbiased simultaneous analysis of nuclear features indicative of normal, senescent, apoptotic and irregular nuclei. It is important to stress that a method based solely on nuclear morphology is simplistic for the definitive identification of such complex cellular processes. Final identification of these processes has to be confirmed with standard techniques for each individual process. However, this is in line with the current way of detecting complex cellular states, which require more than one marker for definitive identification. Notwithstanding, the ability to easily identify these processes simultaneously gives this tool the advantage of an overview of the population dynamics related to these processes.

20There is a need for the development of objective, computer generated, analysis of morphological information of cells and tissues. Recently, a tool to predict the outcome of breast cancer patients based on morphometric analysis of tissues performed by a software developed by a machine learning-based method far outmatched other methods for predicting survival, indicating the potential for computer-based morphometric analysis.

21It is also important to point out that the NMA tool is an easy to use and objective quantification of senescence and nuclear irregularities of adhered cells in culture, two processes normally assessed by subjective analysis of morphology or intensity of β-galactosidase labeling, respectively. Therefore, we believe that this tool can provide important new insights in the biology of complex cellular populations under normal and, most importantly, stressful conditions both in basic cell biology and, potentially, in pathological analysis.

22Protocol uses Image Pro Plus 6.0 (IPP6 - Media Cybernetics, Bethesda, MD, USA), Image J 14.45 (NHI, Bethesda, MD, USA; http://rsbweb.nih.gov/ij/), PASW Statistics 18.0 (formerly known as SPSS - http://www.spss.com/software/statistics/), Microsoft Excel 2003 or 2007 (Microsoft, Redmond, WA, USA) and Open Office Calc (www.openoffice.org). NMA plugin, spreadsheets and sample images are available at http://www.ufrgs.br/labsinal/nma/.

23Brief instructions for using the NMA procedure (see Text S2 file for detailed instructions) are:

24Take DAPI images (with at least 300 dpi; preferentially.TIFF format) of at least 100 nuclei in each condition of experiment; do not take images of fields in high confluence;

25Open an image on IPP6 or Image J (containing the plugin of NMA) and correctly mark the surrounding of nuclei to assess morphometric data to NMA, i.e. area, aspect, area box, radius ratio and roundness;

26After obtaining raw data, choose a group of nuclei from the control condition and exclude nuclei in mitosis or with clear abnormalities. Use the data from these nuclei to set the parameters for the normal population. Paste data from these nuclei in the spreadsheet named “Normal Nuclei and Settings”, from NMA file from Excel or Open Office; choose the number of SD that best defines the ‘normal ellipse’; if using Excel, don’t forget to erase the formula in blank cells of the NII column;

27Paste data from all conditions - control and for each treatment - into the columns named “Treated Nuclei (for setting)”, in the “Normal Nuclei and Settings” spreadsheet of NMA file from Excel or Open Office; thereafter, choose the number of SD to define the thresholds that better separate the populations;

28Create as many spreadsheets as needed (one to each condition) and paste data from analyzed nuclei into them;

29Analyze the percentage of nuclei in each population as well as the averages of the populations.

30U87, C6, HeLa and HCT116 cell lines were purchased from ATCC (American Type Cell Collection, ATCC, Rockville, MD) and were maintained in DMEM supplemented with 10% FCS in 5% of CO2 at 37°C.

31Briefly, cells were pre-incubated with zAsp or vehicle for 1 h, followed by cisplatin treatment for 24 h. After this, for flow cytometry analysis, supernatant and cells were harvested, washed once with PBS 1x and once with annexin binding buffer, and incubated with a solution containing 6 µM of PI and 0.5 µL per sample of annexin V-FITC (Sigma Chemical; St. Louis, MO, USA) for 30 min, as indicated by the manufacturer, followed by flow cytometry. For microscopy, cells were washed once with PBS 1x and once with annexin binding buffer, and incubated with a solution containing 6 µM of PI and 1 µL of annexin per sample for 30 min, as indicated by the manufacturer. After, cells were fixed with paraformaldehyde 2% in PBS, stained with DAPI and images were acquired and overlaid.