↓107 |
Immunological rejection during pregnancy is characterized in the mouse by necrotic tissue and infiltration of immune cells at the site of implantation. This site is usually called “resorption” and resembles human abortion (or miscarriage). Fig. 24 depicts a typical abortion event, near to a healthy fetus at its implantation site. As seen in the picture, resorptions are hemorrhagic necrotic areas of the tissue that are considerable smaller than the normal placentas and their fetuses.
Fig. 24 : Representative picture of the feto-maternal interface in a murine pregnancy | ||
Resorptions and healthy fetuses can be clearly differentiated from each other due to the hemorrhagic and necrotic appearance of the resorption, as well as due to its smaller size as compared to the healthy implantation site. |
↓108 |
Hematoxilin-Eosin staining of healthy placental tissue made it possible to identify the different cell types present in a murine placenta on day 14 of gestation. Fig. 25 shows a typical placenta, staining in which all types of cells can be appreciated.
The outer layer of the placenta, even after dissection from the uterus, is contaminated with maternal decidual cells which usually appear as a thin layer of small cells (Nagy et al. in: Manipulating the mouse embryo, 2003). In the mature placenta, the spongiotrophoblast forms the middle layer of the placenta between the outermost giant cells and the innermost labyrinth. Finally, the last type of cells found in the placenta are the glycogen cells, which appear within the trophoblast layer (reviewed in Cross, 2005).
Fig. 25 : Representative HE staining showing the different cell types in a murine placenta | ||
↓109 |
The application of an adenoviral vector containing HO-1 was employed in order to analyze the effect of a specific up-regulation of HO-1 on the pregnancy outcome of mice undergoing spontaneous immunological abortion.
As expected (Chaouat et al., 1995; Clark et al., 1997), DBA/2J-mated CBA/J mice presented significantly higher abortion rates when compared to the normal pregnant combination CBA/J x BALB/c. Interestingly, abortion-prone mice receiving 1.105 PFU AdHO1/GFP showed a significant diminution in the abortion rate as compared to PBStreated abortion-prone mice (p<0.05) or mice receiving 1.105 AdEGFP (p<0.05), pointing out a beneficial effect of HO1 on pregnancy outcome (Fig. 26). Since mice receiving the adenoviral construct without HO-1 showed similar abortion rates as the abortion-prone group, it is to speculate that the injection of sole adenoviral particles -known to be immunogenic- did not significantly affect the pregnancy outcome (Fig. 26). However, a small augmentation in the abortion rate was observed in the group receiving the AdEGFP vector, indicating that injection of AdHO-1 had a beneficial effect despite the immunological response that the injection of adenoviruses implied.
Fig. 26 : Specific up-regulation of HO-1 improves pregnancy outcome in a murine model of abortion | ||
Data are represented by medians ± 75% quartiles. N.P.= Normal Pregnancy; S.A.= Spontaneous abortion. Circles and/or asterisks above box plots indicate outliers. Significant differences between groups are indicated as asterisks above lines. *p≤ 0.05, **p≤ 0.01 and *** p≤ 0.001, as analyzed by the Kruskall-Wallis non-parametric test between all groups, followed by the Mann-Whitney Utest for two particular groups. |
↓110 |
Interestingly, abortion-prone mice receiving 1.108 PFU AdHO-1/GFP presented a slight and not significant diminution in the abortion rate (n = 8) when compared to mice receiving 1.108 AdEGFP (n = 7), suggesting that a too high HO-1 concentration is not pregnancy-protective. This data is shown as Table 14 in the appendix. These two groups will not be taken into account for the graphics, since its application did not show any effect on the pregnancy outcome. Nevertheless, the data regarding these two groups are shown in the appendix as Table 14.
Since the percentage of abortion rate depends on the number of implantations, the number of implantations was calculated for every group. As it can be observed in Fig. 27, the number of implantations was comparable between all groups (Fig. 27). Considering this, it can be stated that the up-regulation of HO-1 has beneficial effect of pregnancy outcome.
Fig. 27 : The number of implantation sites was similar between all experimental groups | ||
Data are represented by medians ± 75% quartiles. N.P.= Normal Pregnancy; S.A.= Spontaneous abortion; Circles below box plots indicate outliers. No significant differences were found between the groups as analyzed by the Kruskall-Wallis non-parametric test between all groups, followed by the Mann-Whitney Utest for two particular groups. |
↓111 |
In order to analyze whether the effect of HO-1 up-regulation is noticeable in the levels of HO1 at the feto-maternal interface 9 days after treatment, the expression of HO-1 was measured at mRNA and protein levels in placental tissue. Table 7 shows the mRNA levels of mouse HO-1 in all groups, where a slight augmentation of mouse HO-1 mRNA can be seen in the group receiving AdHO-1/GFP. However, this augmentation is not statistically significant. HO-1 mRNA levels where significantly augmented in placentas of mice receiving the higher doses of AdHO-1/GFP in comparison with mice receiving low doses (Table 17, Appendix).
By measuring mRNA of human HO-1, the one coded in the adenovirus used for the in vivo experiment, only few samples showed amplification curves and most of them did not show any amplification after 40 cycles. This may be due to the fact that the mice were killed on day 14 of pregnancy, 9 days after injection of the adenoviruses. This makes sense as it is reported that adenoviruses remain between 5 and 20 days post-infection (Dai et al., 1995).
Table 7 : No significant differences in the placental HO-1 mRNA levels between the groups
N.P (n=5) |
S.A. (n=8) |
S.A. + AdHO-1/GFP (n=5) |
S.A. + AdEGFP (n=7) |
|
Mean |
0.05076 |
0.03590 |
0.04565 |
0.05047 |
SD |
0.04651 |
0.01383 |
0.02169 |
0.01575 |
Median |
0.02906 |
0.02967 |
0.04404 |
0.04443 |
↓112 |
Protein levels of HO-1 were analysed by Western Blot using homogenates of whole placental tissues using b-actin as protein reference. Intensity of the bands was analyzed using the Quantity One Software (BioRad), and a HO-1/β-actin ratio was calculated for each sample. Fig. 28 shows an example of the β-actin and HO1 Western Blot bands obtained.
Fig. 28 : Example of the Western Blot from placental samples | ||
1 and 5) CBA/J x BALB/c placentas 2 and 6) CBA/J x DBA/2J placentas 3 and 7) CBA/J x DBA/2J + AdHO-1/GFP placentas 4 and 8) CBA/J x DBA/2J + AdEGFP placentas |
Bands were analyzed using a densitometer being the densitometric units for HO-1 normalized to those of b-actin. The results are showed in Table 8, showing mean, standard deviation and median for each analyzed group. No differences were obtained in HO-1 protein levels between the groups, as analyzed by the Kruskall Wallis test between all groups, as well as no differences between the groups, as analyzed by the Mann Whitney test. A trend towards a diminution of HO-1 expression in the abortion prone-group when compared to the normal pregnant group could be observed (Table 8), in accordance with previous results from our group (Zenclussen et al., 2005). A tendency towards an augmentation was observed in the group receiving AdHO-1/GFP as well as in the group receiving AdEGFP (Table 8).
