Identification of a Novel Cerebral Malaria Susceptibility Locus (Berr5) on Mouse Chromosome 19

ORIGINAL ARTICLE Identification of a novel cerebral malaria susceptibility locus (Berr5) on mouse chromosome 19

J Berghout1, G Min-Oo1,MTam2, S Gauthier1, MM Stevenson2 and P Gros1 1Department of Biochemistry, McGill University, Montreal, Quebec, Canada and 2Research Institute of the Montreal General Hospital, McGill University Health Center, Montreal, Quebec, Canada

Cerebral malaria (CM) is an acute, generally lethal condition characterized by high fever, seizures and coma. The genetic component to CM can be investigated in mouse models that vary in degree of susceptibility to infection with Plasmodium berghei ANKA. Using survival time to measure susceptibility in an informative F2 cross (n ¼ 257), we identified linkage to chromosome 19 (Berr5 (Berghei resistance locus 5), LOD ¼ 4.69) controlling, in part, the differential response between resistant BALB/c and susceptible C57BL/6 progenitors. BALB/c alleles convey increased survival through the cerebral phase of infection but have no quantitative effect on parasitemia during the later, anemic phase. The Berr5 locus colocalizes with three other immune loci, including Trl-4 (tuberculosis resistance), Tsiq2 (T-cell secretion of IL-4) and Eae19 (experimental allergic encephalitis 19), suggesting the possibility of a common genetic effect underlying these phenotypes. Potential positional candidates include the family of Ifit1–3 (interferon-inducible protein with tetratricopeptide repeats 1–3) and Fas. Genes and Immunity (2010) 11, 310–318; doi:10.1038/gene.2009.79; published online 29 October 2009

Keywords: genetics; malaria; QTL; inflammation; Plasmodium berghei; Ifit

Introduction strains such as C57BL/6 (B6) or CBA developing an acute cerebral syndrome within 6–7 days characterized Malaria is a major global health problem with approxi- by ataxia, paraplegia, seizures and coma, leading to mately 40% of the world’s population living in at risk uniform lethality by day 8–10 after infection. On the areas. In Africa, malaria is responsible for 20% of other hand, resistant mice including the BALB/c and childhood deaths with a cumulative mortality of 1–3 WLA strains do not develop neurological symptoms but million people each year.1 The most severe complication die 2–4 weeks after infection due to severe anemia of Plasmodium falciparum infection is cerebral malaria caused by high levels of blood parasitemia. Current (CM). CM is caused by infected erythrocytes becoming evidence suggests that CM in mice infected with PbA is trapped in brain microvasculature, leading to an caused in part by a robust but detrimental localized excessive, localized host-driven inflammatory response inflammation in response to parasitized erythrocytes and resulting in hypoxia, loss of blood–brain barrier trapped in the brain microvasculature.2 Transcript (BBB) integrity and cell death.2–4 Patients develop profiling experiments have documented a substantial seizures, paralysis, ataxia and unrousable coma that pro-inflammatory response that is more pronounced in often ends in death. The cellular mechanisms of the susceptible mice.7–9 In addition, loss-of-function muta- pathological response are poorly understood and must tions in genes coding for pro-inflammatory molecules be better characterized to develop strategies for preven- such as interferon-g (Ifng), lymphotoxin (Lta), hemolytic tion and treatment of this condition. complement (C5a) and interferon regulatory factor 1 The host genetic background has a major effect (Irf1) have been shown to protect against CM.10–13 These on response to malaria in humans (reviewed by observations support the hypothesis of for a central Kwiatkowski5 and Mazier et al.6). The genetic component function for immune-mediated pathology in CM, with T of susceptibility to CM is multigenic and further cells, platelets and resident macrophages implicated in modified by parasite virulence and other environmental inflammatory response and tissue injury.14–16 determinants. Inbred strains of mice exhibit different Quantitative trait locus (QTL) mapping studies by responses to P. berghei ANKA (PbA), with susceptible whole-genome scanning in mice have detected five chromosomal loci affecting the host response to PbA infection. Using survival as a phenotypic trait—an Correspondence: Dr P Gros, Department of Biochemistry and unnamed locus on chromosome 18 in F2 mice derived Complex Traits Program, McGill University, Bellini Life Sciences from susceptible B6 and resistant DBA/2 was first Building, 3649 Promenade Sir William Osler, Room 366, Montreal, identified.17 Subsequently, QTLs designated Berr1 (chro- Quebec, Canada H3G-0B1. mosome 1), Berr2 (chromosome 11), Berr3 (chromosome 9) E-mail: [email protected] Received 26 May 2009; revised 17 September 2009; accepted 22 and Berr4 (suggestive linkage to chromosome 4) affecting September 2009; published online 29 October 2009 the presence and severity of neurological symptoms Berr5 in cerebral malaria resistance J Berghout et al 311 following infection were mapped in F2 mice derived from BALB/c brains showed very little uptake of the blue dye, C57BL/6 (susceptible) and the resistant wild-derived with levels similar to those seen in uninfected controls WLA strain.18,19 Finally, the Cmsc locus mapping to the H- (data not shown). These results confirm increased PbA- 2 region of chromosome 17 was found to affect suscept- induced BBB permeability as a distinguishing feature ibility in progeny of CBA mice crossed with the resistant between B6 and BALB/c animals. DBA/2 strain.20 In all cases, the detected QTLs span large To study the genetic complexity of the B6 vs BALB/c chromosomal segments, and the causative genes under- differential susceptibility to CM, we infected 257 lying their effect on survival time in P. berghei-infected (C57BL/6 Â BALB/c)F2 mice (CB6F2) with PbA. CB6F2 animals have not been identified. mice segregated into two groups, with a majority of In this study, we have investigated the genetic basis of animals (B75%) dying in the early cerebral phase (before differential susceptibility of the commonly compared B6 day 13), and the remainder surviving past day 17. The strain and the BALB/c strain to infection with P. berghei experiment was terminated at day 22 (Figures 2a and b). ANKA. All animals dying from early CM (susceptible, CM-S) had low parasitemia (o20%), whereas those dying late (resistant, CM-R; including those killed at day 22) developed high parasitemia (40–90%). In addition, the Results predictive value of blood parasitemia at day 6 (Figure 2c) We compared the response of the inbred C57BL/6 and and serum IFNg levels at day 5 (Figure 2d) on overall BALB/c strains of mice to infection with PbA. B6 were susceptibility was determined. We observed that CM found to be susceptible with the majority of animals susceptibility as measured by survival was significantly dying during the early cerebral phase of infection correlated (P ¼ 0.0001) with both higher early parasite- (Figures 1a and b). Neurological symptoms and death mia values and higher IFNg levels. in these mice occurred while levels of blood A genome-wide scan was carried out on all 257 CB6F2 parasitemia were o20% (data not shown). Conversely, mice, using 252 informative markers with survival as a BALB/c mice were resistant to PbA-induced CM, with marker of CM susceptibility. Both a binary (CM-S vs 475–80% of the animals surviving the cerebral phase. CM-R) and a parametric model of survival time (two- Surviving mice did not develop any subsequent cerebral part model) were used in this analysis that yielded symptoms. Instead, they developed high levels of blood similar results. The two-part model takes into account parasitemia (460%; data not shown) and died of severe both the binary trait (probability of survival to the malarial anemia. experimental end point) and the approximately normally CM in susceptible PbA-infected mice has been distributed survival within the cerebral phase (survival previously associated with disruption of the integrity of time contingent on not reaching the end point).22 Both the BBB,21 which we monitored using a dye extravasa- analyses identified a single genetic effect on chromosome tion assay (Figure 1c). Examination of brains from Evan’s 19, with maximum linkage to the CEL-19_32349880 Blue-injected mice revealed that in B6 mice the entire marker at chromosomal position 32.3 Mb (Figures 3a tissue was dyed blue with several darker foci whereas and c). Using the two-part model, we found that the

