1.3. VISIÓN ARTIFICIAL
1.3.4. PROCESAMIENTO Y ANÁLISIS DE IMÁGENES
1.3.4.2. Análisis de Imagen Digital
2:6 VMRhoA 2:5 VMRhoA 1:3
Figure 5.6 Morphology o f D L D -1 cells expressing V 14RhoA
Phase contrast microscopy im ages o f V M R hoA 2:6 (a. b), V M R hoA 2:5 (c, d) and V M R hoA 1:3 (e, f) cells growing on tissue culture plastic. Cells were grown for a total o f 9 days in the absence o f tetracycline. Images in (b), (d) and (f) show (a), (c) and (e) at 2 x higher m agnification. White arrows indicate multmucleate cells, black arrows indicate rounded up cells and the red arrow show s a dying cell Scale bars = 50 p.m.
Western blots are included to illustrate the relative levels o f V M R hoA expression (Figure 4 .6). C ells were grown in the presence (+) or absence (-) o f tetracycline for 3 days prior to lysis and V MRhoA expression detected using an antibody to the 9E10 c -m \c epitope.
Figure 5.7 Actin organisation and V M R hoA protein localisation in V M R hoA 1:3 cells Cells were grown in the absence o f tetracycline for 4 days and stained with antibodies to the 9E10 c-m yc epitope to locate V M R hoA protein (b, d, f). Cells were counterstained with TRITC-conjugated phalloidin to visualise F-actin (a. c, e). Basal (a, b), mid (c, d) and apical (e, f) sections were collected by confocal m icroscopy to illustrate protein localisation throughout the cell. The red arrow indicates a retracted cell, blue arrows indicate vacuoles and white arrows indicate possible new septum formation in nuclei. Scale bar = 20 pm
Figure 5.8 cells
Actin organisation and V M R hoA protein localisation in V M R hoA 2:6
Cells were grown in the absence o f tetracycline for 4 days and stained with antibodies to the 9E10 c-m yc epitope to locate exogenous protein (b, d). Cells were counterstained with TRITC-conjugated phalloidin to visualise F-actin (a, c)
Basal (a, b) and mid (c, d) sections were taken by confocal m icroscopy to illustrate protein localisation throughout the cell.
The red arrow indicates a retracted cell with a dense network o f actin filament bundles. Scale bar = 20 pm
5.2.3.1 : Analysis of cell size bv flow cytometry
To investigate whether the increase in cell area observed for some V14RhoA- expressing cells reflected a change in cell size, the scattering of light by cells from different cell lines was investigated by flow cytometry. Light scattering by cells can indicate a number of things, in particular changes in forward light scatter correlate with changes in cell size, and a larger cell will scatter more light than a smaller cell. In this assay, the percentage of cells falling within a designated range of light scattered, determined by taking the main population of control DT-1 cells as a reference (Figure 5.9), was calculated. This figure gave an indication of how alike the two cell populations were, although it failed to indicate whether cells differed because of an increase or a decrease in cell size. Therefore changes in forward light scatter were also compared to determine cell size. Overall the V14RhoA cells scattered significantly more light in a forward direction than the control DT-1 cells, indicating that there is an increase in the average size of VI4RhoA-expressing cells. The fact that there was no significant decrease in cells falling within the designated range, was most likely due to a large percentage of small DT-1 cells falling outside of the designated area, while a similar percentage of large V14RhoA cells also fell outside the area.
□ V 1 4 R ho 1:3
forward s c a tte r
Figure 5.9 A nalysis o f cell size by flow c\tom etr\
Control DT-1 cells and V M R hoA 1:3 cells were compared for differences in cell size using light scatter (Chapter 2.4,11,1). A box was drawn around the main population o f control DT-1 cells and the percentage o f DT-1 or V M R hoA 1:3 cells falling within this range calculated (area). The percentage o f cells scattering light above a defined threshold level was also calculated to further study cell size (forward scatter). Samples o f raw data for each cell line are show n below and the average o f two sets o f data show n above.