↓113 |
N.P. (n=5) |
S.A. (n=5) |
S.A. + AdHO-1/GFP (n=4) |
S.A. + AdEGFP (n=3) |
|
Mean |
0.3188 |
0.1370 |
0.3924 |
0.3681 |
SD |
0.2296 |
0.1524 |
0.2978 |
0.2267 |
Median |
0.3165 |
0.05028 |
0.2772 |
0.4157 |
Fig. 29 : HO-1 is highly expressed in all murine placental cell types as well as in the decidua as analyzed by immuno-histochemistry in paraffin-embedded tissues. | ||
Immunohistochemical analysis of HO-1 in paraffin-embebbed tissues. A) shows intense HO-1 staining on giant cells; B) depicts HO-1 expression on spongiotrophoblasts. C) HO-1 expression was also found on decidual cells. D) represents the negative control of the staining. All pictures were taken using a 20X magnification of the objective lens. |
Protein levels of HO-1 were also measured by immunohistochemistry in order to analyze the HO-1 expression in each different cell type and have more detailed information regarding local HO-1 expression. HO-2 expression was as well analyzed by immunohistochemistry.
↓114 |
Again, no differences were observed between the groups for murine HO-1, and also no differences were found for HO-2 expression. Representative pictures for HO-1 staining are shown in Figures 29A-C, were it can be seen that HO1 is highly expressed throughout the placenta as well as in decidua, with its highest expression in giant cells. Since giant cells are very important for the onset of placentation, it is tempting to assume that HO-1 may play an important role in placentation. Fig. 29D shows the negative control of the staining. HO-2 expression was found throughout the whole placenta, with similar patterns of expression as for HO-1.
Although no differences were observed in the levels of HO-1 between the groups, it was possible to detect human HO-1 mRNA (derived from the vector) in some placenta samples from animals treated with Ad-HO-1. This suggests that the adenovirus reached the fetomaternal interface. The use of a therapeutic vector containing GFP made possible the analysis of the presence of the construct in the reproductive tract to confirm that the applied adenoviruses reached the feto-maternal interface. GFP expression was analyzed by fluorescence microscopy as well as by real-time PCR.
Fig. 30 : GFP+ cells were found in the placenta and not in the fetus when analyzed by fluorescence microscopy | ||
A) shows GFP+cells at the feto-maternal interface from abortion-prone mice receiving 1 x 105 PFU AdHO-1/GFP; GCs = giant cells; B) depicts GFP+ cells in decidual and placental cells but not in the fetus after the transfer of 1 x 105 PFU AdEGFP. Pictures were taken in a 20X magnification of the objective lens. |
↓115 |
GFP protein was in fact expressed at the feto-maternal interface in decidual cells as well as in giant cells, i.e. in the outer layer of the fetal-maternal interface as analyzed by fluorescent microscopy. Figure 30A shows a representative field of GFP+ cells found in cryo-sections of placental tissue. Figure 30B shows positive GFP expression in decidual and placental tissue, while no expression could be observed in fetal tissue.
Further, a semi-quantitative measurement of adenoviral particles for GFP was performed by real time PCR in liver and placental tissues as well as in fetus homogenates in order to determine the tissue distribution of the adenoviral particles after injection. The higher GFP expression was found in maternal liver, which was expected considering the high hepatic tropism of adenovirus if applied intra-peritoneally (Huard et al., 1995). A lower adenovirus expression was found in placental tissue, which confirms the GFP detection by fluorescent microscopy, and also very low levels of adenoviral particles could be found in fetus homogenates (Fig.31). Low adenovirus expression in fetuses after gene therapy was already described by other authors (Senoo et al., 2000). Because we homogenized fetal tissues using liquid N2, the quantified virus may represent only the virus present in the fetal skin as already reported by other groups (Senoo et al., 2000), since effective gene transfer into the fetal organs e.g. liver or kidney is only possible when applying the vector into the umbilical vein (Senoo et al., 2000). However, this GFP may be not expressed at the fetal tissue, as no GFP expression could be observed in the fetuses by fluorescence microscopy.
Fig. 31 : GFP DNA was mostly found in maternal liver after i.p. injection of AdEGFP as measured by real time PCR | ||
GFP expression (relative to ApoB) measured with real time PCR in mice receiving 1x105 PFU AdEGFP. The highest GFP expression was found in the maternal liver. Placentas and fetuses showed very low GFP expression. |
↓116 |
As HO-1 was previously described to affect the cytokine balance in transplantation models (reviewed in Katori et al., 2002) and considering that a proper Th1/Th2 balance is important for successful pregnancy, systemic and local cytokine expression by T cells was measured by flow cytometry. Spleen and decidual cells were stimulated by PMA/Ionomycin, process which is known to stimulate mainly memory T cells.
Fig. 32 : HO-1 up-regulation could significantly diminish the IL4/IFN-γ ratio in spleen and decidual lymphocytes as analyzed by flow cytometry | ||
Th1/Th2 ratios in spleen (A) or decidual lymphocytes (B) measured by the production of IFN-γ and IL-4 by flow cytometry after PMA/Ionomycin incubation in the presence of monensin. Data are represented by medians ± 75% quartiles. N.P.= Normal Pregnancy; S.A.= Spontaneous abortion; Circles and/or asterisks above box plots indicate outliers. Significant differences between groups are indicated as asterisks above lines. *p≤ 0.05, **p≤ 0.01 and *** p≤ 0.001, as analyzed by the Kruskall-Wallis non-parametric test between all groups, followed by the Mann-Whitney Utest for two particular groups. |
T cells were gated according to their size and granularity, and the ability of Th2 (IL-4 and IL10) and Th1 (IFN-γ and TNF-α) production was analyzed by flow cytometry. The results were expressed as IL-4/IFN-γ or IL10/TNF-α ratio, as it can be observed in Fig 32 for spleen (A) and decidual (B) lymphocytes.
↓117 |
Interestingly, a diminution in the IL-4/IFN-γ ratio could be observed in both, spleen and decidual lymphocytes from abortion mice when compared to normal pregnant mice, nicely supporting the Th1/Th2 paradigm (Fig. 32 A and B). Results are shown in Fig.32A and B, for spleen and decidual lymphocytes, respectively. Very interestingly, only mice treated with 1x105 PFU but not with 1x108 PFU AdHO-1/GFP showed a significant augmentation in the IL-4/IFN-γ cytokines ratio in both, spleen and decidual lymphocytes. Moreover, the IL4/IFNγ levels in this group were comparable to those obtained in the normal pregnancy group. High doses of HO-1 were not able to up regulate the IL-4/IFN-γ ratio, suggesting again that too much HO-1 is not pregnancy compatible.
Fig. 33 : IL-10/TNF-α ratio in spleen and decidual lymphocytes as analyzed by flow cytometry | ||
Th1/Th2 ratios in spleen (A) or decidual lymphocytes (B) measured by the production of TNF-α and IL-10 by flow cytometry after PMA/Ionomycin incubation in the presence of monensin. Data are represented by medians ± 75% quartiles. N.P.= Normal Pregnancy; S.A.= Spontaneous abortion; Circles and/or asterisks above box plots indicate outliers. No significant differences were obtained between groups as analyzed by the Kruskall-Wallis non-parametric test between all groups. |
As depicted in Fig. 33A, a diminution in the IL-10/TNF-α ratio in spleen cells was found in abortion-prone animals compared to normal pregnant animals. HO-1 treatment up-regulated IL-10/TNF-α ratio when compared to abortion-prone mice. Nevertheless, this difference was not significant. Surprisingly, the levels of IL-10/TNF-α in decidual cells were augmented in abortion-prone mice when compared to normal-pregnant mice. Mice receiving AdHO-1/GFP as well as mice receiving AdEGFP showed similar IL-10/TNF-α levels as those obtained for the normal pregnant combination. Kruskall-Wallis analysis between all groups show no significant differences between all groups analysed (Fig.33B).