CM-S 100 100 CM-R 80 80

60 60

40 40 Proportion (%)

Percent survival 20 BALB/c 20 C57BL/6 0 0 0 5 10 15 20 C57BL/6 BALB/c Time (days) n=22 n=25

C57BL/6 BALB/c Figure 1 Differential susceptibility of inbred mouse strains to Plasmodium berghei-induced cerebral malaria. (a) Mice were infected by the intraperitoneal route with 5 Â 104 P. berghei ANKA parasitized erythrocytes and survival was monitored. Neurological symptoms began to appear at day 6 in susceptible (CM-S) C57BL/6 mice. The majority of BALB/c mice were categorized as resistant (CM-R), surviving well past day 12. (b) Proportion of CM-S (filled bars; dead before day 10) and CM-R (empty bars; survive past day 12) mice detected in each strain. (c) Integrity of the blood–brain barrier in PbA-infected B6 and BALB/c mice was evaluated 8 days after infection using an Evan’s Blue dye permeability assay in vivo (see Materials and methods section).

Genes and Immunity Berr5 in cerebral malaria resistance J Berghout et al 312 100 100 20 7000 *p=0.0001 6000 *p=0.0001 80 15 75 5000 4000

60 10 (pg/ml) 50 γ 3000 IFN 2000 40 (%) Parasitemia 5 Proportion (%) 25 1000

Peak Parasitemia (%) Parasitemia Peak 20 0 0 0 /c B CB6F2 L BL/6 0 BALB/c BA CM-S C57BL/6 C57 2 : CM-R 0 3 6 9 12 15 18 21 6F2: CM-S CM-R CB CB6F2: CM-R Survival day CB6F2 CB6F: CM-S Figure 2 Segregation analysis of susceptibility to Plasmodium berghei-induced cerebral malaria in (C57BL/6 Â BALB/c)F2 mice. Daily survival and level of blood parasitemia were recorded for (C57BL/6 Â BALB/c)F2 mice (n ¼ 257) infected with PbA. (a) The level of parasitemia at the peak of infection, or immediately before moribund animals were killed, along with the time of survival are shown. (b) The fraction of CB6F2 mice showing either a CM-S phenotype (dying before day 13 with o25% parasitemia) or a CM-R phenotype (surviving past day 13 with high 40–90% blood parasitemia) is shown. (c) Day 6 blood parasitemia values in parental and F2 mice, classified as either CM-S or CM-R as described in b. CM-S F2s showed significantly higher blood parasitemia at day 6 than CM-R F2s. (d) Day 5 serum interferon-g (IFNg) levels in parental and F2 mice, classified as either CM-S or CM-R.

