The JI
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     
 

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Collette, A.
Right arrow Articles by Pied, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Collette, A.
Right arrow Articles by Pied, S.
The Journal of Immunology, 2004, 173: 4568-4575.
Copyright © 2004 by The American Association of Immunologists

A Profound Alteration of Blood TCRB Repertoire Allows Prediction of Cerebral Malaria1

Alexis Collette2,*,{dagger}, Sébastien Bagot*,{dagger}, Maria E. Ferrandiz*, Pierre-André Cazenave*, Adrien Six3,* and Sylviane Pied*,{dagger}

* Immunophysiopathologie Infectieuse, Centre National de la Recherche Scientifique Unité de Recherche Associée 1961, Institut Pasteur, and Université Pierre et Marie Curie, and {dagger} Institut National de la Santé et de la Recherche Médicale Unité 511, Centre Hospitalier Universitaire Pitié-Salpêtrière, Paris, France


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Cerebral malaria (CM) is one of the severe complications of Plasmodium infection. In murine models of CM, T{alpha}{beta} cells have been implicated in the neuropathogenesis. To obtain insights into the TCRB repertoire during CM, we used high throughput CDR3 spectratyping and set up new methods and software tools to analyze data. We compared PBL and spleen repertoires of mice infected with Plasmodium berghei ANKA that developed CM (CM+) or not (CM–) to evidence modifications of the TCRB repertoire associated with neuropathology. Using distinct statistical multivariate methods, the PBL repertoires of CM+ mice were found to be specifically altered. This alteration is partly due to recurrently expanded T cell clones. Strikingly, alteration of the PBL repertoire can be used to distinguish between CM+ and CM–. This study provides the first ex vivo demonstration of modifications of T{alpha}{beta} cell compartment during CM. Finally, our original approach for deciphering lymphocyte repertoires can be transposed to various pathological conditions.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Malaria is more than ever a health problem. A total of 300–500 million people are infected each year. Approximately two million, mostly African children, die of complications of infection by Plasmodium falciparum, particularly cerebral malaria (CM).4 This situation could soon worsen as resistance to chemotherapies are spreading and mosquito resistance to insecticides has hampered programs to diminish vector prevalence (1).

Pathogenesis of CM is a complex process involving both host and parasite factors. Several cytokines and mediators have also been associated with CM (2). However, no mechanism has been clearly established. Several lines of evidence implicate T lymphocytes in the development of CM. In humans, a class I MHC allele as well as a class II MHC haplotype have been associated with protection from severe forms of malaria (3). In mouse CM models that mimick only partially the neuropathology in humans (4), the immune system plays an active deleterious role, as evidenced by the protective effect of immunosuppressive treatments (5, 6, 7). T{alpha}{beta} cells, both CD4+ and CD8+ subpopulations, are necessary for the development of experimental CM as shown in nude, knockout, and T cell-depleted mice (6, 8, 9, 10, 11, 12). T{gamma}{delta} cells could also play a role in the neuropathogenesis because depletion by anti-C{delta} Ab prevents CM (11). However, they are not required because C{delta} knockout mice still develop CM (11, 12).

Balance between deleterious and protective roles of immune responses is a common feature in many infectious diseases in which pathogens manipulate the host immune system (13, 14, 15, 16). In such complex situations, deciphering lymphocyte subpopulation implication in pathology vs protection requires the exhaustive description of lymphocyte populations in clinical trials involving numerous individuals. Lymphocyte responses can be described by their Ag-specific receptor diversity, which is produced by somatic DNA rearrangements of V, (D), and J segments later spliced to C segments (17). The product of the V(D)J joining, called the CDR3, is in contact with the Ag. This region is imprecise in the number and the nature of nucleotides that are removed or added, and is therefore variable in amino acid length and composition. The CDR3 spectratyping method precisely describes diversity of a lymphocyte population repertoire by the analysis of the CDR3 length use (18, 19).

In this study, we set up a new strategy associating large scale CDR3 spectratyping, original software tools, and multifactorial statistical analyses that enabled us to decipher the TCRB repertoire during infection by Plasmodium berghei Anvers/Kasapa (ANKA) and distinguish between pathogenic and protective repertoires. We show that the PBL repertoire of CM+ mice is altered and identify potentially pathogenic recurrent clones expanded in CM+ mice, but not in infected mice that do not develop CM (CM–).