V U R lio A 1:3
Forward scatter Forward scatter
5.3 ; Cell proliferation
To establish whether expression of constitutively active or dominant negative RhoA altered the growth rate of DLD-1 cells, a number of complimentary approaches were used to analyse their growth properties.
5.3.1 : Measurement of cell growth by optical density
Multi-well growth assays were initially chosen to analyse cell proliferation because they provided a rapid method for screening a large number of different cell populations (Chapter 2.4.9). The relative growth rates of control cells were measured to determine whether any changes affecting growth regulation had occurred as a result of multiple rounds of cloning and selection (Figure 5.10). Cells were grown in the presence and absence of tetracycline to establish whether the antibiotic was inhibitory to cell growth. The relative growth rates of each control cell population were also studied under different serum conditions to compare transformed growth properties, since growth factor-independent growth is characteristic of a transformed phenotype.
From the combined data, it was observed that although tetracycline had a small inhibitory effect on the growth of each cell line, under all conditions, this was not significant (Figure 5.10). Most striking was that in the absence of tetracycline and under each set of growth conditions, DLD-1 cells increased in cell density more rapidly than either DT-1 or pSVz cells. In the absence of tetracycline and again under each set of growth conditions, both DT-1 and pSVz cells showed similar patterns of growth and reached similar final cell density readings. Since no significant difference between growth and morphology of DT-1 and pSVz cells was detected, the DT-1 cells, which represented the parental control for all Rho-expressing clones, were used as the control for all further assays. The data also indicated that each control cell line was able to grow in the absence of serum and thus showed one of the characteristics of transformed growth. However the results for cells grown in the complete absence of serum illustrated a problem of this assay system (Figure 5.10). The data showed a sudden increase in the growth of DLD-1 cells between days 9 and 12, when the DT-1 and pSVz cells continued to increase in cell density at a rate similar to that shown over the previous 9 days. The DLD-1 data is somewhat misleading because the spectrophotometer was observed to measure readings from 0 to 3 and then jump to readings of 9 and above, indicating intensity saturation, with no readings falling
between 3 or 9. Thus the assay was less sensitive to changes in cell number at high density.
Similar multi-well cell density assays were carried out with V14RhoA and N19RhoA cell lines (Figure 5.11). The most striking result was the greatly inhibited growth exhibited by the V14RhoA 1:3 line. These cells already showed a reduced density when compared to control DT-1 cells at the beginning of the assay and by the end of the assay had failed to reach even one tenth of the final density of DT-1 cells. The observation that V14RhoA 1:3 cells showed an increased rate of expansion at nine days indicated that these cells had probably attained a density at which they could successfully grow. Such a pattern of growth had been observed during routine cell culture of this cell lines, where cells could only be successfully cultured if maintained at a density of at least 50% confluence. The two lower expressing clones 2:5 and 2:6, grew significantly more than the 1:3 cells and attained final densities in the same range as those of control DT-1 cells (Figure 5.11). When cells were grown in the absence of serum, growth of V14RhoA 1:3 cells was almost completely inhibited (Figure 5.11). These results further indicate an inability of VI4RhoA-expressing cells to grow in the absence of stimuli, either from other cells or from the growth medium. Clones 2:5 and 2:6, which expressed similar levels of protein, showed similar patterns of growth in the presence of serum, indicating that effects were due to V14RhoA expression rather than to clonal variation, although the small increase in expression observed in the 2:5 clone resulted in reduced growth in the absence of serum. Despite western blot analysis having indicated that all three V14RhoA cell lines expressed similar levels of V14RhoA in the presence of tetracycline, they did not all grow at the same rate when the antibiotic was included in the growth medium. One possible explanation is that the V14RhoA 1:3 cells have a more leaky expression system where levels of V14RhoA expression very quickly increase with degradation of tetracycline. It is possible that the long-term expression of differing levels of VMRhoA resulted in cells with different phenotypes that were not reverted in the presence of tetracycline.
In contrast to the effects of V14RhoA, expression of N19RhoA had little effect on cell growth rates (Figure 5.11). All cells reached a similar final density to control DT-1 cells by the end of the two week period, and the presence or absence of tetracycline had no affect on this.