↓118 |
Medians of the cytokine levels for each measured cytokine are shown as a Table in the Appendix (Tables 15 and 16).
Since a Th2 shift in the cytokine production was observed in the AdHO-1/GFP treated group, it was interesting to analyze whether this treatment also provokes an augmentation in the number of lymphocytes reaching the feto-maternal interface. The marker CD3, expressed in all T cells, was analyzed at mRNA and protein levels.
Levels of CD3 mRNA were measured by real time RT-PCR in placentas (tissue of fetal origin) as well as in decidua (maternal tissue). As can be observed in Table 9, the CD3 mRNA levels were very low, confirming low amounts of the immune cells at the placenta. Besides, no major differences were obtained regarding CD3 expression in placental tissue between the groups. This suggests that the lymphocytes present in the placenta are prone to produce more Th2 and less Th1 cytokines after HO-1 treatment.
↓119 |
In decidual tissue (of maternal origin) CD3 expression was augmented in animals receiving AdHO-1/GFP. As can be seen in Fig. 34, the expression of CD3 mRNA in HO-1 treated animals was significantly augmented when compared to normal pregnant animals. It is tempting to speculate that this augmentation is due to a migration of lymphocytes into the feto-maternal interface. As we could observe higher Th2 than Th1 levels it is then postulated that more Th2 lymphocytes generated in the periphery (Fig. 33A and 34A) reach the feto-maternal interface (Fig. 33B).
N.P (n=3) |
S.A. (n=5) |
S.A. + AdHO 1/GFP (n=5) |
S.A. + AdEGFP (n=4) |
|
Mean |
0.0003161 |
0.0002290 |
0.0003148 |
0.001146 |
SD |
0.0005072 |
0.0002357 |
0.0003607 |
0.001094 |
Median |
0.0000277 |
0.0001757 |
0.0002286 |
0.0007203 |
Fig. 34 : The systemic application of AdHO-1/GFP led to an augmentation of CD3 mRNA in the decidua as analyzed by real time PCR normalized to β-actin. | ||
Decidual expression of CD3 mRNA as analyzed by real time PCR normalized to b-actin. Data are represented by medians ± 75% quartiles. N.P.= Normal Pregnancy; S.A.= Spontaneous abortion. Significant differences between groups are indicated as asterisks above lines. *p≤ 0.05, as analyzed by the Kruskall-Wallis non-parametric test between all groups, followed by the Mann-Whitney Utest for two particular groups. |
↓120 |
CD3 expression at protein level was analyzed in placental tissue by immunohistochemistry. Unfortunately, CD3+ cells were rare to found. Since in many samples no cells could be found, it was not feasible to quantify the number of CD3+ cells. Nevertheless, samples of animals receiving AdHO-1/GFP did not show massive infiltration of CD3+ lymphocytes, supporting the RT-PCR data with no differences in CD3 expression in placental tissue between the groups.
Many authors have proposed a link between HO-1 and regulatory T cells (Treg) (Pae et al., 2003; Choi et al., 2005), although this is still controversial. In order to clarify whether in our animal model HO1 was able to up-regulate the number of Treg at the feto-maternal interface, the markers TGF-β and FoxP3 were analyzed by real time RT-PCR in placental tissue. FoxP3 is considered to be a specific marker that indicates the presence of Treg as it is constitutively expressed in Treg but not in effector T cells. TGF-β is normally secreted by certain types of Treg and augmented levels of any of these markers or of both markers together may indicate the presence of Treg. As shown in Table 10 no augmentation of TGF-β mRNA was observed when injecting AdHO-1/GFP, but there was an augmentation in the levels of TGF-β in the group receiving AdEGFP when compared with the group receiving AdHO-1/GFP. Although these results were not expected as it would mean that the sole injection of an adenovirus is protective, one have to keep in mind that TGFβ is mainly regulated post-transcriptionally, and mRNA levels may not reflect protein levels, hence may not count for the in vivo situation. Unfortunately, it was impossible to measure protein levels by IHC or Western Blot.
N.P (n=5) |
S.A. (n=6) |
S.A. + AdHO 1/GFP (n=5) |
S.A. + AdE FP (n=5) |
||
TGF - |
Mean |
0.1278 |
0.1041 |
0.05937 |
0.1624 |
SD |
0.1106 |
0.06807 |
0.05241 |
0.02397 |
|
Median |
0.1344 |
0.08931 |
0.04123 |
0.1560* |
↓121 |
Levels of FoxP3 mRNA were found to be comparable between all groups analyzed (Table 11). No relationship could be found between the application of the adenoviral vector containing HO-1 and Treg. However, more studies should be carried out before excluding a connection between these two systems in this particular model.
Table 11 : No differences in the mRNa levels of FoxP3 in decidual tissue between the groups
N.P (n=3) § |
S.A. (n=2) § |
S.A. + AdHO-1/GFP (n=3) § |
S.A. + AdEGFP (n=3) § |
||
FoxP3 |
Mean |
4.00 x 10-5 |
1.24 x 10-4 |
1.23 x 10-4 |
9.59 x 10 -5 |
SD |
2.88 x 10-5 |
1.91 x 10-5 |
6.35 x 10-5 |
4.29 x 10 -5 |
|
Median |
2.43 x 10 -5 |
1.24 x 10 -4 |
9.06 x 10 -5 |
9.61 x 10 -5 |
↓122 |
HO-1 and its products are known to exert anti-apoptotic effects in many different models. In order to see if HO-1 also has an anti-apoptotic effect at the feto-maternal interface, apoptosis was measured by two different methods, namely caspase-3 activity in tissue homogenates and TUNEL staining in tissue slides. The measurement of Caspase-3 activity of whole placental homogenates showed a slight augmentation in the caspase-3 activity, thus in apoptosis, in placentas from abortion-prone animals when compared to placentas from normal pregnant mice. The caspase-3 activity could be diminished after injection of AdHO1/GFP, suggesting an antiapoptotic effect of the HO-1 therapy that may involve a caspase3 dependent pathway. The levels of caspase-3 activity were also diminished when compared to the levels measured in mice receiving an AdEGFP control vector. Besides, the placental homogenates from animals receiving the AdEGFP control vector showed even a slight augmentation when compared to the abortion-prone group. Results for caspase-3 activity are shown in Fig. 35. The measurement of caspase-3 activity gives an idea of the rate of apoptosis in the tissue that occurred in a caspase-3 dependent way. However, other mechanisms of apoptosis may occur, and this was the reason why apoptosis was also measured by TUNEL technology.
Fig. 35 : Diminished caspase-3 activity in placental tissue of animals receiving the AdHO-1/GFP vector | ||
Data are represented by medians ± 75% quartiles. N.P.= Normal Pregnancy; S.A.= Spontaneous abortion; Circles and/or asterisks above box plots indicate outliers. No significant differences were found as analyzed by the Kruskall-Wallis non-parametric test between all groups. |
By using the TUNEL technology, the number of apoptotic cells/mm2 in the placenta could be determined by light microscopy based on the morphology of the stained nuclei of the cells, and results are depicted in Fig.36. Placentas from abortion-prone mice presented significant higher number of apoptotic cells when compared to placentas from normal pregnant animals.