peak LOD score was 4.69 (Po0.05) whereas the expression in response to PbA infection for the genes in binary analysis yielded a peak LOD score of 3.51 at the minimal Berr5 interval using transcript profiling the same marker (Po0.05). This locus controlling PbA- information from B6 and BALB/c brains published by induced CM was given the temporary identification Lovegrove et al.9 Examination of their data showed that Berr5 (Berghei resistance locus 5). The 95% confidence the expression of 15 genes within the 95% confidence interval for the Berr5 QTL was estimated to be interval for Berr5 interval are differentially modulated in approximately 10–20 Mb (25.6–45.1 Mb based on a 1.5 brain tissue in response to PbA infection and/or in a LOD decrease; 29.3–41.3 Mb using the 95% Bayesian strain-specific manner (two-way analysis of variance credible interval) (Figure 5a). Berr5 was estimated to (ANOVA) with a 1.4-fold cutoff, Bonferroni correction). explain between 5.9 and 8.8% of the phenotypic variance. Figure 5b compares the level of expression of these genes We also used day 6 parasitemia as a quantitative in B6 and BALB/c in response to infection (presented as measure in whole-genome scanning. This analysis a ratio compared to levels in uninfected BALB/c revealed a single suggestive linkage peak on chromo- controls). Of particular interest are the three members some 19 in the Berr5 interval (LOD ¼ 3.35; Figures 3b and of the interferon-induced protein with tetratricopeptide c). These results show that Berr5 alleles can affect both repeats (TPRs; Ifit) gene family. By microarray, these early parasitemia and CM susceptibility in CB6F2 mice. three genes show a dramatic upregulation (between 8- This colocalization agrees with the results in Figure 2c, and 30-fold) in infected B6 brains compared to BALB/c. showing an association between day 6 parasitemia levels To confirm this observation, we performed semiquantita- and CM-associated death. tive reverse transcription (RT)-PCR for Ifit1–3 (Figure 6), The effect of Berr5 alleles on the phenotypic markers of which verified significant differential induction of the CM susceptibility was examined in the CB6F2 mice genes controlled by strain-specific genetic factors over (Figure 4). Homozygosity for B6 (susceptible) parental the course of infection. In addition, the tumor necrosis alleles at the peak marker was associated with CM-S, factor (TNF) receptor family member-encoding gene, Fas, with only 9% of these animals surviving the cerebral maps within the Berr5 interval. These two groups of phase whereas homozygosity for BALB/c (resistant) genes constitute excellent candidates for the Berr5 effect, alleles was associated with increased CM-R and 39% and thus to explain the differential host response of the survival through the cerebral phase (Po0.005, Kruskal– B6 and BALB/c strains. Wallis). Heterozygosity was also associated with increased resistance (31% survival), suggesting that resistance alleles at Berr5 are inherited dominantly (Figures 4a Discussion and c). Similarly, B6-derived susceptibility alleles at Berr5 were associated with higher day 6 parasitemia in CB6F2s CM is a complex pathology that involves interactions when compared with either homozygotes or heterozy- between host genetics, the Plasmodium parasite and the gotes for BALB/c-derived resistance alleles at Berr5 type and magnitude of the host response to infection (Po0.01, t-test; Figure 4b). (reviewed by Haldar et al.23). Although CM symptoms The Berr5 interval spans 25 Mb and contains 167 may appear while levels of blood parasitemia are still annotated genes (Supplementary Table A) that could be low, host-driven inflammatory response at the site of reduced to 14 Mb and 84 genes by superimposing the infection in the form of increased pro-inflammatory linkage peak obtained using day 6 parasitemia mediators (TNFa, IFNg), has been proposed to contribute (Figure 5a). We analyzed possible modulation of mRNA to CM pathology (reviewed by Hunt and Grau24).

Genes and Immunity Berr5 in cerebral malaria resistance J Berghout et al 313 p=0.05 levels (LOD ¼ 3.35). Although maximum LOD scores are 4 obtained for both phenotypes with the same marker 3 (CEL-19_32349880), the strength of the linkage is greater when survival is used as the marker of CM susceptibility. 2 In both cases, B6-derived Berr5 alleles are associated with 1 increased susceptibility among the CB6F2s. The weaker linkage observed when using parasitemia as a quantita- 0 tive trait may be the result of the fairly narrow numerical 1234567891011121314151617 1819 X range for this trait, and the asynchronous aspect of the rise of blood parasitemia in infected mice. Therefore, it is 3 possible that survival and parasitemia may be influenced by the same gene but with different penetrance/ 2 expressivity or differences in ascertainment for the two 1 traits. It is also possible that both traits are influenced by

