Heat stress reduces sexual development and affects pathogenesis of Eimeria maxima in meat-type chickens

Coccidiosis, caused by Eimeria spp. presents a self-limiting intestinal infection of poultry. Intestinal replication of the parasite causes severe morphological alterations to the host gastrointestinal tract, marked, among others, by the disruption of the intestinal barrier. We have previously reported a significant reduction in merozoite replication and oocyst shedding in E. tenella in vitro and in vivo. The objective of this study was to investigate the pathogenesis of E. maxima infection in broiler chickens under heat stress (HS) and mRNA expression of host cytokines that might affect the curtailed development of the parasite. We herein demonstrate that there is a significant detrimental effect of HS on the pathogenesis of E. maxima infection in broilers. There was a restricted replication of the parasite in HS chickens evidenced by significantly reduced oocyst shedding and disruption of the intestinal blood barrier. Gene expression of parasite genes demonstrated curtailed sexual reproduction of E. maxima in HS chickens. There was downregulation of Eimeria spp. genes related to gamete fusion, oocyst shedding, mitosis and spermiogenesis. Host gene expression indicates alterations in the cytokine expression that could be related to reduced parasite development in vivo.


Eimeria maxima detection by histopathology.
There was no histological evidence of coccidia infection in sections from uninfected control chickens (TNc and HSc) throughout the experiment. Mild to severe coccidial enteritis was observed at 7 dpi in chickens from the TNi group (Fig. 6). The diagnosis was marked by the presence of E. maxima in several stages of development in the lamina propria and by high numbers of organisms present in each section of intestine, varying from asexual to sexual stages (Figs. 6, 7). On the other hand, chickens from the HSi group were diagnosed as rare to minimal coccidial enteritis (Fig. 8), marked by the presence of scattered schizonts in the lamina propria (Fig. 8) and rare sexual stages (Fig. 9).
The mRNA expression of host IL-1, TNF-α, IL-6, IL-10, TGF-B1, TGF-B2, NF-κB-1 and NF-κB-2 was assessed at 5 to 7 dpi, and results depicted on Fig. 11. There was a significant downregulation of IL-1 in HSi as compared to TNc, TNi and HSc at 5 dpi, followed by upregulation of IL-1 in TNi at 6 dpi. IL-1 was downregulated in HSc as compared to HSi at 7 dpi, without statistical differences between TNc, TNi and HSC, or between TNc, TNi and HSi. There were no statistical differences in the expression of TNF-α at 5 dpi, however this gene was significantly upregulated in TNi at 6 dpi, and in HSc at 7 dpi. There were no statistical differences in the expression of IL-6 at 5 and 6 dpi between the treatments. At 7 dpi, IL-6 was upregulated in HSc as compared to TNi, however no statistical differences were recorded between TNc, TNi and HSi, or between TNc, HSc and HSi. IL-10 was . Quantification of Eimeria maxima mRNA in ileum samples of chickens infected with E. maxima and housed at heat stress (HSi; red) environment as compared to infected chickens housed at thermoneutral (TNi; green) environment, assessed by RT-qPCR from 5 to 7 dpi and expressed as 2 −∆∆Ct . Mean and standard error of the mean (SEM) are depicted. Data were analyzed by GLM procedure with multiple comparisons corrected with Tukey's method at 5% level of significance (p < 0.05). Significant differences between the groups within the same day are indicated by the superscript asterisk.
