1 Exploring the Effects of Anastasis, the Reversal of Late Stage Apoptosis Melanie Abou Jaoude, Jun Yin, Stephen Arce, and Yiider Tseng Department of Chemical Engineering, College of Engineering, University of Florida A nastasis refers to the recovery of cells from the late stages of a p optosis, even after DNA damage. It has been found that cells which under go anastasis display phenotypic changes characteristic of genomic change including focus formation [ 1 ] Here NIH 3T3 fibroblasts were used to successfully reproduce this result by treating them with 10% dimethyl sulfoxide (DMSO) Mutated cell phenotypes were compared to a control utilizing both morphology and cell migration behavior. T h e treated cells were found to have smaller areas and greater variance in cell area. They also had smaller average displacements and small er centroid step si zes, but larger nucleus centroid step sizes This s uggests that the mutant cells have greater migration ability b ut less adhesive ability which may be a result of anchorage independent growth, which is worth further investigation More work is needed to characterize the changes in the se cells more thoroughly but future research into anastasis co uld provide fascinating insight into carcinogenesis and cancer re c urr ence Introduction Apoptosis, also known as programmed cell death, is a specific type of cell death characterized by cell contraction and blebbing follo wed by cell structure breakdown, nuclear condensation and DNA fragmentation  It happens at predictable points during development to dispose of unwanted and abnormal cells, but can also be triggered in response to certain drugs, irradiation, and other d amage to remove injured cells [ 1 ] Many malignant tumors have developed a resistance to apoptosis [ 3 ] while too much apoptotic behavior may contrib ute to degenerative diseases [ 4 ] Apoptosis had been thought to be irreversible once the caspase 3 cascade h ad been triggered, releasing active proteases that destroy the structural and functional components of the cell. However, a 2012 paper by Tang et al. from Johns Hopkins School of Medicine reported the recovery of a number of different cell types even after the activation of caspase 3 and DNA damage in a mechanism they called [ 1 ] Interestingly, the cells that undergo ana stasis exhibit higher rates of genomic change from errors in DNA repair In the case of NIH 3T3 fibroblasts, these changes wer e characterized by focus formation, indicating a loss of contact inhibition of growth, and proliferation in soft agar, indicative of anchorage independent growth. These phenotypes were not found in control cells, but are known to be hallmarks of cancer cel ls [ 5 ] These results point to anastasis as mechanism that contributes to evolution as a source of genetic diversity
2 tha t also has the potential to be carcinogenic. The first goal of this project is to duplicate the results of this study to verify its find ings using NIH 3T3 fibroblasts The second goal is to further investigate the mutations cause d by anastasis by comparing size and displacement data Methods Cell Culture The cells used in this experiment were NIH 3T3 fibroblasts, cultured in DMEM conta ining 10% fetal bovine serum. All cells were maintained in a humidified incubator at 37 C in a 10% CO 2 environment Inducing Reversible Apoptosis To induce apoptosis, the plate of fibroblasts at roughly 70% confluence were treated with 10% dimethyl sulfoxide ( DMSO ) for 20 hours, at which point they appeared to be in the final phases of apoptosis. The inducer was then washed aw ay and replaced with fresh DMEM During the late stages of apoptosis the cells ha d detached from the ir plate s, so a significant number we re lost when the inducer is washed away. The cells that remained were further incubated for three weeks before there were enough for experimental use and a line of these DMSO treated cells was saved for further use. Fluorescence and Microscopy The cells were transfected with green fluorescent protein DNA using Lipofectamine2000 ( Invitrogen, Carlsbad, CA ), following the standard protocol. Glass bottom plates were bombarded with green/blue fluorescent beads using the Biolistic Particle Delivery System (Bio Rad, He rcules, CA ) to provide a static reference when tracking cell motion. These plates were then coated with fibronectin to provide a framework for cellular adhesion. Ce lls were passed onto the prepared plates 24 hours before microscopy, then their nuclei were stained with HT33342 nucleus dye( Sigma Aldrich, St. Louis, MO ). Images were taken using Nikon TE 2000 microscope (Nikon, Melville, NY), equipped with a X Cite 120 PC fluorescent light source (EXFO, Ontario, Canada), Cascade:1K CCD camera (Roper Scientific, Tucson, AZ) and an on stage incubator with CO2 supplementary system (In Vivo Scientific, St. Louis, MO). During imaging, cells were maintained in an atmosphere co ntaining 10% CO 2 The images were sequentially ta ken using a 20x objective lens, a 200ms exposure time and 50% light source power. 3x3 binning was used, and photos were taken in 1 minute time loops Image processing and data analysis Images were process ed in MATLAB (The Mathworks, Natick, MA) using custom made programs. These programs obtained cell nucleus and centr oid displacement to compare cell motion as well as the area of the cells to compare relative size. T tests and histograms were used to determine the significance of the results. Results and Discussion In agreement wi th the literature, after 20 hours of being treated with 10% DMSO, al l the cells appeared to have undergone apoptosis though DNA damage was not confirmed via microscopy (figure1a) Also in agreement with the literature, upon washing away the DMSO with fresh medium, a significant number of cells recovered. After culturing for three more weeks focus formation was visible (figure1b.) which confirms the presence of genomic change as one of the phenotypic changes noted in the literature