↓123 |
The application of AdHO-1/GFP led to a significant diminution of the number of apoptotic cells when comparing to placentas from abortion-prone mice receiving PBS or AdEGFP, reinforcing the hypothesis of an antiapoptotic effect of the HO-1 therapy. The fact that a big diminution in the number of apoptotic cells was found in the group receiving AdHO-1/GFP, whereas the difference obtained in the caspase-3 activity was only slight and not significant may suggest that the AdHO-1/GFP could have acted also in a caspase-3 independent pathway. A representative picture of the TUNEL staining in placental tissue is shown in Fig. 37A. Fig. 37B shows the negative control of the staining.
Fig. 36 : The up-regulation of HO-1 led to a significant diminution in the number of apoptotic cells in the placenta as measured by TUNEL | ||
Number of apoptotic cells/mm2. For each section, positive cells were counted in at least 20 fields using a 200x magnification (20x objective and 10x ocular). Data are represented by medians ± 75% quartiles. N.P.= Normal Pregnancy; S.A.= Spontaneous abortion. Significant differences between groups are indicated as asterisks above lines. *p≤ 0.05 as analyzed by the Kruskall-Wallis non-parametric test between all groups, followed by the Mann-Whitney Utest for two particular groups. |
Fig. 37 : Representative example of the TUNEL staining in placental tissue | ||
A) sample containing apoptotic cells; B) negative control; pictures taken at 20X objective and 10X ocular amplification. |
↓124 |
Since mice receiving AdHO-1/GFP presented diminished apoptosis, it was interesting to go into a deeper analysis and try to elucidate by which mechanism this prevention of apoptosis occurred. For that, mRNA levels of Bag-1 and further anti-apoptotic molecules such as Bcl-xl and Bcl-2 in decidual and placental samples were analysed by real time RT-PCR. As depicted in Fig. 38, Bag-1 mRNA levels were slightly diminished in decidual tissues from the abortion-prone group when compared to the normal pregnancy group. Interestingly, AdHO1/GFP treatment could revert the Bag-1 down-regulation (p<0.05), confirming an anti-apoptotic effect of the HO-1 therapy. Levels of mRNA for Bcl-xl and Bcl-2 molecules were not affected by the AdHO-1/GFP treatment and no differences between the groups were found (data not shown).
Fig. 38 : Up-regulation of HO-1 led to a significant augmentation of the mRNA levels of the anti-apoptotic molecule Bag-1 as measured by real time RT-PCR. | ||
Data are represented by medians ± 75% quartiles. N.P.= Normal Pregnancy; S.A.= Spontaneous abortion. Significant differences between groups are indicated as asterisks above lines. *p≤ 0.05 as analyzed by the Kruskall-Wallis non-parametric test between all groups, followed by the Mann-Whitney Utest for two particular groups. |
The HO-1 molecule is known to be related to angiogenesis. In this regard, it has been shown by others (reviewed in Dulak et al., 2004) that an up-regulation of HO-1 leads to an augmentation of the angiogenic factor VEGF and viceversa (Bussolati et al., 2004). In order to analyze whether the beneficial effect obtained by the up-regulation of HO-1 was related to angiogenesis, the expression of VEGF was measured at a single cell level in placental tissue by immunhistochemistry. The results obtained with this method are shown in Fig. 39, where it can be seen that animals receiving low doses of AdHO-1/GFP showed a slight augmentation in the number of cells expressing VEGF, when compared to all other groups, suggesting that HO-1 may have also angiogenic effects in this particular model.
↓125 |
Fig. 39 : Slight augmentation in the number of VEGF+ cells in placentas of animales treated with AdHO-1/GFP as analyzed at single cell level by IHC. | ||
Number of VEGF+ cells/mm2. For each section, positive cells were counted in at least 20 fields using a 200x magnification (20x objective and 10x ocular). Data are represented by medians ± 75% quartiles. N.P.= Normal Pregnancy; S.A.= Spontaneous abortion. No significant differences were found between groups as analyzed by the Kruskall-Wallis non-parametric test between all groups. |
An example of the immunohistochemical staining for VEGF is shown in Fig. 40, where a sample containing VEGF+ cells (A) is depicted next to the negative control of the staining (B).
Fig. 40 : Representative example of the VEGF expression in placental tissue as measured by immunohistochemistry in paraffin-embedded tissue. | ||
A) sample containing VEGF positive cells; B) negative control of the staining. Pictures taken at 20X objective and 10X ocular amplification. |
↓126 |
The Rcho-1 cell line is a trophoblastic stem cell line that differentiates into giant cell line upon different culture conditions. This cell line is useful to study whether HO-1 is necessary for these cells to survive and differentiate into giant cells in the context of this thesis. For this purpose, the effect of CoPPIX and ZnPPIX was analyzed in the culture of undifferentiated Rcho-1 cells. ZnPPIX is known to down-regulate HO-1 in vivo as well as in v i tro. CoPPIX is known to up-regulate HO-1 in vivo, whereas in vitro is controversial if this up-regulation takes place (reviewed in Ryter et al., 2006). Since no reports of the effect of CoPPIX or ZnPPIX on trophoblast cells were found in the literature, different conditions were tested in order to use the optimal concentration of ZnPPIX and CoPPIX for these cells. The results on HO-1 expression using the different concentrations and incubation time are shown in Fig. 41.
As it can be observed in the western blot example (Fig 41), Rcho1 cells express very high basal levels of HO-1. A down-regulation of HO-1 can be observed already with 50 mM ZnPPIX at 24h. Unlike ZnPPIX, CoPPIX had no effect on HO-1 levels in these cells as HO-1 expression remained at basal levels. Cells treated with CoPPIX were used then as a control for the assay, in order to analyze whether the effects observed with ZnPPIX are due to the down-regulation of HO-1 or to a putative toxic effect of the porphyrin.
Fig. 41 : Example of the Western Blot analysis of Rcho-1 cells treated with CoPPIX or ZnPPIX | ||
A to D: 12 h after ZnPPIX treatment (0, 25, 50 and 100 μM respectively). E to H: 24 h after ZnPPIX treatment (0, 25, 50 and 100 μM respectively) I to L: 12 h after CoPPIX treatment (0, 25, 50 and 100 μM respectively). M to P: 24 h after CoPPIX treatment (0, 25, 50 and 100 μM respectively). It is unknown the reason why the antibody against HO-1 gave two bands in these samples. Nevertheless, the upper band (corresponding to 32 kDa) was taken into consideration. |
↓127 |
In a first assay, cell viability of the trophoblast stem cell was analyzed after 48 h of treatment with medium, ZnPPIX or CoPPIX by tryphan blue exclusion.
Fig. 42 : HO-1 down-regulation leads to diminished viability of the trophoblast stem cells as measured by trypan blue exclusion. | ||
Data are represented by median & range. Significant differences between groups are indicated as asterisks above lines. *p≤ 0.05 as analyzed by the Kruskall-Wallis non-parametric test between all groups, followed by the Dunn´s Multiple Comparison Test for two particular groups. |
As it can be observed in Fig. 42, a significant diminution in the viability of the cells was obtained 48 h after cells were treated with 50 μM or 100 μM ZnPPIX. This effect cannot be due to the toxic effect of the porphyrin, since cells treated with CoPPIX did not show a significant diminution in the cell viability. Since ZnPPIX is known to block the expression of HO-1, it can be assumed that HO-1 is necessary for the survival of trophoblast precursor cells.
↓128 |
The effect of CoPPIX and ZnPPIX in the differentiation of Rcho-1 cells into giant cells was analysed applying CoPPIX or ZNPPIX during the differentiation process. Differentiated cells show a characteristic morphology, and they can be well recognized from undifferentiated cells. Differentiated cells (giant cells) are considerably bigger that undifferentiated cells, and they are normally multinucleated or bear a big nucleus. Differentiation of these cells is normally obtained after 1 week of culture in medium containing 10% horse serum.