LOD Score distinct but closely linked genes. Additional experimen- 0 tations will be required to distinguish the two possibi- 12345678910111213141516171819X Chromosome lities. Nevertheless, the fact that Berr5 alleles appear to influence both early parasitemia and early death merits 5 further discussion. One explanation would be that Berr5 Survival: Two part influences the extent PbA replication early during Survival: Binary infection, with a threshold effect triggering cerebral 4 Parasitemia (d6) symptoms and early death. Although possible, this explanation is incompatible with the observation that 3 all mice surviving the cerebral phase go on to develop high parasitemia without CM. A second explanation would be that Berr5 has little effect on the early rate of LOD Score2 LOD Score replication, but rather influences the host response to rising parasitemia. In this case, host inflammatory 1 response would be proportional to the extent of early parasitemia (for example, measured at day 6), and would CEL-19_32349880 * be associated with CM in mice homozygous for B6 alleles 0 at Berr5 at a much higher rate than mice carrying the 0 5 10 15 20 25 30 35 40 45 50 55 60 BALB/c allele. Finally, it is possible that Berr5 haplotype Marker Position (Mb) somehow influences the balance between the beneficial Figure 3 Mapping of genetic effects (Berr5) influencing suscept- and immunopathological effects of the early inflamma- ibility to Plasmodium berghei-induced cerebral malaria. Whole- tory response. genome scanning was carried out in 257 (C57BL/6 Â BALB/c)F2 There are a total of 167 positional candidates in the mice, using time of survival after infection with P. berghei ANKA estimated genetic interval for Berr5. The information in or day 6 parasitemia. Interval mapping was carried out using the R/qtl software package, and results for all chromosomes are Supplementary Table A shows that 46 of these genes are plotted. (a) Survival time was analyzed according to the two-part expressed in the brain and 27 in the hemolymphoid model of Boyartchuk et al.22 A significant linkage peak (LOD ¼ 4.69) system (22 expressed in both; Mouse Genome Informatics was detected on chromosome 19. Permutation testing (1000 tests) Gene Expression Database; http://www.informatics.jax. revealed the cutoff for genome-wide significance (Po0.05) in this org/expression.shtml), two systems known to be involved model to be 4.19 (dashed line). (b) A suggestive linkage to the same in CM. Examination of transcript profiling data9 revealed portion of chromosome 19 was also detected using day 6 parasitemia as a quantitative trait. (c) LOD score plots were generated for strain-specific factors and/or PbA infection regulated the chromosome 19, using either time of survival in the two-part model expression of 15 of these 167 genes (Figure 5; 1.4-fold (black line) or binary analysis of CM-S and CM-R phenotypes threshold). When analyzing both the absolute level of (dashed line), and day 6 parasitemia (gray line), which all point to infection-induced modulation, and interstrain differences the same linkage peak. in this modulation between B6 (CM-S) and BALB/c (CM-R), most differences are modest, with the notable exception of the Isg56/Ifit gene family (Isg56/Ifit1, Isg54/ Ifit2 and Isg49/Ifit3) that were much more highly Genetic studies in humans have shown that functional upregulated in B6 mice. The Ifit gene family has three promoter polymorphisms affecting the level of expression members in mice and four members in humans, each of TNFa mRNA can affect the incidence of CM in field clustered on small chromosomal segments, likely having studies.25 Likewise, studies in the PbA mouse model of arisen by two successive gene duplication events. In the CM have shown that loss-of-function mutations in pro- mouse, the three Ifit genes are each organized into two inflammatory molecules or their receptors have protec- exons and code for highly similar proteins.27 Ifit proteins tive effects against CM.12,26 are highly hydrophilic cytoplasmic proteins28 containing Here, we have studied the differential susceptibility of multiple TPR motifs that represent degenerate protein inbred mouse strains B6 (CM-S) and BALB/c (CM-R) to interaction modules. All three mouse Ifit genes show two PbA-induced CM. We have identified a locus on functional interferon-stimulated response element motifs chromosome 19 that we have designated Berr5, which in their proximal promoters (À60 to À100 position), which influences both survival time (two-part model; have a critical function in the regulation of mRNA LOD ¼ 4.69) and the early (day 6) blood parasitemia expression.27 Expression of Ifit genes is part of an early

Genes and Immunity Berr5 in cerebral malaria resistance J Berghout et al 314 *p=0.0091 20

100 CC *p=0.001 BC 80 BB 15

60 10 40 Survival (%)

20 (%) Parasitemia 5

0 0 3 6 9 12 15 18 21 0 Day CC BC BB

BALB/c allele Heterozygote B6 allele 100

50 Parasitemia (%) Parasitemia 25

0 10 20 10 20 10 20 Day Day Day

0 0 0

50 50 50 CM-S

Proportion (%) CM-R 100 100 100 61%39% 69% 31% 91% 9% Figure 4 Effect of haplotype combination at Berr5 on susceptibility to Plasmodium berghei-induced cerebral malaria. Effect of haplotype combination at Berr5 on survival (a) and day 6 parasitemia (b)inP. berghei-infected CB6F2 mice. In panels a and b, ‘C’ refers to the BALB/c allele at Berr5 and ‘B’ refers to the B6 allele. (c) Distribution of CM-S and CM-R phenotypes in CB6F2 mice according to their Berr5 haplotype. Mice homozygous for the C57BL/6 allele (91%) were CM-S, whereas only 61% of mice carrying the BALB/c allele were CM-S (Po0.0001, w2-test).