Scientific RepoRtS | (2020) 10:10736 | https://doi.org/10.1038/s41598-020-67330-w www.nature.com/scientificreports/ upregulated in HSi as compared to TNc and HSc at 5 dpi. There was a significant upregulation of IL-10 in TNi at 6 as compared to HSi, TNc and HSc. Similarly, there was a significant upregulation of IL-10 in HSi as compared to the uninfected controls at 6 dpi. The expression of IL-10 remained upregulated in HSi at 7 dpi, however there were no statistical differences in the expression of this gene in TNc, TNi and HSc at 7 dpi. The expression of TGF-β1 was unaltered at 5 dpi, downregulated in HSi at 6 dpi, and upregulated in TNi, and HSi at 7 dpi. The expression of TGF-β2 was downregulated in TNi at 5 and 7 dpi, and in HSi at 6 dpi. The expression of NF-κβ-1 was downregulated in TNi as compared to TNc at 5 dpi, with no statistical differences between TNi, HSc and HSi. There was a significant downregulation of NF-κβ-1 at 6 dpi. NF-κβ-1 was significantly upregulated in HSc at 7 dpi as compared to TNc and TNi. There were no statistical differences in the expression levels of NF-κβ-1 between HSc and HSi, between TNc and HSi, or between TNc and TNi. The expression of NF-κβ-2 was statistically similar between the groups at 5 and 6 dpi. There was a significant upregulation or NF-κβ-2 in HSi at 7 dpi as compared to TNc, but this upregulation was not significant when compared to TNi and HSc. Figure 5. Eimeria spp. gene expression analysis in ileum samples collected from chickens infected with E. maxima and housed at heat stress (HSi; red) environment as compared to infected chickens housed at thermoneutral (TNi; green) environment, assessed by RT-qPCR from 5 to 7 dpi and expressed as 2 −∆∆Ct . Genes analyzed were inner membrane complex 1 (IMC), elongation factor 2 (EF-2), gametocyte protein 56 (GAM56), gametocyte protein 82 (GAM82), hapless 2 (HAP-2) and protamine (Prot). Data were analyzed by GLM procedure with multiple comparisons corrected with Tukey's method at 5% level of significance (p < 0.05). Mean and standard error of the mean (SEM) are depicted. Significant differences between the groups within the same day are indicated by the superscript asterisk.

Discussion
In poultry, Eimeria spp. replicates in the intestine causing extensive tissue damage. E. maxima causes congestion and edema, cellular infiltration, thickening of the mucosa, with the later generations of schizonts and sexual stages developing deeper in the tissues, causing considerable disruption of the mucosa 2 , forming a port of entry for secondary infections such as necrotic enteritis 11 . We have previously reported the shortened life cycle of E. tenella under HS in vitro and in vivo, as evidenced by the reduction in merozoite replication and consequently oocyst shedding 9 . The objective of this study is to understand some of the underlying mechanisms of the detrimental effect of HS on Eimeria spp. and on its interactions with the host.
In the current study, HS was confirmed by physiological response of the animals, as marked by panting and opened wings. Our results demonstrate that, at the level of infection tested, E. maxima infection in meat-type chickens is as detrimental as HS when infection occurs at 14 days of age, as demonstrated by the similarity in   The recovery from infection coincided with significant reduction in oocyst counts, lesion scores and microscores in TNi as compared to HSi. We also demonstrate that HS by itself exerts severe detrimental effects for poultry production, independent of coccidiosis infection, such as lower development and weight gain. These results are in concordance with literature reports for uninfected heat-stressed broilers, indicating decrease in BW and FI 12 .
Although not statistically significant, we also demonstrate that TNi chickens have higher FCR as compared to HSi. Interestingly, our data also reveals that there are no statistical differences in FCR between TNc, HSc and HSi, indicating an effect of heat stress on parasite development and consequently, FCR.  www.nature.com/scientificreports/ The peak of intestinal lesions was at 6 dpi, and at that day, both TNi and HSi reached similar levels of lesions, indicating that both groups presented significant parasite replication in the intestine. Moreover, at 7 dpi, TNi showed a tendency towards less severe intestinal lesions, as compared to HSi, indicating a faster recover from infection. Following a similar trend, the peak of microscopic scores was at 6 dpi, with a similarity in scores between both infected groups (TNi and HSi). At 7 dpi, microscopic scores tended (non-significant) to be less severe in TNi chickens as compared to its counterpart, HSi. Altogether these data demonstrate that, at the level of infection tested, heat stressed chickens reach similar levels of intestinal damage caused by E. maxima and are rendered unprepared to overcome infection as fast as chickens raised in a thermoneutral environment.
Intestinal damage was also assessed by disruption of the intestinal blood barrier determined by the concentration of FITC-d in the serum. While both groups of infected chickens reached similar levels of intestinal barrier disruption at 6 dpi, only TNi chickens showed significant disruption of the intestinal blood barrier at 5 dpi, indicating a slower replication of E. maxima in heat stressed chickens as opposed to the faster replication of the parasite in chickens housed in a thermoneutral environment.