Fig. 43 : Stem cells treated with ZnPPIX were unable to differentiate into giant trophoblast cells as analyzed by light microscopy. | ||
Pictures taken after 72 h, 5 days and 7 days of culture under differentiation media. Cells treated with normal media as well as cells treated with CoPPIX are able to differentiate into giant cells, whereas cells treated with ZnPPIX are unable to differentiate into a mature state. |
As observed in Fig. 43, the morphology of the cells start changing already at 72 h of culture and the differentiation is complete at 7 days of culture with medium including horse serum. When applying CoPPIX (Fig. 43), no significant changes in the morphology of the cells are observed when compared to cells treated with only differentiation medium. However, when applying ZnPPIX (Fig 43), which blocked HO-1, the cells are unable to differentiate into giant cells, pointing out that HO-1 is indispensable for trophoblast precursor cells to differentiate into a mature phenotype. This effect could be due to the toxic effects of ZnPPIX, but this is not rather the case, since cells receiving CoPPIX, where the same toxic effect is applied, did not show alterations in their differentiation capacity.
↓129 |
With the aim to transduce lymphocytes with HO-1 for testing their in vitro or in vivo ability to suppress maternal T cell responses, we proceed to first generate packaging cell lines expressing HO-1. Since retroviral vectors are suited for transduction of T cells, two different packaging cell lines overexpressing the HO-1 transgene were obtained. These cell lines showed the following viral titers:
GP+E HO-1: 6.106 cfu/ml.
GP+E HO-1iresEGFP: 4.106 cfu/ml.
↓130 |
After limiting dilution, 14 clones of the packaging cell line HO-1 and 9 clones of the HO1iresEGFP packaging cell line were selected. The selection criterion was based on the Western Blot intensity of HO-1 observed for each sample. Representative examples of different HO-1 protein expression patterns are shown in Fig.44. In A) it can be observed the difference in the HO-1 expression between the packaging cell line without the HO-1 transgene, and both packaging cell lines obtained. These packaging cell lines were further cloned, and examples of the different HO-1 expression in the clones are illustrated in B).
After selecting the highest producer clones, the viral titers were determined as already described and the results are summarized in the following table (Table 12).
Since there is not always a correlation between the transgene expression and the viral titer, it was necessary to select those clones that show not only a high expression of the transgene but also have a high viral titer.
↓131 |
Fig. 44 : Cells from the packaging cell line were efficiently transfected to produce high amounts of HO-1 as observed by Western Blot analysis. | ||
A) HO-1 expression in the packaging cell lines prior and after transfection. B) HO-1 expression in the different clones from the HO-1 and the HO1iresEGFP packaging cell lines. |
Table 12 : Viral titers of the different clones of the HO-1 overexpressing packaging cell lines
A |
Clone |
Viral titer (cfu/ml) |
B |
Clone |
Viral titer (cfu/ml) |
1D10 |
3.107 |
1H7 |
4.105 |
||
1E12 |
3.106 |
2A10 |
2.107 |
||
1H12 |
3.105 |
2D1 |
6.106 |
||
2A11 |
2.107 |
2D7 |
1.108 |
||
2C11 |
3.105 |
2G2 |
1.106 |
||
2C3 |
7.106 |
3B3 |
1.108 |
||
2D12 |
1.106 |
3E1 |
1.107 |
||
2E8 |
9.106 |
3E10 |
8.107 |
||
3A2 |
1.107 |
3G8 |
8.107 |
||
3B10 |
1.106 | ||||
3B6 |
2.107 | ||||
3B7 |
4.107 | ||||
3D8 |
1.107 | ||||
3H1 |
1.107 |
After careful analysis of the results from Western Blots and viral titers determination, the clone 3B3 was selected as an appropriate one to be used in the transduction of the Tlymphocytes with the HO-1iresEGFP, since it shows a high expression of the protein, and an excellent viral titer. In addition, it harbours the advantage of the EGFP as a reporter gene to control the transduction efficiency. From the HO-1 packaging cell line, the clone 3D8 was selected.
↓132 |
Many conditions for mouse T cell transduction with the retroviral particles were tested. The highest transduction efficiency (7.91%) was obtained when CD4+ T cells were transduced with 500 ml of virus concentrate, in a final volume of 1.5 ml in a 12-well plate. The virus concentrate was obtained by means of centrifugation in a Vivaspin column. An example of the efficiency of virus transduction, as measured by flow cytometry, is shown in Fig. 45. All other conditions tested (explained in materials and methods) were unsuccessful or led to even lower transduction efficiency (around 2%).
Fig. 45 : The highest transduction efficiency was obtained by using a virus concentrate as analyzed by flow cytometry | ||
FL-1H measures the cells that are positive for HO-1iresEGFP (transduced cells). In this example, 7.49% of the total cells were efficiently transduced. |
These transduced cells could theoretically be selected by G-418. Although the selection seemed to work, the number of cells remaining alive was to low to be maintained in culture, and the cells died after few days. This was repeatedly observed in several approaches. Since the G-418 selection was not feasible, the obtainment of a sufficient amount of transduced cells would have required the sacrifice of more than one animal in order to obtain a sufficient initial amount of cells. Additionally, an excessive production of virus supernatant would have been necessary in order to obtain enough amount of virus concentrate. All these inconvenient led us to conclude that the transduction of mouse T cells with a retroviral vector was not a feasible for our proposed experimental settings, at least using this retroviral vector. We decided then to try alternative methods, like protein transfection, in order to obtain T cells over-expressing HO-1.
↓133 |
Protein transfection is still not described in the literature for mouse primary T cells as it is for other cells like murine and human fibroblasts. In order to establish the better condition for transfection, a positive control using β-galactosidase was used (suggested by the manufacturer of the protein transfection reagent). As mentioned in Materials and Methods, different cell number, protein content and incubation times were analyzed. Fig.46 depicts the different results obtained.
Fig. 46 : Up to 80% of the cells were efficiently transfected with the b-gal protein using the protein transfection technology as analyzed by flow cytometry. | ||
Data are shown by histogram. FL1-Height measures the cells that show bgalactosidase activity. Fig. A shows the different conditions tested for 1x105cells/well, whereas Fig. B shows the conditions tested for 3x105cells/well. Up to 80.95% of the cells (purple line, diagram B) were efficiently transfected. |
As mentioned in Materials and Methods, the condition mentioned in 2.2.2.10 was selected as optimal for protein transfection. Transfection with HO-1 protein could unfortunately not be measured by flow cytometry, and the optimal protein amount was determined by western blot. As can be seen in Fig. 47, transfected cells showed higher expression of HO-1 protein when compared to untransfected cells, and the amount of 2 μg of protein was selected as optimal for the further experiments.