response to infectious agents or products derived from Also mapping within the Berr5 interval is the Fas gene. them, and can be induced by many stimuli including viral Although the differences in Fas mRNA levels noted were infection (Sendai virus, herpes simplex virus, human not as impressive as those noted for Ifit genes, the known cytomegalovirus, West Nile virus), double-stranded RNA, functions of the Fas protein (member of the TNF-receptor type I interferon (a, b), lipopolysaccharide (LPS), Toll-like superfamily) in inflammation and the pathogenesis of receptor ligands and several others.29 Their molecular various malignancies of the immune system make it an function is not fully understood, but they have been attractive positional candidate for the Berr5 effect.31 reported to inhibit translation by binding to subunits of Hypomorphic Faslpr mouse mutants show increased the eIF3 complex.29 In addition, it has been shown that resistance to PbA-induced CM indicating a pathological overexpression of Ifit2 in mouse macrophages (RAW264.7) function for this molecule.32 We have noted lower Fas modulates LPS-induced expression of TNFa, interleukin-6 RNA expression in PbA-infected brains from BALB/c (IL-6) and MIP-2.30 These data suggest that Ifit proteins (CM-R) compared with B6 (CM-S) (Figure 5). Interest- may have a function in early host response to different ingly, a functional promoter polymorphism (rs50344448; types of infection, which supports the possibility that they pst À375) has been detected in the mouse Fas gene: B6 may be responsible for the Berr5 effect on CM. carries an A at that position whereas BALB/c carries a

Genes and Immunity Berr5 in cerebral malaria resistance J Berghout et al 315 25Mb

Vldlr B6 d0 BALB/c d0 B6 d6 BALB/c d6 B6 d0 BALB/c d0 B6 d6 BALB/c d6

Cd274 Il33 Ifit3

Pten Fas Ifit2 Ch25h CEL-19_32349880 Parasitemia QTL Lipa Ifit2

Berr5 Ifit1 Ifit3 Eae19 Ifit1

HPRT

Tsiq2 Blnk Exosc1 Cnnm1 Cutc 20 BALB/c d0 Scd2 Trl4 B6 d0 45Mb 15 BALB/c d6 B6 d6 Vldlr BALB/c d0 10 B6 d0 Cd274 BALB/c d6 Il33 B6 d6 Fold increase over BALB/c d0 5 Pten Fas 0 Ifit3 Ifit2 Ifit1 Ch25h Figure 6 Confirmation of differential Ifit1–3 upregulation in the Lipa brains of PbA-infected mice. Semiquantitative reverse transcription (RT)-PCR was carried out for each transcript in control and PbA- Ifit2 infected brains as described in the Materials and methods section. Gene ID (a) Amplicons for each target were run on agarose gels and then Ifit3 hybridized with specific [32P]dATP-labeled probes. The intensities Ifit1 of the hybridization signals were quantified using a phosphorima- ger. (b) Each replicate was normalized to Hprt and then expressed Blnk as a relative ratio over the signal for BALB/c brains at day 0. Exosc1 G. The latter allele alters a putative AP-1 binding site33 Cnnm1 and reduces Fas mRNA expression. Finally, Berr5 overlaps three previously mapped loci Cutc affecting host response to infectious or inflammatory stimuli (Figure 5a). The Trl-4 (tuberculosis resistance; Scd2 41–52 cM) locus affects susceptibility to pulmonary 0 102030 tuberculosis and modulates replication of Mycobacterium tuberculosis in the lungs of (C57BL/6 Â DBA/2)F2 mice.34 Ratio over BALB/c d0 B6 alleles at Trl-4 are associated with increased resistance Figure 5 Analysis of positional candidate gene expression in the to infection in the form of a robust early immune minimum genetic interval Berr5 and associated quantitative trait response to a mounting bacillary load. Another QTL loci. (a) Approximate position of three previously mapped loci overlapping Berr5 is the Tsiq2 (T-cell secretion of IL-4) associated with host response to inflammatory or infectious stimuli, 35 38 34 locus (24–43 Mb) that regulates the capacity of spleen including Tsiq2, Eae19, Trl-4 and Berr5 (this study). The interval þ 35 mapped using survival is indicated with a black box whereas the CD4 T cell to secrete IL-4 in response to stimulation. suggestive interval determined using day 6 parasitemia is marked BALB/c alleles at Tsiq2 are associated with high IL-4 in the stippled box. The marker that gave the strongest linkage for production (Th2 polarization of early T-cell response) all analysis (CEL-19_32349880) is also plotted. Differentially whereas B6 alleles are associated with lower IL-4 expressed positional candidate genes are also identified along with production (pro-inflammatory, Th1 polarization). Th1 their chromosomal positions. (b) Among the 167 genes in the Berr5 interval, the expression of 15 of them in brain was regulated response is characterized by expression of pro-inflam- in a C57BL/6J vs BALB/c manner after infection with P. berghei matory mediators including IFNg, TNFa, lymphotoxin (two-way analysis of variance (ANOVA), Bonferroni and Hoch and IL-12 all previously implicated in CM immuno- correction with a 1.4-fold cutoff). The relative expression by pathology.12,36,37 In contrast, IL-4-driven Th2 response is microarray of each of these 15 genes is shown calculated as a ratio associated with low-level expression of pro-inflamma- (fold expression) over expression levels measured in uninfected tory cytokines but elevated levels of IL-4, IL-6, IL-10 and brains from BALB/c mice. TGF-b, which are thought to have an immune-regulatory,