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Mice and parasites

Eight-week-old B10.D2 mice were purchased from Harlan (Oxon, U.K.). The clone 1.49L of P. berghei ANKA (20) was kindly given by D. Walliker (Institute of Genetics, Edinburg, U.K.) and is maintained in our laboratory on C57BL/6J female mice. This clone induces in mice a neurological syndrome partly mimicking the one of human CM. Erythrocytic stages of the parasite were cryopreserved in liquid nitrogen as stabilates in Alserver’s solution containing 10% glycerol. Infection was induced by i.p. injection of 106 parasitized RBC. Between days 7 and 10 after infection, 90% of the mice develop CM characterized by ataxia, paralysis, deviation of the head, and convulsions, followed by deep coma and death. These mice constituted the CM+ group. CM– mice were sacrificed between days 11 and 16 after infection, not having shown signs of CM during the critical period and exhibiting a parasitemia above 20%.

Cell preparation

Blood was obtained on heparin by retroorbital punction. Mononuclear cells were isolated on Ficoll-Hypaque gradient (Pharmacia, Orsay, France). Spleen was removed, and cells were suspended in 3% FCS-PBS. RBC were lysed with ammonium chloride buffer for 5 min at room temperature. The brain lymphoid cells were obtained after tissue homogenization in RPMI 1640 medium and centrifugation at 500 x g for 20 min in 35% (v/v) Percoll (Pharmacia). The brain lymphoid cells of four CM+ mice were pooled, stained with PE-conjugated anti-mouse CD3{epsilon} mAb (145-2C11), and sorted by use of a MoFlo cell sorter (DakoCytomation, Fort Collins, CO, and BD Pharmingen, San Diego, CA). Dead cells were excluded using propidium iodide at a final concentration of 1 µg/ml (Sigma-Aldrich, Saint Quentin Fallavier, France).

Cell preparations were then washed twice with PBS. Lymphoid cells were counted using Malassez cell in presence of eosin to exclude dead cells.

TCRB repertoire

Total RNA was extracted from >90,000 mononuclear cells for each sample using the TRI REAGENT kit (Molecular Research Center, Cincinnati, OH). A total of 20 µg of glycogen (Roche, Meylan, France) was used to ensure optimal precipitation of RNA and pellet visualization. Protocols for TCR BV-BC and BV-BJ CDR3 spectratyping have already been explained elsewhere in detail (18, 21). BC, BV, and BJ primer sequences were as described earlier (18), except BV8.3 (TGCTGGCAACCTTCAAATAGGA) and BV13 (AGGCCTAAAGGAACTAACTCCAC). As BV5.3 (22), BV17 (23), and BV19 (24) are not functional in B10.D2 mice, they were not amplified. BV-BJ repertoires were analyzed for BV2, BV3, BV4, BV5.1, BV6, BV7, BV8.1–3, BV9, BV14, and BV16 genes. PCR products were loaded on a 36-well ABI373 automated sequencer (Applied Biosystems, Foster City, CA) and separated according to their nucleotide length, forming a profile of peaks for each primer combination, spaced by 3 nt as expected for in-frame transcripts. Each peak corresponds to a CDR3 length. The Immunoscope software (18) was used to obtain peak area and nucleotide length and CDR3 profile displays from sequencer raw data.

BV-BJ direct sequencing

Direct sequencing was performed with BV and BJ primers for peaks representing between 65 and 100% of the BV-BJ profile, following the recommendations of the ABI Prism Dye Terminator Cycle Sequencing Ready Reaction kit (Applied Biosystems). BV-BJ PCR products were first reamplified. PCR products were then incubated with 0.5 U of shrimp alkaline phosphatase (United States Biotechnology, Cleveland; OH) and 5 U of exonuclease I (United States Biotecnology) at 37°C for 40 min, followed by 20 min at 80°C. DNA alignments were performed using the GCG package (Genetics Computer Group, Accelrys, Cambridge, U.K.)

Methods and software tools for CDR3 spectratype analyses

We developed the ISEApeaks software package (©2000–2002; Institut Pasteur, France, Paris) to extract, smooth, manage, and analyze the large amount of data obtained in this study (25, 26). As the intensity of CDR3 peaks is not comparable between different amplifications, we considered the percentage of use of each CDR3 length, obtained by dividing the area of CDR3 peaks by the total area of all peaks within the profile.

Perturbation of BV-BC repertoire was compared with a control group, as explained elsewhere (27, 28). Sample perturbation (µDBV-BC) is the mean of perturbations of each BV segment (DBV-BC). We extended this perturbation index to BV-BJ repertoires by computing µµDBV-BJ, the mean of the DBV-BJ for all BV and BJ segments. BV-BC and BV-BJ perturbations range from 0, identical profiles, to 100, completely different profiles.