The reduction in oocyst shedding of E. acervulina 10 and E. tenella 9 under conditions has been previously reported. Similarly, HSi chickens also presented a significant reduction in oocyst shedding at 7 dpi as compared to TNi chickens. Parasite replication was further quantified by qRT-PCR. Our results demonstrate that E. maxima replicate faster in TNi chickens as compared to HSi, indicating that of the host also exerts a detrimental effect on E. maxima. There is a significant upregulation in 18 s at 7 dpi in HSi chickens, as compared to TNi, indicating that the life cycle of E. maxima in HS chickens might be in fact extended for a period longer than usual. This delay in parasite development at 7 dpi is also noticeable on histology sections of the small intestine. While moderate to severe coccidiosis was reported for TNi, marked by the presence of several stages of development of E. maxima in intestinal sections, including macrogametocytes and microgametocytes, only rare asexual stages of E. maxima were seen in sections from HSi chickens. These are indicatives of the extended asexual development of E. maxima in HS chickens. The significant increase in expression of 18 s at 7 dpi overlaps with the excretion of E. maxima oocyst in the feces. Therefore, this increase is most likely an artifact observed due to the excretion of E. maxima from the intestines of TNi chickens.
We evaluated the development of the parasite with a panel of genes related to oocyst production and macrogametocyte development (GAM56 and GAM82), Two putative genes (HAP-2 and Prot) previously related to microgametocyte development in E. tenella 13 , and two genes (IMC and EF-2) expressed throughout the development of E. maxima 14 . Gene expression analysis reveals a significant upregulation of IMC in HSi at 5 dpi and similar expression of EF-2 between HSi and TNi at 5 and 6 dpi, indicating an overall similar level of E. maxima asexual reproduction in both treatments. Our analysis also reveals significantly lower expression of Eimeria spp. genes related to the sexual reproduction (GAM56, GAM82, HAP-2 and Prot) in HSi as compared to TNi.
IMC is a component of the glideosome, the locomotory system believed to be used by E. maxima to achieve their characteristic gliding motility 14 . The upregulation of this gene in HSi at 5 dpi indicates higher motility of Figure 10. T lymphocyte cell populations and ratio in the spleen of chickens infected with E. maxima and housed at thermoneutral (TNi; green) or heat stress (HSi; red) environment, as compared to uninfected chickens housed at thermoneutral (TNc; blue) or heat stress (HSc; yellow) environments. CD4 + :CD8 + ratio at 7 dpi (a), CD4 + percentage at 7 dpi (b), CD8 + percentage at 7 dpi (c), CD4 + :CD8 + ratio at 14 dpi (d), CD4 + percentage at 14 dpi (e), CD8 + percentage at 14 dpi (f). Data were analyzed by GLM procedure with multiple comparisons corrected with Tukey's method at 5% level of significance (p < 0.05). Mean and standard error of the mean (SEM) are depicted. Significant differences between the groups are indicated by different superscript letters.
Scientific RepoRtS | (2020) 10:10736 | https://doi.org/10.1038/s41598-020-67330-w www.nature.com/scientificreports/ E. maxima in the ileum of these chickens as compared to TNi. There was a significantly lower expression of IMC at 6 and 7 dpi, indicating that at those time points, motility of merozoites is compromised in HSi chickens. EF-2 is a common immunodominant antigen 15 expressed in E. maxima 14 . EF-2 plays a crucial role in the elongation stage of mRNA translation in eukaryotes, by mediating the translocation of the ribosome relative to the mRNA after addition of each amino acid to the nascent chain 16,17 . The family of proteins is highly conserved and expressed in all eukaryotic cells, playing important functions in signal transduction during mitosis, apoptosis, control of the cell cycle, and utilization of proteins 18,19 . The downregulation of this gene in HSi at 7 dpi strongly indicates reduced E. maxima replication on this group of chickens. It is important to note, however, that the functions of IMC and EF-2 are putative and have not been yet confirmed at the protein level. The genes GAM56 and GAM82 code for formerly characterized gametocyte antigens and oocyst wall proteins, components of the wall-forming bodies of the macrogametes and integrated into the oocyst wall of E. maxima [20][21][22] . The upregulation of these two genes in TNi agrees with the augmented oocyst production as compared to HSi. HAP-2, a microgametocyte-specific gene previously reported in E. tenella 13 and Plasmodium, was also downregulated in parasites from HSi chickens. The expression of this gene is required for gamete fusion and subsequent fertilization of Plasmodium [23][24][25] and is believed to have similar functions in E. tenella, mediating the fusion of the membrane between mating gametes 26 . The downregulation of HAP-2 in HSi indicates that there is a significant reduction in parasite fertilization when hosts are exposed to HS. Prot