Genes and Immunity Berr5 in cerebral malaria resistance J Berghout et al 316 protective function against CM.24,35,36 It is interesting to (Montreal, QC, Canada) using the low-density Illumina speculate that, by analogy with Tsiq2, Berr5 may GoldenGate mouse panel that contained 246 informative modulate CM susceptibility by altering the Th1/Th2 markers distributed across the genome. Fine mapping balance after infection with PbA. The third overlapping and filling gaps larger than 40 Mb was carried out using QTL (Eae19 (experimental allergic encephalitis 19); microsatellite markers obtained from the Mouse Genome 26–53 cM) affects the severity of EAE that is used as a Informatics Database (www.informatics.jax.org).39 mouse model of multiple sclerosis, a known inflammatory disease of the central nervous system (CNS). Eae19 Data analysis (26–53 cM) is specifically associated with brain demyeli- Survival of CB6F2 mice was analyzed by the Kaplan– nation in response to injection of heterologous CNS Meier log-rank test and survival fractions were homogenates, a process thought to involve inflammatory compared using the w2 statistic (GraphPad Prism 5, San cytokines, demyelinating antibodies and apoptosis of Diego, CA, USA). Blood parasitemia and serum IFNg affected cells.38 It is interesting to speculate that Trl-4, levels were analyzed using ANOVA, followed by a Tsiq2, Eae19 and Berr5 may be pleiotropic manifestations Bonferroni post hoc test. LOD scores were calculated with of the same gene effect. The formal identification of the R/qtl software version 2.7.0 using survival time as a gene responsible for Berr5 and associated QTLs on this phenotype. Because the data were not normally dis- portion of chromosome 19 will involve the creation and tributed, we used the two-part model22,40 that was characterization of congenic and subcongenic lines to developed for phenotypes with a spike in the distribu- narrow down the physical interval of Berr5, as well as the tion. This method has been used to analyze similar data formal evaluation of positional candidates, including the sets.19 Data were also analyzed using the binary model, creation and testing of loss-of-function mutations in vivo. segregating mice into two categories based on survival past day 13. The cutoff used for genome-wide significance (Po0.05) was 4.19 for the two-part model and 3.50 Materials and methods for the binary model based on 1000 permutations. Animals Pathogen-free, 8- to 10-week-old inbred C57BL/6 (B6) Microarray analysis 9 and BALB/c mice were purchased from Charles River Gene expression data sets published by Lovegrove et al. Laboratories (Wilmington, MA, USA). 257 (C57BL/6 Â were downloaded (GEO database; record GSE7814) and BALB/c)F2 progeny (CB6F2) were bred by systematic analyzed with the GeneSifter software package. These brother–sister mating of (C57BL/6 Â BALB/c)F1 pro- transcript profiles correspond to total brain RNA from B6 genitors (Charles River Laboratories). All animal experi- and BALB/c mice at day 0, 1, 3 and 6 after PbA infection, ments were carried out according to Canadian Council which had been hybridized to Affymetrix GeneChip on Animal Care guidelines and approved by the Animal Mouse Genome 430A 2.0 Array chips (Affymetrix, Santa Care Committee of McGill University. Clara, CA, USA). Expression was compared using two- way ANOVAwith a 1.4-fold cutoff, Bonferonni and Hoch Parasite and infection correction. In the Berr5 interval, 15 genes showed a P. berghei ANKA was obtained from the Malaria difference in expression level due to either strain-specific Reference and Research Reagent Resource Centre variation or over the course of infection. Ten genes (Ifit1, (MR4) and maintained as frozen stocks at À80 1C that Ifit2, Ifit3, Cd274, Ch25h, Blnk, Fas, Il33, Cutc and Vldlr) were passaged in B6 mice before use in infection. All were regulated by the interaction of genetic background experimental mice were injected intraperitoneally with and infection. Genes that were differentially regulated 5 Â 104 parasitized red blood cells. Daily blood para- between strains either during infection (day 6) or at basal sitemia was determined on Diff-Quick (Dade Behring, levels (day 0) were sorted according to chromosomal Newark, DE, USA) stained thin blood smears. Neurolo- location to identity positional candidates. gical symptoms and survival were monitored and CM was diagnosed by the presence of ataxia, weak gripping Semiquantitative RT-PCR and righting responses, paralysis, seizures and coma Uninfected control and day 6 PbA-infected C57BL/6 and between days 6 and 10. Serum was collected at day 5 BALB/c mice were killed and brains were harvested. after infection from all CB6F2s and IFNg levels were RNA from four replicate brains was pooled and 2 mgof measured using a commercial enzyme-linked immuno- the pooled sample was reverse transcribed using oligo sorbent assay kit (eBiosciences, San Diego, CA, USA). d(T) and MMLV RT (Invitrogen). PCR amplification of target genes was carried out using Taq DNA polymerase Evan’s Blue dye extravasation assay (Invitrogen, Burlington, ON, Canada) with 5 min dena- To examine BBB permeability during infection in vivo,we turing (94 1C) and 24 cycles amplification (30 s denatur- injected control and PbA-infected C57BL/6 and BALB/c ing at 94 1C, 30 s annealing at 55 1C and 30 s extension at mice (at day 8 after infection) intraperitoneally with 72 1C) with a final 7 min extension. The RT-PCR primers 0.2 ml of 1% Evan’s Blue dye (E2129; Sigma-Aldrich, used were as follows: Ifit1,50-CAGAAGCACACATT Oakville, ON, Canada) in phosphate-buffered saline GAAGAA and 30-TGTAAGTAGCCAGAGGAAGG; Ifit2, 0 0 (PBS). Ninety minutes later, mice were killed by CO2 5 -GGGAAAGCAGAGGAAATCAA and 3 -TGAAAGT inhalation, perfused with PBS and brains were harvested TGCCATACAGAAG; and Ifit3,50-GCCGTTACAGGG and photographed. AAATACTGG and 30-CCTCAACATCGGGGCTCT. Relative gene expression was quantified using Genotyping Quantity One software following overnight hybridiza- All 257 CB6F2 mice were genotyped at the McGill tion of a [a-32P]dATP-labeled probe specific for each University and Genome Quebec Innovation Centre amplicon as described by Marquis et al.41