Recurrence of CDR3 expansions was assessed quantitatively with OligoScore (25), which scores each peak in each group of samples. The maximum score of the control group is used as a threshold for other groups. Peaks with a score above this threshold are considered recurrently expanded.

Statistics

Multivariate statistics were used to analyze repertoire data. In a first approach, each repertoire was considered as a vector in n-dimensional space, where n is the number of variables that describe the repertoires. Missing values were replaced by the overall mean of the variable (29). The number of variables (230 peaks for BV-BC repertoires) was too high for theoretical constraints of discriminant analysis (DA) and was thus reduced using principal components analysis (PCA). PCA extracts new variables from the data set that retains the variability contained in the original data set. Linear DA was performed on the new data set to compare the different groups. DA computes discriminant functions that maximize intergroup variation and minimize intragroup variation. Significance of each discriminant function was tested using {chi}2 approximation of Wilks’ statistics (29).

Univariate one-way or two-way ANOVA was used to analyze sample perturbation data (µDBV-BC or µµDBV-BJ). When significant, comparison between two categories was performed with Fisher’s protected least significant difference. One-way multiple ANOVA (MANOVA) was used to compare DBV-BJ. MANOVA F-statistics approximation of Wilk’s {lambda}, Roy’s greatest root, Pillai trace, and Hotelling-Lawley trace multivariate statistics were calculated. As the power of these four statistics are not equivalent (29), we considered MANOVA significant when all four statistics were significant. Indicated p value corresponds to the maximum value of the p value obtained for these four statistics (30). When significant, a one-way ANOVA was conducted on each variable (BV-BC perturbation) to assess for differences in the considered groups. This enables minimization of global error rate of the univariate tests (29).

To further assess similarity between samples, k-mean clustering with Euclidean distance was used. Significance of the clusters was assessed by computing the probability of observing by chance the number of samples of a particular group within the cluster (31).

PCA, DA, and k-mean clustering were performed with SYSTAT 10 (SPSS, Chicago, IL). ANOVA and MANOVA were performed with StatView 5.0 (SAS Institute, Cary, NC). Statistics were considered significant when p < 0.01.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Analysis of TCRB repertoire during CM by the Immunoscope method

The TCRB repertoire during CM induced in B10.D2 mice by P. berghei ANKA was analyzed using the exhaustive CDR3 length spectratyping approach (18).

BV-BC repertoires were studied with total cDNA, prepared from noninfected control (CTR) mice and infected mice that did (CM+) or did not (CM–) develop CM. In PBL and spleen of naive mice, CDR3 profile for each BV-BC combination is bell shaped, indicative of a diverse polyclonal repertoire (Fig. 1, and data not shown). The TCRB repertoires of the CM+ and CM– mice are profoundly altered: almost all CDR3 profiles are different from a bell-shaped profile, and multiple expansions are evidenced (Fig. 1). Among these complex repertoire modifications, some could contribute to the protective response against P. berghei ANKA infection, while others are involved in pathology. The comparison between CDR3 profiles from CM+ and CM– mice enables the identification of potentially pathogenic clones associated with CM that would be present in CM+ mice, but absent in CM– mice.



View larger version (53K):
[in this window]
[in a new window]
  		FIGURE 1. Typical PBL repertoires of CTR, CM–, and CM+ mice. cDNA were obtained from PBL of control (CTR), infected without (CM–), and infected with neurological signs of CM (CM+) mice. BV-BC CDR3 spectratyping was performed by PCR amplification of cDNA with combinations of BV-specific primers and a BC-specific primer. After runoff with a fluorescent BC-specific primer, PCR products were size separated on an automated DNA sequencer. Sequencing gels were analyzed with the Immunoscope software. Typical samples are represented. Horizontal axis represents the nucleotide size, and vertical axis the fluorescence intensity, in arbitrary units.

 
Specific alterations of TCRB repertoire during CM

To identify CM-associated alterations, we studied ex vivo the PBL and splenocyte TCRB repertoires of several individuals for CTR, CM–, and CM+ groups. A total of 1150 BV-BC CDR3 profiles was obtained for 55 samples. The analysis of this set of BV-BC profiles, each including 6–8 peaks, required the use of an original approach that combines bioinformatic tools and multivariate statistics.