G ro w th in m edia c o n ta in in g 10% PCS 100 D L D -1 — D L D -1 * DT-1 p S V 2 - DT-1 p S V 2 0.01 0 2 4 6 8 10 D a y s p o s t r e m o v a l o f t e t r a c y c l i n e 12 14
G row th in m edia co ntainin g 1% PCS 100 E C 10 § 1 DLD-1 0 1 "5 O ■O' ■ DT-1 p S V 2 0 01 0 2 4 6 8 10 12 14 D a y s a f t e r r e m o v a l of te t r a c y c l in e G row th in m edia c o n ta in in g 0% PCS 100 DLD-1 D L D -1 — -D T-1 - - DT-1 p S V 2 p S V 2 0 01 D a y s p o s t re m o v a l o f t e t r a c y c l in e
F igure 5.10 Proliferation o f control DLD-1, DT-1 and pSV^ cells, assayed by spectrophotometric analysis o f cell density changes overtim e.
Cells were seeded in growth media containing 10% PCS and tetracycline. Tw o days later the cells were washed once and fresh media added, with varv ing amounts o f PCS as indicated, and w ith or without tetracycline Pollow ing this, the m edia were changed ever} 2-3 days and plates stained with methylene blue on the same days for spectrophotometric analysis (Chapter 2.4.9 1). Dotted lines indicate growth in the presence o f tetracycline and continuous lines show growth in the absence o f antibiotic. Error bars indicate standard deviation over 3 or more expenm ents
100 00 ■ DT - 1 10% serum V M R h o A 1 : 3 V U R h o A 2 : 5 V M R h o A 2 : 6 10.00 ™ 1 00 0 10 0.01 0 2 4 6 e 10 12 14
Days post rem oval of te tracycline
10 00 DT-1 VM RhoA 1:3 VMRhoA 2:5 VMRhoA 2:6 0% serum VM RhoA 1:3 VM RhoA 2:5 VMRhoA 2:6 E Q O I 03 o 0.01 0 2 4 6 8 10 12 14 D a y s p o st re m o v a l o f te tra c y c lin e 100 0 0 DT-1 N 1 9 R fio A 3 :3 N 1 9 R fio A 3 :3 N 1 9 R fio A 2 :1 1 0 .0 0 N 1 9 R h o A 2 :1 0 10 0 01
D a y s p o s t rem o v a l of tetrac y c lin e
V H R llo A 2:6 + V M R hoA 2:5 V M R hoA 1:3 + N MRhoA 2:1 + N M RhoA 3:3 + À A
Figure 5.11 Proliferation o f control and RhoA-expressing DLD-1 cells, assayed by spectrophotom etnc analysis o f cell densit> changes overtim e.
Cells were seeded in the presence o f tetracycline Two days later w ells were washed and re-fed w ith media w ith or w ithout tetracycline (day 0). Following this the media were changed every 2-3 days and plates stained with m ethylene blue on the same days for spectrophotometric analysis (Chapter 2.4.9.1). Dotted lines indicate growth in the presence o f tetracycline and continuous lines show grow th in the absence o f antibiotic. DT-1 cells grown in the absence o f tetracycline were used as a control. Error bars indicate standard deviation over a minimum o f three independent assays. Western blots are as show n previously (Figures 5.4 and 5 6).