Genes and Immunity Berr5 in cerebral malaria resistance J Berghout et al 317 Conflict of interest 17 Nagayasu E, Nagakura K, Akaki M, Tamiya G, Makino S, Nakano Y et al. Association of a determinant on mouse The authors declare no conflict of interest. chromosome 18 with experimental severe Plasmodium berghei malaria. Infect Immun 2002; 70: 512–516. 18 Bagot S, Campino S, Penha-Goncalves C, Pied S, Cazenave PA, Acknowledgements Holmberg D. Identification of two cerebral malaria resistance loci using an inbred wild-derived mouse strain. Proc Natl Acad Work in PG’s lab was supported by a Team Grant (CTP Sci USA 2002; 99: 9919–9923. 79842) from the Canadian Institutes of Health Research 19 Campino S, Bagot S, Bergman ML, Almeida P, Sepulveda N, (CIHR) to K Kain (U Toronto) and PG (McGill Uni- Pied S et al. Genetic control of parasite clearance leads to versity), and by a CIHR operating Grant (FRN79343) to resistance to Plasmodium berghei ANKA infection and confers immunity. Genes Immun 2005; 6: 416–421. MM Stevenson, and to PG. PG is a James McGill 20 Ohno T, Nishimura M. Detection of a new cerebral malaria Professor of Biochemistry, and JB is supported by a susceptibility locus, using CBA mice. Immunogenetics 2004; 56: Doctoral Award from CIHR. 675–678. 21 Penet MF, Viola A, Confort-Gouny S, Le Fur Y, Duhamel G, Kober F et al. Imaging experimental cerebral malaria in vivo: significant role of ischemic brain edema. J Neurosci 2005; 25: References 7352–7358. 22 Boyartchuk VL, Broman KW, Mosher RE, D’Orazio SE, 1 WHO. Malaria. In: World Health Organization 2007. Starnbach MN, Dietrich WF. Multigenic control of Listeria 2 Lamb TJ, Brown DE, Potocnik AJ, Langhorne J. Insights into monocytogenes susceptibility in mice. Nat Genet 2001; 27: the immunopathogenesis of malaria using mouse models. 259–260. Expert Rev Mol Med 2006; 8: 1–22. 23 Haldar K, Murphy SC, Milner DA, Taylor TE. Malaria: 3 Coltel N, Combes V, Hunt NH, Grau GE. Cerebral malaria—a mechanisms of erythrocytic infection and pathological corre- neurovascular pathology with many riddles still to be solved. Annu Rev Pathol 2 Curr Neurovasc Res 2004; 1: 91–110. lates of severe disease. 2007; : 217–249. 4 Nyangoto EO. Cell-mediated effector molecules and compli- 24 Hunt NH, Grau GE. Cytokines: accelerators and brakes in the 24 cated malaria. Int Arch Allergy Immunol 2005; 137: 326–342. pathogenesis of cerebral malaria. Trends Immunol 2003; : 5 Kwiatkowski DP. How malaria has affected the human 491–499. genome and what human genetics can teach us about malaria. 25 Clark TG, Diakite M, Auburn S, Campino S, Fry AE, Green A Am J Hum Genet 2005; 77: 171–192. et al. Tumor necrosis factor and lymphotoxin-alpha poly- 6 Mazier D, Nitcheu J, Idrissa-Boubou M. Cerebral malaria and morphisms and severe malaria in African populations. J Infect immunogenetics. Parasite Immunol 2000; 22: 613–623. Dis 2009; 199: 569–575. 7 Sexton AC, Good RT, Hansen DS, D’Ombrain MC, Bucking- 26 Campanella GS, Tager AM, El Khoury JK, Thomas SY, ham L, Simpson K et al. Transcriptional profiling reveals Abrazinski TA, Manice LA et al. Chemokine receptor CXCR3 suppressed erythropoiesis, up-regulated glycolysis, and inter- and its ligands CXCL9 and CXCL10 are required for the feron-associated responses in murine malaria. J Infect Dis 2004; development of murine cerebral malaria. Proc Natl Acad Sci 189: 1245–1256. USA 2008; 105: 4814–4819. 8 Delahaye NF, Coltel N, Puthier D, Flori L, Houlgatte R, Iraqi 27 Bluyssen HA, Vlietstra RJ, Faber PW, Smit EM, Hagemeijer A, FA et al. Gene-expression profiling discriminates between Trapman J. Structure, chromosome localization, and regula- cerebral malaria (CM)-susceptible mice and CM-resistant tion of expression of the interferon-regulated mouse Ifi54/ mice. J Infect Dis 2006; 193: 312–321. Ifi56 gene family. Genomics 1994; 24: 137–148. 9 Lovegrove FE, Gharib SA, Patel SN, Hawkes CA, Kain KC, 28 Guo J, Peters KL, Sen GC. Induction of the human protein P56 Liles WC. Expression microarray analysis implicates apopto- by interferon, double-stranded RNA, or virus infection. sis and interferon-responsive mechanisms in susceptibility Virology 2000; 267: 209–219. to experimental cerebral malaria. Am J Pathol 2007; 171: 29 Fensterl V, White CL, Yamashita M, Sen GC. Novel character- 1894–1903. istics of the function and induction of murine p56 family 10 Hansen AM, Ball HJ, Mitchell AJ, Miu J, Takikawa O, Hunt proteins. J Virol 2008; 82: 11045–11053. NH. Increased expression of indoleamine 2,3-dioxygenase in 30 Berchtold S, Manncke B, Klenk J, Geisel J, Autenrieth IB, Bohn murine malaria infection is predominantly localised to the E. Forced IFIT-2 expression represses LPS induced TNF-alpha vascular endothelium. Int J Parasitol 2004; 34: 1309–1319. expression at posttranscriptional levels. BMC Immunol 2008; 11 Senaldi G, Shaklee CL, Guo J, Martin L, Boone T, Mak TW et al. 9: 75. Protection against the mortality associated with disease 31 Strasser A, Jost PJ, Nagata S. The many roles of FAS receptor models mediated by TNF and IFN-gamma in mice lacking signaling in the immune system. Immunity 2009; 30: 180–192. IFN regulatory factor-1. J Immunol 1999; 163: 6820–6826. 32 Ohno T, Kobayashi F, Nishimura M. Fas has a role in cerebral 12 Engwerda CR, Mynott TL, Sawhney S, De Souza JB, Bickle malaria, but not in proliferation or exclusion of the murine QD, Kaye PM. Locally up-regulated lymphotoxin alpha, not parasite in mice. Immunogenetics 2005; 57: 293–296. systemic tumor necrosis factor alpha, is the principle mediator 33 Villa-Morales M, Santos J, Fernandez-Piqueras J. Functional of murine cerebral malaria. J Exp Med 2002; 195: 1371–1377. Fas (Cd95/Apo-1) promoter polymorphisms in inbred mouse 13 Patel SN, Berghout J, Lovegrove FE, Ayi K, Conroy A, strains exhibiting different susceptibility to gamma-radiation- Serghides L et al. C5 deficiency and C5a or C5aR blockade induced thymic lymphoma. Oncogene 2006; 25: 2022–2029. protects against cerebral malaria. J Exp Med 2008; 205: 34 Mitsos LM, Cardon LR, Ryan L, LaCourse R, North RJ, Gros P. 1133–1143. Susceptibility to tuberculosis: a locus on mouse chromosome 14 Rogerson SJ, Grau GE, Hunt NH. The microcirculation in 19 (Trl-4) regulates Mycobacterium tuberculosis replication in severe malaria. Microcirculation 2004; 11: 559–576. the lungs. Proc Natl Acad Sci USA 2003; 100: 6610–6615. 15 Pais TF, Chatterjee S. Brain macrophage activation in murine 35 Choi P, Xanthaki D, Rose SJ, Haywood M, Reiser H, Morley BJ. cerebral malaria precedes accumulation of leukocytes and Linkage analysis of the genetic determinants of T-cell IL-4 CD8+ T cell proliferation. J Neuroimmunol 2005; 163: 73–83. secretion, and identification of Flj20274 as a putative candi- 16 Renia L, Potter SM, Mauduit M, Rosa DS, Kayibanda M, date gene. Genes Immun 2005; 6: 290–297. Deschemin JC et al. Pathogenic T cells in cerebral malaria. Int J 36 Yanez DM, Manning DD, Cooley AJ, Weidanz WP, van der Parasitol 2006; 36: 547–554. Heyde HC. Participation of lymphocyte subpopulations in the