is a histone-like protein that constrains sperm DNA, resulting in the condensation of the genome into an static state 27 . The gene has been shown to be upregulated in microgametes of E. tenella 13 . The downregulation of Prot in the parasite population from HSi chickens is in agreement with the downregulation of the other genes whose expression is specific to the sexual stage of Eimeria spp., indicating lower sexual reproduction of E. maxima in HSi as compared to TNi. At this point, however, the functionality of Prot and HAP-2 in E. maxima is only putative, based on data from other apicomplexan parasites. Further studies should confirm these genes' functions at the protein level. A previous study shows the effect of Eimeria spp on the reduction in white blood cell count, antibody production, lymphocytes numbers, spleen weight, macrophage activity; CD4 + and CD8 + lymphocyte counts 12 . Moreover, Khajavi, et al. 28 demonstrates that the increase in CD4 + :CD8 + ratio in HS chickens occurs despite decrease in both CD4 + and CD8 + numbers. Our data demonstrate that the CD4 + :CD8 + ratio increases in chickens upon infection and/or HS, without additive effects. Moreover, HS and E. maxima infection cause a significant increase in CD4 + at 7 dpi, while E. maxima infection causes a decrease in CD8 + cells at 7 dpi but there were no differences in CD4 + and CD8 + populations between TNi and HSi chickens. Furthermore, our data indicates that the detrimental effects of the host on the outcome of E. maxima infection is not due to immunosuppression, as demonstrated by the comparable CD4 + :CD8 + ratios in both groups of infected chickens, TNi and HSi, during the first and second weeks of infection. Further studies should determine whether HS during primary infection limits the host's immune response to challenge. Due to the systemic effects of HS in the host, we decided to assess the systemic variation in the CD4 + :CD8 + ratio. Future studies should also focus in the intestinal evaluation of T-lymphocytes.
It has been shown that Eimeria spp. activates cytokine release and the migration of the inflammatory cells that modulate the host immune system in different ways according to the species 29,30 . Stressors have also been reported to modulate the production and release of cytokines and other inflammatory mediators 31 . To better understand the changes in disease pathogenesis on the current study, we conducted a comprehensive panel analysis of anti-inflammatory and pro-inflammatory cytokines. Heat stress significantly downregulates the expression of pro-inflammatory cytokines (IL-1 and TNF-α) in the ileum of chickens infected with E. maxima, with the simultaneous downregulation of NF-κβ-1 and upregulation of IL-10. There is the possibility, however, that this diminished cytokine gene expression is a result of the reduced parasite development in HSi chickens. However, it remains unclear at this point the exact mechanism of cytokine response elicited by HS that limits E. maxima development. There was an upregulation of IL-10 at the peak of intestinal lesions (6 dpi) in TNi and HSi. Interestingly, there is an upregulation of IL-10 only in HSi at 7 dpi, putatively indicating continuation of the replication of this parasite population, corroborating the quantification of E. maxima population in the ileum.
Results from our previous studies have reported the significant reduction in merozoite replication and oocyst shedding in E. tenella in vitro and in vivo. We, herein, also show that there is a significant detrimental effect of HS in the outcome of E. maxima infection in broiler chickens, as demonstrated by reduced oocyst output post infection. Moreover, the restricted replication of the parasite in HS chickens is related to reduced expression of genes related to gamete fusion, parasite fertilization, and sexual reproduction, resulting in an overall diminished parasite development in chickens reared under HS conditions. Further studies should assess if these changes are translated at the protein level. We also conclude that, at the level of infection tested, E. maxima induces a systemic downregulation of CD8 + lymphocytes, whereas HS and E. maxima infection induce upregulation of CD4 + lymphocytes. Also, a very diverse cytokine response shows indicatives of reduced inflammatory response during concomitant HS and infection, suggesting this to be one of the mechanisms resulting in reduced sexual replication of E. maxima in heat stressed chickens. www.nature.com/scientificreports/ one volume of feces was solubilized in 9 volumes of saturated salt solution and the supernatant used to verify for the presence of oocysts in a McMaster chamber. The oocysts recovered from one of the chickens were sporulated as previously reported 32 , and used to infect a second (p2) and third (p3) passages in chickens. The oocysts from the third passage (p3) were later sporulated and used for the experimental infections of chickens. The purity of the E. maxima clone was verified by PCR following protocol reported by Jenkins, et al. 33 .