5.3.2 : Proliferation rates of RhoA-expressing DLD-1 cells
Since the data above measured changes in optical density of cells over time and not actual cell numbers, it was decided that the growth changes observed should be confirmed using another assay. Changes in cell numbers over the same time-course and cultured in the same 96-well plate format, were therefore determined by cell counting. The growth rate of DT-1 cells slowed during the second week of the assay suggesting that the cells showed some growth regulation in the form of contact inhibition (Figure 5.12). However, as shown previously, cells continued to overgrow each other (Figure 5.2), but at a reduced rate of growth compared to sub-confluent cells. The rate of growth may also appear slower than it really is, if cells growing on top of other cells are less adherent than when growing on plastic and thus are dislodged more easily during medium changes. As had been seen in the optical density assay, VMRhoA 1:3 cells showed a significantly slower rate of growth compared to DT-1 cells (Figure 5.12). The lower expressing V14RhoA 2:6 cells showed a similar rate of growth to DT-1 cells, with the only difference observed towards the end of the assay when the V14RhoA- expressing cells continued to proliferate (Figure 5.12). Although VMRhoA 1:3 cells grew consistently slower than 2:6 cells the greatest difference was seen over the first two days after tetracycline withdrawal and may once again be the result of more leaky regulation in the 1:3 cells combined with higher levels of expression. Such high levels of VMRhoA expression inhibited the growth of DLD-1 cells at low density, as had been seen previously (Figure 5.11). Interestingly, the final density of V14RhoA 1:3 cells was observed to be equal to that of V14RhoA 2:6 cells around day six, indicating that the rate of proliferation of the 1:3 cell population was around half that of the 2:6 cell population. This correlates with the observation that approximately twice as many cells of the 1:3 clone expressed detectable levels of V14RhoA protein than of the 2:6 clone (Figures 5.7 and 5.8).
As had been observed by cell density analysis of cell growth, both N19RhoA cell lines showed a similar rate of proliferation to DT-1 cells (Figure 5.12). N19RhoA 2:1 cells showed an almost identical pattern of growth to control DT-1 cells, also slowing their rate of growth as cells became confluent. NMRhoA 3:3 cells began to slow down after the first few days, with cell numbers remaining constant between days 7 and 9. Cell numbers then increased after day 9, a phenomenon that had also been observed in the optical density assay (Figure 5.11).
1 000000 100000 <v a . 1/1 10000 01 Ü DT-1 o 01 E 3 C V 1 4 R h o A 1 3 1000 V 1 4 R h o A 2:6 100 0 2 4 6 8 10 12 14
days after removal of tetracycline
1000000 100000 --- 10000 -- DT-1 N 1 9R hoA 2:1 1000 N 1 9 R h o A 3:3 100
days after removal of tetracycline
V M R hoA 2:6 + V M R hoA 1:3 + N 19R hoA 2:1 + N 19R hoA 3:3 +
F igure 5.12 Rates o f proliferation o f R hoA-expressm g DLD-1 cells
Cells were seeded in the presence o f tetracycline. Tw o days later w ells were washed to remove tetracycline and re-fed with medium lacking tetracycline (day 0). Follow ing this the medium was changed on days 2, 5, 7 and 9. Cell counts were made on d a \s 0, 2, 5, 7, 9 and 12.
Error bars indicate standard deviation over 3 or more assays Western blots are as shown previously (Figures 5 4 and 5.6).
5.3.3 ; Flow cytometric analysis of cell cycle distribution
To study the effects of RhoA activity on the growth of DLD-1 cells further, the distribution of the total cell population across the cell cycle was investigated by flow cytometric analysis of propidium iodide-stained cells. This method was chosen primarily to determine whether there was a block in the cell cycle of V14RhoA 1:3 cells, leading to a build-up of cells in one phase of the cycle and hence the reduction in growth rate observed. It was also hoped that the assay could be used to detect a sub-Gi hypodiploid apoptotic cell population, if present, as has been shown previously (Esteve et al, 1998).
A significantly larger percentage of the total VMRhoA 1:3 cell population was found to be in Gz/M-phase of the cell cycle than control DT-1 cells (Figure 5.13) and occurred at the expense of cells in Gi-phase. Although none of the results for V14RhoA 2:6 cells differed significantly from that of DT-1 cells, there was a very small increase in the percentage of cells in Gz/M-phase. Ideally the high-expressing cells would be sorted with an anti 9E10 c-myc antibody first and then these cells assayed for DNA content. However, it was not possible to use the same fixation protocol to detect both the 9E10 c-myc epitope and DNA content simultaneously. In contrast to VMRhoA cells, it was observed that there was generally a greater percentage of N19RhoA 3:3 cells in Gi with a corresponding decrease in cells in S-phase (Figure 5.13), although there was some variation between assays.