Genes and Immunity Berr5 in cerebral malaria resistance J Berghout et al 318 pathogenesis of experimental murine cerebral malaria. 39 Fortin A, Diez E, Rochefort D, Laroche L, Malo D, Rouleau GA J Immunol 1996; 157: 1620–1624. et al. Recombinant congenic strains derived from A/J and 37 Yoshimoto T, Takahama Y, Wang CR, Yoneto T, Waki S, C57BL/6J: a tool for genetic dissection of complex traits. Nariuchi H. A pathogenic role of IL-12 in blood-stage murine Genomics 2001; 74: 21–35. malaria lethal strain Plasmodium berghei NK65 infection. 40 Broman KW. Mapping quantitative trait loci in the case J Immunol 1998; 160: 5500–5505. of a spike in the phenotype distribution. Genetics 2003; 163: 38 Butterfield RJ, Blankenhorn EP, Roper RJ, Zachary JF, Doerge 1169–1175. RW, Teuscher C. Identification of genetic loci controlling the 41 Marquis JF, Nantel A, LaCourse R, Ryan L, North RJ, Gros P. characteristics and severity of brain and spinal cord lesions in Fibrotic response as a distinguishing feature of resistance and experimental allergic encephalomyelitis. Am J Pathol 2000; 157: susceptibility to pulmonary infection with Mycobacterium 637–645. tuberculosis in mice. Infect Immun 2008; 76: 78–88.

Supplementary Information accompanies the paper on Genes and Immunity website (http://www.nature.com/gene)

Genes and Immunity