After inoculation with live tachyzoites of the Nc-1 isolate at 210 days of gestation, all dams showed evidence of infection although no abortions occurred and only minimal to mild necrotic and inflammatory lesions were observed in placental and foetal tissues . These findings are similar to previous studies where dams intravenously inoculated with Nc-Liv at the same gestational age gave birth to asymptomatic congenitally infected calves [23, 30]. In comparison, the outcome of infection in early gestation is often fatal with extensive lesions [17, 23, 25, 30] while infection in mid-gestation may result in abortion or the birth of persistently infected calves with mild lesions [18, 23, 25, 26, 28, 35, 36].
This paper describes a novel scoring methodology (similar to that described by Tekin and Hansen ), in order to allow a more objective analysis and comparison of the immune responses in fixed tissue samples. The protocol could be standardised because we used the same IHC protocols in all the samples and the scoring was undertaken by the same observer. Furthermore, this technique could be modified and applied to other studies with the aim of characterising and compare cellular immune responses in fixed tissue samples.
The results of the phenotypic analysis of the immune cell infiltrate in the placentas of the present study have shown that, although differential infiltration was observed between N. caninum-inoculated and negative control animals, no statistically significant differences were found for macrophages. Initially, macrophages were observed in large numbers in the placentas of N. caninum-inoculated dams culled at 14 dpi. Although Neospora was not detected by PCR or IHC in the placentas of these animals  the macrophage infiltration could have been involved in the initiation of an immune response to the parasite challenge in these N. caninum-inoculated animals. Monocytes/macrophages are one of the principal cellular components of innate immunity, acting as antigen presenting cells and consequently influencing the functional direction of the subsequent adaptive immune response [37, 38]. Indeed Rosbottom et al.  demonstrated that endometrial macrophage populations were increased in pregnant cows after experimental infection with N. caninum. At 28 dpi, macrophage infiltration was minimal to mild and an increased number of positive cells was observed at 42 and 56 dpi. This infiltration is associated with the presence of more severe pathological changes in the placenta . Macrophages are not only involved in anti-parasitic activity but also play a key role in the tissue repair process, since they are the principal cell type responsible for wound debridement . This may help to explain their presence after the appearance of tissue damage.
Statistical differences between the N. caninum-inoculated and the negative control animals were observed for the infiltration of CD3+ T cells and, a similar phenomenon was observed for CD4+, CD8+ and γδ-T cells. A positive correlation was observed between these cellular infiltrates and the presence of pathological changes. The mild inflammatory infiltrates observed in the analysed placentomes of N. caninum-inoculated dams primarily expressed CD4+ and γδ-T cell markers, indicating a predominant Th1 response. Mean CD8+ T cell scores were lower than those for γδ T cells and T helper cells. The time pattern for the appearance of inflammatory cells in the placenta was related to the presence of Neospora (PCR and IHC) in the placental samples . Orozco et al.  found scattered and fewer CD4+ and CD8+ T cells in the uterus of pregnant cows naturally infected with N. caninum and found no differences when compared with seronegative cattle. However, it has been established that Neospora is largely controlled by cell-mediated immune mechanisms and specifically, CD4+ T lymphocytes have a significant protective role, demonstrable by the direct lysis of N. caninum-infected cells and production of IFN-γ, which can significantly inhibit multiplication of the parasite [41–45]. Following infection with N. caninum in cattle, CD4+ T lymphocytes are principal components of the Th1 response and produce pro-inflammatory Th1 type cytokines including IFN-γ, TNF-α and IL-12, which have an essential role in protective immunity against the parasite [23, 43, 45–49]. On the other hand, this Th1 response can be detrimental to pregnancy and can compromise foetal survival [19, 38, 50]. In vitro studies have shown that treatment of ruminant cells with IFN-γ significantly inhibited intracellular multiplication of N. caninum.
In humans and ruminants, γδ-T cells are one of the immune cells associated with mucosal surfaces and in the placenta they may be part of the first line of defence against pathogens [20, 31, 51]. In the peripheral blood and lymphoid organs of young ruminants, they represent up to 50% of all T cells , but their role in combating N. caninum infections is not yet known . In murine models, γδ-T cells may also have the capacity to trigger foetal losses by reacting against the foetal trophoblast . After N. caninum infection in early gestation, placentas contain moderate numbers of γδ-T cells, which can increase to large numbers if dams are carrying dead foetuses . The mild γδ-T cell infiltration that was detected in the present trial supports the hypothesis that, during pregnancy, an anti-Neospora maternal immune response at a later stage is less harmful than one that occurs in early gestation.