All peak data of TCRB repertoire were extracted, formatted, and edited with the ISEApeaks software to construct the peak database. PCA was used to reduce the initial number of variables. The new data set retains 97% of the original information. We used linear DA to evidence global modifications due to infection in CM+ and CM– mice. DA determines whether multivariate observations of different groups are samples of the same statistical population. As shown on Fig. 2, all members of a given group are closely clustered together. Moreover, sample groups can be separated in four statistically significant clusters, as only the first three discriminant functions are significant: CM+ PBL, CM– PBL, CM– spleen, and a group containing CM+ spleen, CTR PBL, and CTR spleen.



View larger version (51K):
[in this window]
[in a new window]
  		FIGURE 2. DA separates CM+ PBL from CM+ spleen, CM– PBL, and CM– spleen BV-BC repertoires. DA was performed on the new set of variables obtained with PCA. Bivariate graphics represent the value of two discriminant functions. Discriminant functions are linear combination of the variables that try to separate groups. F1 to F5 stands for each discriminant function, sorted decreasingly by the associated eigenvalue. Samples (8 CTR PBL, 6 CTR spleen, 10 CM– PBL, 8 CM– spleen, 13 CM+ PBL, and 10 CM+ spleen) are represented on the bivariate plots, but are scarcely visible because DA groups samples very well. Confidence (0.95) ellipses, centered on group centroid, are overlaid. The density normal curve of each group is shown on the diagonal. Vertical and horizontal axes represent canonical scores for each function. The density normal curves of F4 and F5 show that groups are not well separated, consistent with the nonsignificance of these functions. All CDR3 profiles of this study can be obtained on the supplement web page (www.pasteur.fr/recherche/unites/Imchan/addit2.html).

 
To further characterize the repertoires, we used another method that computes CDR3 spectratype perturbation index for each sample (27). All repertoires were compared with the spleen CTR repertoires to obtain a perturbation index for each BV-BC combination (DBV-BC). µDBV-BC, the average of BV-BC perturbations, were compared by two-way ANOVA between compartments (PBL and spleen) and the different groups of mice (CTR, CM–, and CM+) (Fig. 3A). ANOVA compares the effect of qualitative factors on a quantitative dependent variable. Infection by P. berghei ANKA leads to significant alteration of the TCRB perturbation both due to the groups of mice (p < 0.0001) and to the lymphoid compartment (p = 0.0002). PBL repertoires of CM+ mice (mean = 20.2) were more perturbed than in CM– mice (mean = 15.2; t test, df = 21, p < 0.0001). To strengthen this observation, we used k-mean clustering to see whether groups could be separated on the basis of BV-BC perturbations without knowledge of group composition, which was the case of DA and ANOVA. For k = 3 (or k = 4), all 13 CM+ PBL samples clustered together in a group containing 17 (or 16) samples that had the probability p = 1.64E-9 (p = 3.86E-10) to occur by chance (Fig. 3B). Because PBL BV-BC repertoires appear to be specifically altered in CM+ as compared with CM– mice, further analyses were performed on PBL only.



View larger version (22K):
[in this window]
[in a new window]
  		FIGURE 3. PBL CM+ repertoire are significantly more perturbed than CM– PBL or spleen repertoires, and can be clustered separately. A, BV-BC perturbations (DBV-BC) were computed with ISEApeaks using the CTR spleen group as control. Mean sample DBV-BC (µDBV-BC) with their SE are shown for CM+, CM–, and CTR mice in the PBL and spleen. DBV-BC range from 0, identical with the reference repertoire, to 100, completely perturbed. B, Schematic representation of sample clusters obtained with k-mean clustering on DBV-BC with k = 4 and 3. BV-BC perturbation of each sample was analyzed without prior knowledge of group composition. For k = 5 or 6, PBL CM+ samples were split in two clusters. C, The BV-BJ repertoires of PBL samples (6 CTR, 5 CM–, and 6 CM+) were correctly grouped by k-mean clustering with k = 3 on DBV-BJ data.

 
BV-BC perturbation of PBL repertoires allows prediction of CM

We investigated whether the alteration of PBL BV-BC repertoire could enable us to classify samples. Because the sample number is not large enough to divide it in training and testing sets, we used the jackknife method: each sample is left out in turn and DA is performed on the remaining samples to obtain classification functions. These functions are then used to determine to which group the left-out sample belongs. Table I shows that 85% of CM+ PBL samples were correctly classified, indicating that perturbation of PBL repertoire can be used to discriminate between CM+ and CM– mice.