↓134 |
Fig. 47 : HO-1 protein was efficiently transfected into CD4+ cells as determined by Western Blot analysis | ||
1) Untransfected CD4+ T cells 2) CD4+ T cells transfected with 1 μg of HO-1 protein 3) CD4+ T cells transfected with 2 μg of HO-1 protein |
The idea to use cells over-expressing a given molecule as therapy for transplantation (Hammer et al., 2000) gave us the idea of using CD4+ T cells in an abortion-prone setting. For that reason, the effect of CD4+ T cells over-expressing HO-1 on cells from an abortionprone female was analyzed. Since HO-1 has been described as having anti-proliferative effects, the proliferation of these cells (called from now on “control cells”) was analyzed by flow cytometry using the CFSE membrane dye. For doing this, cells from a CBA/J female previously mated with a DBA/2J male were isolated from lymph nodes, spleen, and thymus and stained with CFSE. From the thymus, CD4+CD25+ and CD4+CD25- were isolated and separately tested. As previously mentioned, the female was pregnant on day 5 from a DBA/2J male, meaning that cells from this female have already encountered male antigens and are supposed to generate an antigen specific reaction. In order to stimulate the culture and to simulate the continuous encounter with male antigens, DBA/2J cells previously treated with mitomycin C were used as stimulators.
As depicted in Fig 48, and contrary to our expectations, no differences in the proliferation of control cells could be found when CD4+ T cells over-expressing HO-1 were put in contact with cells from lymph nodes, spleen or thymus of a CBA/J pregnant female. A significant difference was observed in the proliferation of spleen cells, but this diminution was also observed when TMOCK cells were used. The presence of these cells (both THO-1+ or TMOCK cells) did also not influence the expression of different markers as CD69 or CD25 (data not shown).
↓135 |
Fig. 48 : No effect of CD4+ T cells overexpressing HO-1 on the proliferation of control T cells as measured in a MLC. | ||
Data are represented by mean ± SD. Significant differences between groups are indicated as asterisks above lines. *p≤ 0.05 and ***p≤0.001 as analyzed by the Kruskall-Wallis non-parametric test between all groups, followed by the Dunn´s Multiple Comparison Test for two particular groups. |
The effect of these cells was also tested in vivo. For this purpose, DBA/2J-mated CBA/J females were intravenously injected with CD4+ T cells over-expressing HO-1 or with TMOCK cells in day 5 of pregnancy. Also in vivo, no effect of these cells could be observed, since the abortion rates remained similar between the groups.
From this part of the study, it can be concluded that in this particular model, a cell therapy with lymphocytes over-expressing HO-1 was not adequate for the purpose of reducing the abortion rate. However, one has to keep in mind that even when HO-1 was expressed in the T cells, the transfected protein may not have been active. Unfortunately we were not able to measure HO-1 activity in these cells to confirm our hypothesis. Other possibility is that T cells are not the right target for over-expressing HO-1. Dendritic cells would be a rather better choice since it is known that HO-1 is expressed in them and that an up-regulation of HO-1 in these cells is capable of maintaining them in an immature state (Chauveau et al., 2005).
↓136 |
The use of transgenic animals lacking a given molecule represents a useful tool to analyze the effect that this produces in the organism. In pregnancy, it was essential for the purpose of this thesis to analyze the effect of the lack of HO-1 on different stages of pregnancy.
As already mentioned in the literature, breeding between homozygous Hmox1 deficient mice does not yield progeny (Poss and Tonegawa, 1997), and mice genetically deficient in Heme Oxygenase-1 (Hmox1 deficient mice, or Hmox1 -/- mice) are partially embryonic lethal (Yet et al., 1999), suggesting a key role of HO-1 in the onset of pregnancy. Additionally, when heterozygous mice (Hmox1 +/- ) are mated for maintaining the colony, very low percentage of knockout animals (Hmox1 -/- ) are usually born, usually between 5 and 10%, according to the literature and to personal communication of Prof. Soares. This unexpected variation from the 25% expected Mendelian rate may be due to different reasons:
↓137 |
For the purpose of analyzing if the Hmox1 -/- fetuses are rejected and to know if there is a difference in the number of implantations or in the abortion rates regarding the different expression of Hmox1, the outcome of different combinations of Hmox1 -/- , Hmox1 +/- and Hmox1 +/+ mice was analyzed. This was tested in a syngeneic as well as in an allogeneic context, whereas the female was from a different genotype, namely C57/BL6. Unfortunately, for this combination it was only possible to use wild type females (Hmox1 +/+ ) as no knockout in C57/BL6 background were available at the moment of performing the experiments. The results obtained are shown in Table 13.
It has been described previously (Poss and Tonegawa, 1997; Yet et al., 1999) that no progeny can be achieved when mating Hmox1 -/- mice. However, these studies did not analyze whether this phenomenon is due to no implantation or due to complete rejection of the implanted embryos. Due to the difficulties in the obtainment of Hmox1 -/- animals, we did not perform groups using Hmox1 -/- females.
Mating genotype (female x male) |
Number of co n cepti (number of an i mals) |
Abortion rate (m e dian) only in those an i mals with visible i m plantations on day 12-14 |
Mean of Impla n tation sites (plugged f e males) |
Number of an i mals with visible c o cepti (plugged female) |
Syngeneic (BALB/c x BALB/c) |
||||
Hmox1 +/- x Hmox1 -/- |
44 (11) |
38.7 % |
4 (11) |
4 (11), 36.4 % |
Hmox1 +/- x Hmox1 +/- |
82 (7) |
22.2 % |
11.7 (7) |
7 (7), 100 % |
Hmox1 +/- x Hmox1 +/+ |
46 (6) |
18.4 % |
7.7 (6) |
4 (6), 66.7 % |
Hmox1 +/+ x Hmox1 -/- |
32 (5) |
9.4 % |
6.4 (5) |
3 (5), 60 % |
Hmox1 +/+ x Hmox1 +/- |
42 (5) |
11.1 % |
8.4 (5) |
5 (5), 100 % |
Hmox1 +/+ x Hmox1 +/+ |
43 (8) |
0 % |
5.45 (8) |
5 (8), 62.5% |
Allogeneic (C57/BL6 x BALB/c) |
||||
Hmox1 +/+ x Hmox1 -/- |
58 (8) |
8,6 % |
7.25 (8) |
7 (8), 87.5% |
Hmox1 +/+ x Hmox1 +/+ |
43 (8) |
0 % |
5.40 (8) |
5 (8), 62.5% |
↓138 |
When females partially deficient in Hmox-1 (Hmox1 +/- ) are mated with Hmox1 -/- males, 4 out of 11 females that showed insemination plug were indeed pregnant, and showed an abortion rate of 38.7%. The fact that only 4 out of 11 females were pregnant suggests that Hmox1 -/- male may present fertilization problems. Similarly, when Hmox1 +/+ females were mated with Hmox1 -/- , 3 out of 5 females mated were pregnant, suggesting here that Hmox1 -/- males may have fertilization problems. However, Hmox1 +/+ males also showed fertilization problems when fertilizing Hmox1 +/- or Hmox1 +/+ females, since not all females showing insemination plug were pregnant. This suggests that these fertilization problems may not be exclusively associated with the lack of Hmox1 in the male.
When analyzing the abortion rates from the heterozygous females, an interesting result is observed, namely that the abortion rate is higher when less HO-1 is present in the mating combinations. When these females were mated with Hmox1 +/+ males, they presented an abortion rate of 18.4 %, with Hmox1 +/- males the abortion rate was of 22.4 %, whereas with Hmox1 -/- males the abortion rate was of 38.7 %. Although these differences are not significant, a tendency towards an augmentation on the abortion rate with the diminution of HO-1 in the system can be appreciated.
When Hmox1 +/+ females were mated with all three types of males, they showed very similar abortion rates in all three types of combinations, and the abortion rate was normal for syngeneic combinations (between 0 and 10% of abortion). This data strongly suggests that HO-1 is very important in the maternal site for a successful pregnancy, whereas its absence in the paternal site may not be as important as like in the maternal site.