Methods
experimental design. Three hundred 14-days old Ross708 broiler chickens were divided into 30 groups of 10 chickens each and infected via gavage with 2 × 10 5 Eimeria maxima sporulated oocysts suspended in water and housed at two different temperatures: 15 groups were housed at 20 °C (I-20) and 15 groups at 35 °C (I-35). Similarly, another 300 chickens were mock infected with water and housed at 20 °C (C-20) and 35 °C (C-35). All chickens were raised in batteries with wired floor, with ad libitum access to water. Individual body weights and feed consumption were recorded at the day of infection (day 0), at 7 and 14 days post infection (dpi). From 4 to 7 dpi, 10 chickens from each treatment had blood samples collected and were euthanized by cervical dislocation, following with collection of histology samples of liver, spleen and bursa, intestinal content from jejunum, ileum and caeca. Intestinal lesions and the presence of E. maxima developmental stages (micro-scores) were scored from 4 to 7 dpi. Oocyst shedding was estimated on feces collected from underneath the pens from 4 to 7 dpi and at 14 dpi. Spleen samples collected at zero, 7 and 14 dpi were used to assess T-cell immune response. Quantification of E. maxima developmental stages in the ileum was assessed by qPCR from 1 to 7 dpi, and by merozoite collection from 2 to 4 dpi.
intestinal permeability. Disruption of the intestinal blood barrier was assessed as previously reported 34 .
Briefly, FITC-d was gavaged to chickens at 2.2 mg of FITC-d/bird. Two hours post administrations, blood was collected from the jugular vein and stored in tubes protected from the light and kept at room temperature for 3 h to allow clotting, following with centrifugation (1000 g for 15 min) to separate serum.  www.nature.com/scientificreports/ measured at the end of each cycle. In addition, at the end of the reaction, the melting temperature curve of each RT-qPCR was determined. Each biological sample was runt in triplicates using StepOnePlus (Applied Biosystems, Carlsbald, CA). Relative expression of chicken genes was calculated from the amount of the gene-specific cDNA (target) normalized to the amount of chicken β-actin (endogenous control), using the 2 −∆∆ct method 36 , in which differential mRNA expression was expressed as treatment groups relative to the control group (TNc). Relative expression of Eimeria-specific genes was calculated from the amount of the gene-specific cDNA (target) normalized to the amount of 18 s (endogenous control), in which differential mRNA expression was expressed as treatment groups relative to the TNc group. Relative quantification of E. maxima in ileum samples was calculated from the amount of 18 s cDNA normalized to the amount of chicken β-actin, in which differential mRNA expression was expressed as treatment groups relative to the TNc group. Ileum samples were collected in a daily basis, fixed in 10% buffered formalin, trimmed into cassettes, routinely processed and embedded in paraffin. Tissue sections were cut and stained with hematoxylin and eosin (H&E), coverslipped and analyzed by light microscopy.   www.nature.com/scientificreports/ 10 °C, followed by removal of the supernatant and fixation of the cells in 200 μL of 2% formaldehyde for 15 min. Post fixation, the plate was centrifuged (800 rcf, 4 min, 10 °C), the fixing solution was removed and the cells were resuspended in 200 μL of flow cytometry buffer. Samples were analyzed by flow cytometry in up to 24 h post fixation. Flow cytometry was performed on a CMM CytoFLEX (Beckman Coulter, Indianapolis, IN) with 5,000 events acquired per sample. Samples were analyzed in 2 channels: FITC emission using a 488-525 nm filter, and PE emission using a 561-585 nm filter. A gate was set to encompass the majority of single cells using the images generated by bright field. The FITC and PE gates were drawn based on the intensity of FITC vs. the intensity of PE in single-stained and non-stained controls. Cell populations in the samples were determined post compensation with single-stained and non-stained controls. The CD4 + :CD8 + ratio was calculated by dividing the number of CD4 + T cells by the number of CD8 + T cells.
performance parameters. Body weight (BW), feed intake (FI), feed conversion ratio (FCR) and relative growth (RG) were assessed at 7 and 14 dpi. Feed conversion ratio (FCR) was calculated using the following formula: FCR = FI/BW gain. Relative growth was calculated using the following formula: Relative growth = body weight/initial weight × 100, and results expressed as g/g.
Statistical analysis. Statistical analysis of BW, FI, FCR, RG, lesion scores, microscores, FITC-d and OPG was conducted in Prism 7.0 software (Graphpad Software Inc., San Diego, California). D' Agostino and Pearson test was applied to determine normality and determine whether data should be analyzed by parametric or non-parametric tests. Kruskal-Wallis test with multiple comparisons corrected using Dunn's method was used to compare non-parametric data. Parametric data were analyzed by two-way-ANOVA test with multiple comparison's corrected by Tukey's method. Gene expression and lymphocyte population analysis was conducted in SAS software (SAS Institute, Cary, North Carolina) by GLM procedure with multiple comparisons corrected with Tukey's method. All tests were performed at the 5% level of significance. All results are expressed as mean ± standard error of the mean.