3 Visual inspection of the images showed extreme variation in shape This variation is difficult to characterize and more precise analysis is needed (figure 2) Figure 2 : a)DMSO ce ll with many protrusi ons b) a more round ed DMSO cell c and d ) control cells Area comparisons showed that he cells that had been treated wi th DMSO had a smaller mean are a with a variance more than twice that of the control cells ( t able 1 ) The overall distributions were significantly differen t (figure 3 ) with m an y outliers, so the median better represents the data The median cell area of the DMSO cells was 750 pixels, while the median of the control cells was 1060 pixels. Table 1 : T test of area measurements Figure 3 : Histograms of c ell area measurements. The DMSO group ha s both smaller are as and greater variance in area. Nucleus and cell centroid displacement data both showed that the treated cells had smaller displacements over 1 hour periods (figure 4 ) with excellent agreement between the two parameters I nterestingly the me dian nucleus centr oid step size is greater for the DMSO cell s ( 0 56 ) than the control ( 0 .45 ) while the me dian cell centroid step size is smaller ( 0 .4 7 0 .6 5 ) 0 200 400 600 200 600 1000 1400 1800 2200 2600 3000 DMSO Cell Areas (pixels) 0 200 400 200 600 1000 1400 1800 2200 2600 3000 Control Cell Areas (pixels) Figure 1 : a) Cells undergoing apoptosis and b) an example of focus formation (right). a. b
4 Figure 4 : histogra ms of step siz e data. The DMSO treated cells had larg er median nucleus s tep size, but smaller median control step size. T hese results may be because the t reated cells form weaker attach ments to t he fibronectin coated plates, preferring instead to g row independent o f anchorage These result s T tests showed more variance in the DMSO cell data in all categories except for the control centroid step size Table 2 : Nucle us centroid step size t test Table 3 : Ce ll centroid step size t test. Conclusions The result of the Tang et al. p aper was successfully reproduced; after treatment with DMSO, NIH 3T3 cells achieved apoptotic states o nly to recover with phenotypic changes consistent with what was previ ously reported Fibroblasts that have undergone anastasis after treatment wit h DMSO have smaller areas, and greater variance in area measurement supports the exist ence of greater genetic diversity in these recovered cells. These cells also have a s maller average displacement and median centroid step size d espite a larger median nucleus centroid step size I t is suspected but not confirmed that the cells that have undergone anastasis h ave a reduced ability to adhere to cell culture plates. This might explain their sma ller average size and greater nucleus st ep size; Figure 4 : Nucleus a n d cell centroid displaceme nt comparison 0 5 10 15 20 25 1 2 Nucleus Displacement (m)*** 0 5 10 15 20 25 1 2 Centroid Displacement (m)*** Control n=24 DMSO n=30 Control n=24 DMSO n=30
5 without as many anch ors to the plate, the cell body will not spread out as much on the plane of the plate, and th e nucleus can mov e more quickly with the rest of the cell i nstead of remaining more fixed as sampling occurs. Future work The sample size o f this data, particularly the displacement data, is regrettably small. Better conclusions might be drawn from this work with a larger sample size. Un fortunately, there is a significant amount of noise in the data used Some cells fluoresced particularly bright ly creating disks of light noise that interfered with the segmentation, making it through the filters i n the code used. Adjustments were made, but not all the noise was removed (figure 5 a d) This noise from bright fluorescence was presen t in both sets of cells, so statistically our results are still likely to have significance but a number of the DMSO cells had extreme protrusions that could not be captured in d etail (figure 5 e) This is likely to have the greatest impact on the area measurements and the centroid displacement data. Because of this noise, no attempt to measure perimeter or circularity was made. Improv ing the segmenting further might allow better data to be obtained improving the observati ons here and allowing the comparison of parameters like circularity in addition to those compared here. I mmunostaining tec hniques could be used to compare the actin cytoskeletons of these cells to b etter investi gate the potential of anastasis causing reduced adhesion to cell culture plates It might also be worthwhile to repeat the reverse apoptosis process on t he DMSO treated cell l ine to amplify the effects of anastasis and increase the percentage of mutated cells. It would be interesting to see whether repeated anastasis might lead to an overall resistance to apoptosis. Experiments along this vein might also be tried with other ce ll lines. For example, OSE10 epithelial cells could provide fascinating insights into carcinogenesis, as skin cancers have been linked to irradiation from sun exposure and irradiation can cause apoptosis. A number of epithelial tumors are solid and resistant to apoptosis  and it would be inter esting to see if this might be because of anastasis However, this particular cell line was not tested in the literature, and further research would be required to determine an appropriate treatment to achieve high rates of anastasis a. b c d e Figure 5 : Noise pollution and segmentation issues a)ImageJ threshold r esult. b) Initial programming threshold. c and d) different frames from the final segmentation code used. e) initial (left) an d impr oved(right) cell outlines for DMSO treated cells Even the improved code had mixed results
6 References [ 1 ] H. Tang, H. Tang and D. Montell, '"Stress Induced Mutagenesis, Genetic Diversification, and Cell Surviva l via Anastasis, the Reversal of Late Stage Apoptosis," Stress Induced Mutagenesis 01/01, pp. 223 241.  M.L. Coleman and M.F. Olson, '"Rho GTPase signalling pathways in the morphological changes associated with apoptosis," Cell Death Differ. vol. 9, no. 5, May, pp. 493 504. [ 3 ] H.L. Tang, '"Cell survi val, DNA damage, and oncogenic transformation after a transient and reversible apoptotic response," Mol.Biol.Cell vol. 23, no. 12, pp. 2240 -2252. [ 4 ] J.M. Brown and L.D. Attardi, '"The role of apoptosis in cancer developmen t and treatment response," Nature reviews.Cancer JID 101124168 no. 1474 175; 1474 175. [ 5 ] H. Rubin, '"Cell cell contact interactions conditionally determine suppression and selection of the neoplastic phenotype," Proceedings of the National Academy of Sciences vol. 105, no. 17, April 29, pp. 6215 6221.