↓139 |
When doing an allogeneic combination using Hmox1 +/+ C57/BL6 females mated either with Hmox1 -/- or Hmox1 +/+ BALB/c males, most of the females presenting insemination plug were pregnant. When the male was deficient in Hmox1, an 8.6% of abortion was observed, against 0% of abortion obtained with the Hmox1 +/+ males, confirming again that the absence of HO1 on the paternal side may not be as important as it is in the maternal site.
Analysis of the genotype of the fetuses and resorptions were performed in the groups of heterozygous females mated either with knockout or heterozygous males. Some of the fetuses from the other groups were randomly analysed and confirmed to be heterozygous, as expected. An example of the genotyping analysis performed by PCR is shown in Fig. 49.
The results of the analysis are shown in Fig. 50 for the Hmox1 +/- x Hmox1 -/- combination, and in Fig. 51 for the Hmox1 +/- x Hmox1 +/- mating combination.
↓140 |
Surprisingly, the mating of Hmox1 +/- females with Hmox1 -/- males lead only to heterozygous offspring. These results may suggest that Hmox1 -/- blastocysts were unable to implant, since fetuses as well as resorptions were heterozygous. The fact that in this group only 4 out of 11 females showing insemination plug were pregnant suggests that Hmox1 -/- sperm may present fertilization problems.
Fig. 49 : Example of amplification bands obtained in the genotyping of fetuses as analyzed in an Ethidium Bromide-stained gel after electrophoresis. | ||
Amplification products of the PCR reactions for the E3/EI3 fragment (present in the wild type gene) and the Neo-1/E4 fragment (present in the modified gene) as analyzed in a 1% agarose gel containing ethidium bromide under ultraviolet transillumination. |
↓141 |
Fig. 50 : all fetuses and resorptions from the Hmox1 +/- x Hmox1 -/- combinations were heterozygous. | ||
In the Hmox1+/- x Hmox1-/- combination, fetuses and resorptions from 2 pregnant females were genotyped. |
On the other side, the mating of Hmox1 +/- females with Hmox1 +/- males led to the typical percentage of knockouts obtained from these matings, and this percentage is lower as the expected Mendelian rate of 25%. According to personal communication from Prof. Soares and to personal observation from the maintenance of the colony, this percentage is always between 3 and 10%. This reinforces the idea of inability of knockout embryos to implant, or inability of females lacking partially HO-1 of providing essential conditions for fetuses to implant.
Fig. 51 : All resorptions from the Hmox1 +/- x Hmox1 +/- combination were heterozygous | ||
In the Hmox1+/- x Hmox1+/- combination, fetuses and resorptions from 3 females were genotyped. |
↓142 |
Interestingly, the resorptions were again heterozygous (Fig 51) suggesting that the lower percentage of Hmox1 -/- fetuses obtained from the Hmox1 +/ x Hmox1 +/- mating is due to the lack of implantation of the Hmox1 -/- blastocysts rather than to rejection of these fetuses.
In the former experiment it could be observed that HO-1 is very important especially in the mother, for successful pregnancy to occur. This does still not explain why the mating of heterozygous mice does not give the expected Mendelian rate, since the fetuses that were rejected were heterozygous and not homozygous as expected, and does also not explain why homozygous mice do not give viable progeny. In order to assess if this can be due to sterility problems of the homozygous mice or to a defect or delayed ovulation, Hmox1 -/- and Hmox1 +/+ females were first stimulated with hormones in order to see if they are able to produce viable oocytes. As described in 3.2.4.2.4, females were sacrificed 13-14 h after hCG injection. Oocytes from 8 females were placed in two or three fertilization plates. This procedure was carried out for wild type females, here referred as Hmox1 +/+ (n = 8) and for knockout (Hmox1 /- ) females (n = 8). Interestingly, the number of oocytes obtained differed greatly between the Hmox1 +/+ and Hmox1 -/- females, although they were treated under the same conditions and injected with the same lot of hormones. Age differences between the animals can not play a role since the animals used were agemated. The total number of oocytes obtained from the wild type females (n=8) was 208, while the number obtained from the Hmox1 -/- females (n=8) was 137. This difference is depicted in Fig. 52A. A similar graphic was obtained when calculating the number of oocytes per female (calculated taking into account the average of oocytes/female in each plate), as can be seen in Fig. 52B.
Fig. 52 : Hmox1 -/- females produce less oocytes than Hmox1 +/+ after hormonal stimulation. | ||
A) Total number of oocytes (n=8 wild type females; n=8 Hmox1-/- females); B) Number of oocytes per female. |
↓143 |
Since the procedure of oocyte collection from the oviduct has to be performed very quickly and in a very short period of time (less than 5 minutes per female), and considering that newly ovulated oocytes are surrounded by cumulus cells, it was not possible to count the number of oocytes obtained per female as they were retreated from the oviducts. Once the oocytes were washed and liberated from cumulus cells, it was possible to determine the exact number of oocytes per fertilization plate. As the number of oocytes obtained from the Hmox1 -/- females was less than the ones from the Hmox1 +/+ females, their fertilization was performed only in two fertilization plates instead of three. For this reason, it was not possible to determine of the difference in the number of oocytes was statistically significant.
In order to analyze the origin of the difference observed in the oocyte number between Hmox1 +/+ and Hmox1 -/- females, the ovaries of the females used as oocytes donors Hmox1 +/+ and Hmox1 -/- ) where analyzed histologically. Differences in the ovaries due to different sexual maturity between Hmox1 -/- and Hmox1 +/+ can be discarded since, as mentioned before, the animals used for this study were age-mated.
Histological analysis of the ovaries of females from both groups showed that, even though ovaries from both types of females showed similar morphology, they presented differences regarding their follicle development. In order to quantify these observations, the number of follicles in different stages of maturation was analyzed under the microscope in all samples obtained (in both ovaries from each animal, counting at least three slides per ovary). As depicted in Fig. 53, the total number of oocytes per ovary is similar between both types of females.
↓144 |
Fig. 53 : Ovaries from Hmox1 +/+ and Hmox1 -/- females present the same total number of follicles as analyzed under light microscope. | ||
Data are represented by medians ± 75% quartiles. No significant differences were found between the groups Mann-Whitney U-test between both groups. |
However, when analyzing the number of follicles in the different maturation stages (Fig 55), some differences were found. Fig 54 is showing an example of follicles in different maturation stages. As it can be observed in Fig. 55, Hmox1 -/- ovaries show similar number of secondary follicles as Hmox1 +/+ ovaries. The number of primordial, primary and mature follicles differs slightly between both groups, but these differences were not significant. However, when analyzing the number of corpus luteum between both groups, Hmox1 -/- ovaries showed a significant diminution when compared to Hmox1 +/+ ovaries. It is important to mention that the total number of follicle per ovaries does not differ between the groups (Fig. 54), pointing out that the difference in the number of corpus luteum account are compensated by a slight augmentation in the number of primordial and secondary follicles in Hmox1 -/- ovaries when compared to Hmox1 +/+ ovaries.
Fig. 54 : Representative pictures of HE staining showing the different stages of follicle maturation as seen under light microscope. | ||
Pictures were taken with a 20X magnification. The differences in the size of the follicles showed in these pictures are in accordance with the difference in sizes observed in the samples. A) Primordial follicle B) Primary follicle C) Secondary follicle D) Mature follicle E) Corpus luteum |
↓145 |
The significant diminution in the number of corpus luteum as shown in Fig. 55 is also in accordance with the fact that Hmox1 -/- females produced less oocytes than Hmox1 +/+ females (Fig. 52), because the corpus luteum is the structure remaining after the release of the oocyte.