Natural killer (NK) cells play an important role in the early response to a wide variety of pathogens, including N. caninum, and also direct the adaptive immune response towards a Th-1 response. Furthermore, after N. caninum infection during early gestation, dams carrying live foetuses show a lower number of NK cells when compared with dams with dead foetuses, suggesting a role for these cells in the immunopathogenesis of neosporosis . During the present trial, mean scores of NK infiltration were relatively constant throughout the period of study, unlike that observed with T cells (CD3+, CD4+, CD8+ and γδ-T cells). The degree of infiltration of NK cells was similar when compared with observations at early gestation in dams carrying live foetuses .
In previously reported work, B-cell-deficient mice were shown to be increasingly susceptible to cerebral N. caninum infection, suggesting an important role for B cells in host immunity against N. caninum. Nevertheless in the present study, only rare or single CD79αcy+ cells were observed with no differences between negative control and N. caninum-inoculated dams, similarly to infection in early gestation  and after recrudescence of infection in mid and late-gestation . These data suggest that these cell types are probably not involved in the immunopathogenesis of neosporosis in pregnant cows. However, further analysis will be required in order to elucidate the true identity of the CD79αcy+ cells, because positive labelled cells morphologically and histologically resembled trophoblast cells instead of B cells. They were located in the trophoblast layer and even though the majority were mononucleated cuboidal cells, occasional binucleated cells were also labelled. More studies using other mAbs need to be carried out in order to establish the presence or absence of B cells in these placentas, and to resolve the identity of these CD79αcy+ cells.
The observed differences in the pattern of cellular responses at different stages of gestation may be attributable to differences in the immunological environment which allow or hamper the multiplication of the parasite within the placenta. This is further reflected in the clinical outcome of infection at different stages of pregnancy. Previous studies have reported N. caninum specific cellular proliferation responses and a corresponding production of IFN-γ early in gestation; while in mid-gestation these immune responses are not as powerful, allowing Neospora transmission. This suggests that biological changes associated with pregnancy may allow reactivation of tissue cysts of N. caninum leading to the release of bradyzoites [28, 55]. The host immune response may also be influenced by hormones produced during pregnancy. Progesterone and prostaglandin E2 are known to bias a T-cell response towards a Th2 phenotype during human pregnancy [56, 57]. Steadily increasing levels of progesterone in the plasma of pregnant cattle have been documented from early to mid-gestation; then these values significantly declined in the last few weeks of gestation . Collectively, these observations tend to favour a bias towards a more regulatory Th2-type cytokine milieu during normal pregnancy, especially in mid-gestation . In accordance with previous works [19, 28], a stronger Th1 response was expected in this study, whereas we observed a milder Th1-type cellular balance compared with responses to infection during early stages of pregnancy , possibly resulting in the minor clinical consequences of the disease in late gestation. However, it is important to consider that the immunological maturity of the foetus may also play an important role in influencing the pathogenesis of a Neospora infection.
Immunological changes have been reported during normal pregnancy, seemingly playing a critical function in the protection of the foetus and in the “housekeeping” of the placenta [34, 60, 61]. In the current study no phenotypic differences were observed in the placentas from the negative control animals over a period of 42 days (see Additional file 2) so it seems unlikely that the immune cell infiltrate observed in the placentas of dams inoculated with Neospora was simply due to physiological changes of gestation.
No differences were observed for each of the immune cell markers between the negative control animals culled at 28, 42 and 56 dpi when compared with the one negative control animal culled at 14 dpi, which was N. caninum positive by PCR . This observation is in agreement with other immunological findings, since this animal was serologically negative and do not respond to Neospora antigen in the lymphoproliferation assays (Paul Bartley, unpublished observations).
In conclusion, following infection of pregnant cattle at 210 days of gestation, the immune cell infiltrate in placental tissues was milder than that observed in similar studies that investigated infection at earlier gestational stages and can partially explain the milder clinical outcome, i.e. congenital transmission but no abortions.
Previous reports on neosporosis show an association between production of inflammatory cytokines, such as IFN-γ, TNF-α and IL-12 and disease pathogenesis [28, 41–44, 46–48, 55]. However, it remains unclear which cells are responsible for the production of these cytokines and how the cells and cytokines relate to pathogenesis, particularly at different stages of gestation. Further studies are needed in order to clarify this aspect of the disease.