Taken all these data together, it is tempting to speculate that Hmox1 -/- females react differently than Hmox1 +/+ females to hormones, since the follicle development of the ovaries from both differ greatly, even though both groups showed similar number of oocytes in the ovaries. This difference in the response to hormonal treatment may also account for the difference found in the number of oocytes secreted by Hmox1 -/- and Hmox1 +/+ females. Since the difference in the number of corpus luteum between both groups is significant, it is tempting to speculate that the difference in the number of oocytes showed in Fig. 52 may also be significant.
Fig. 55 : Different follicle development in ovaries from Hmox1 +/+ and Hmox1 -/- females after hormone stimulation as analyzed by light microscopy. | ||
Data are represented by median & interquartile range. Significant differences between groups are indicated as asterisks above lines. *p≤ 0.05 as analyzed by the Mann-Whitney U-test between both groups. For each animal, the mean value of follicles in different stages was calculated analyzing at least 3 slides for each ovary and used as number of follicles per ovary. For each group, 8 animals were analyzed. |
↓146 |
Since Hmox1 -/- do not yield progeny, it was imperative to determine whether this problem could be due to an impaired fertilization. Briefly, Hmox1 -/- oocytes were in vitro fertilized with Hmox1 /- sperm whereas Hmox1 +/+ oocytes were in vitro fertilized with Hmox1 +/+ sperm. The fertilization rate was assessed by counting the number of oocytes in two-cell stage, and the percentage of fertilization was calculated taking into account the initial number of oocytes. As can be observed in Fig. 56, a statistically significant difference in the fertilization rate of Hmox1 +/+ and Hmox1 -/- oocytes was obtained.
Fig. 56: Hmox1 -/- oocytes are poorly fertilized with Hmox1 -/- sperm when compared to Hmox1 +/+ oocytes with Hmox1 +/+ sperm compared to Hmox1 +/+ oocytes by counting the number of oocytes in two cell stage. | ||
Percentages of fertilization rate of Hmox1+/+ and Hmox-/- oocytes. The results are expressed as the mean of the fertilization rate obtained in each fertilization plate. ***p≤ 0.0001 as analyzed by the Fisher´s exact test between both groups comparing the total number of fertilized and unfertilized oocytes. |
A high percentage (61.70%) of the Hmox1 +/+ oocytes were successfully fertilized in vitro, whereas only 19.78% of the Hmox1 -/- oocytes could be fertilized using the exact same conditions (Fig. 56).
↓147 |
The percentage of fertilization obtained for Hmox1 +/+ oocytes was in accordance with the percentages normally obtained using this methodology. The percentage of fertilization of Hmox1 -/- was extremely low, even though the conditions of fertilization were identical in both cases. The statistically significant difference obtained in the fertilization rates between Hmox1 +/+ and Hmox1 -/- oocytes is outstanding, since works claiming the importance of other molecules in fertilization present a difference in percentage of 14.1% between wild type and knockout oocytes (Hefler and Gregg, 2002), whereas in this work the difference was as big as 41.92%. The data obtained here clearly shows that deficiency in Hmox1 has a strong effect in fertilization and opens up an unexpected role of HO-1 in the reproductive field.
Another possible explanation for the lack of progeny of Hmox1 -/- could be related to implantation problems beside the impaired fertilization ability of the oocytes and sperm. In order to test this hypothesis, the embryos in two-cell stage obtained by in vitro fertilization were transferred into recipient females. The recipient females were hormonally treated and mated with vasectomised males in order to assure that they were in a receptive phase. Only females showing plug with the vasectomised males on the day of the transfer were used for this part of the study.
In order to test whether Hmox1 is necessary in the mother, in the fetus or in both for implantation to occur, Hmox1 -/- oocytes were transferred to a Hmox1 +/+ as well as to a Hmox1 -/- recipient female. Additionally, Hmox1 +/+ oocytes were transferred either to a Hmox /- or to a Hmox1 +/+ female. In this context, it was also interesting to analyze the participation of T cells in HO-1-dependent impaired pregnancies. As we know from a personal communication, the mating of Hmox1 +/- SCID animals gives more Hmox1 -/- progeny mice than the mating of Hmox1 +/- immune competent animals (personal communication of Prof. Soares). Therefore, we also tested whether Hmox1 +/+ are able to implant in SCID Hmox1 -/- recipient females.
↓148 |
Results for the transfer of Hmox1 +/+ oocytes are schematized in Fig. 57 and in Fig. 58 for the transfer of Hmox1 -/- oocytes.
Two Hmox-1 -/- females and two SCID/Hmox1 -/- females received 10 two-cell stage Hmox1 +/+ embryos in each uterine horn. One Hmox1 +/+ female and one SCID Hmox1 +/+ female received as well 10 two-cell stage embryo in each uterine horn.
Unfortunately, one of the Hmox1 -/- females died two days after the embryo transfer due to unknown reasons. The other Hmox1 -/- female was sacrificed on day 14 of pregnancy, and showed no sign of embryo implantation. The uterus was neither swallowed nor was it different from the uterus of a virgin female.
↓149 |
One of SCID Hmox1 -/- females showed no visible implantations after transfer of Hmox1 +/+ oocytes, but inflamed uterus. However, the other SCID Hmox1 -/- female presented two implantations after the transfer of Hmox1 +/+ oocytes. It is possible that the first recipient female rejected the embryos after implantation, therefore the inflamed uterus. Unfortunately, the SCID Hmox1 +/+ female, which was supposed to serve as a control, showed no sign of visible implantations.
Hmox1 +/+ oocytes
Fig. 57 : Hmox1 +/+ embryos were not able to implant in Hmox1 -/- uteruses | ||
Implantation success was assessed 14 days after embryo transfer. |
↓150 |
The Hmox1 +/+ female showed 2 implantations after transfer of Hmox1 +/+ oocytes.
These results suggest that in immune competent animals, the lack of Hmox1 can lead to implantation failure, as Hmox1 +/+ oocytes were not able to implant in Hmox1 -/- uterus while these oocytes were able to implant in Hmox1 +/+ uterus. The fact that these embryos were able to implant in SCID animals lacking Hmox1 suggests that in immune competent animals, HO1 is necessary to counteract the inflammation in the uterus due to the presence of immune cells of maternal origin, which may attack paternal antigens.
Hmox1 -/- oocytes
↓151 |
Fig. 58 : Hmox1 -/- embryos are able to implant in Hmox1 +/+ uterus | ||
Implantation success was assessed 14 days after embryo transfer. |
Due to the low number of Hmox1 -/- fertilized oocytes, the transfer could be only performed into two females (one Hmox1 +/+ and one Hmox1 -/- ).
The Hmox1 +/+ female implanted three Hmox1 -/- embryos, showing that Hmox1 -/- can normally implant if the uterus express HO-1. However, the Hmox1 -/- female showed no implantation. The uterus was nor inflamed or different from the uterus of a virgin female. This suggests again need of uterine HO-1 expression for implantation to occur.
↓152 |
All these data together suggest that HO-1 is essential in the female uterus for implantation to occur beside the already analyzed role of HO-1 in prevention of rejection.
© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme. | ||
DiML DTD Version 4.0 | Zertifizierter Dokumentenserver der Humboldt-Universität zu Berlin | HTML generated: 29.09.2014 |