View this table:
[in this window]
[in a new window]
  		Table I. Perturbation of BV-BC repertoires allows prediction of CMa

 
BV-BJ PBL repertoire analysis during infection

To describe more precisely the TCRB repertoire during CM, we performed BV-BJ repertoire analysis for 9 BV genes of 21, which represent more than two-thirds of the total TCRB repertoire. A total number of 2300 BV-BJ profiles was generated.

To compare perturbation results obtained from BV-BC and BV-BJ repertoire analysis, we first computed the average DBV-BJ for all BV and BJ genes for each sample (µµDBV-BJ). The three PBL groups (CTR, CM–, and CM+) were significantly different (ANOVA, p < 0.0001). Again, CM+ PBL repertoires (mean = 32.3) were more altered than those of CM– PBL (mean = 24.5; p = 0.0035). By using k-mean clustering with DBV-BJ data, all three groups could also be reconstituted without prior knowledge of group composition (Fig. 3C). Five of six CM+ PBL samples (p = 0.0001) clustered together apart from CTR and CM– samples.

To determine which BV gene(s) contributed to this difference, we computed the perturbation of BV genes on the basis of BV-BJ repertoires (µDBV-BJ perturbation). MANOVA analysis showed that the three groups are statistically different (p = 0.01). However, only µBV8.1-BJ (ANOVA, p = 0.01) was significantly more altered in CM+ mice compared with CM– mice. Finally, analysis of the contribution of BJ segments in BV8.1-positive T cells by MANOVA, followed by ANOVA, implicated only DBV8.1-BJ1.1, DBV8.1-BJ1.6, DBV8.1-BJ2.1, and DBV8.1-BJ2.2.

ISEApeaks was used to compute BJ percentages for each BV gene in PBL samples. These percentages were compared by MANOVA. BJ use was not significantly different among the CTR, CM–, and CM+ groups (data not shown). Thus, perturbations of BV-BJ CDR3 spectratypes are not correlated with modifications of the BJ use.

Identification of pathogenic recurrent clones

We searched the PBL BV-BC and BV-BJ repertoire data for pathogenic T cell clones associated with CM. We concentrated on the identification of pathogenic clones that are recurrently present in CM+, but not in CM– mice. We used OligoScore, which scores peaks for their recurrence in each group (25). For BV-BC repertoires, no peak in CM– or CM+ groups has a score higher than the threshold of CTR groups (data not shown). Hence, no BV-BC peak is found recurrently expanded.

We used the same scoring approach to identify recurrent clones in BV-BJ repertoires. A total of 122 peaks in the PBL CM+ mice has a score above the corresponding threshold defined with the CTR PBL group. They were compared with those of the CM– group by subtracting the corresponding peak scores. Peaks were sorted decreasingly to identify the most recurrent in CM+, but absent in CM–. The three most differentially expressed recurrent peaks belong to BV8.1-BJ1.5 and BV8.1-BJ2.2 profiles (Table II and Fig. 4). BV2-BJ1.3 peak is given as an example of a peak that is recurrently expanded both in CM+ and CM– mice and thus low ranked in Table II (this point will be discussed later). BV8.1 peaks were directly sequenced. For each of the two BV8.1-BJ combinations, similar amino acid sequences were found in several CM+ individuals (Table III). On the contrary, direct sequencing for the PCR products in CM– samples yielded no readable CDR3 sequence (data not shown).


View this table:
[in this window]
[in a new window]
  		Table II. Identification of recurrent expansions in CM+ mice in BV-BJ CDR3 profiles by OligoScorea

 


View larger version (40K):
[in this window]
[in a new window]
  		FIGURE 4. BV-BJ CDR3 profiles of recurrent expansions identified by OligoScore and perturbation analysis. PBL cDNA were amplified with BV8.1- or BV2-specific primers with a BC-specific primer. PCR products were then subjected to runoff with appropriate BJ-specific primers. Products were separated on an automated sequencer and analyzed with Immunoscope and ISEApeaks. All PBL samples for which those combinations were analyzed are represented. Horizontal axis represents the nucleotide size, and vertical axis the fluorescence intensity, in arbitrary units.

 

View this table:
[in this window]
[in a new window]
  		Table III. CDR3 sequences of BV-BJ expansions in CM+ micea

 
Finally, we have explored the representation of the identified BV-BJ combination in the brain of CM+ mice. Due to cell number limitation and to obtain a global view of BV-BJ repertoire in the brain, we have pooled PBL, spleen, and brain samples from four CM+ mice. BV08.1-BJ1.5, BV08.1-BJ2.2, and BV02-BJ1.3 were analyzed together with BV08.1-BJ2.3 and BV02-BJ2.3 as controls (Fig. 5). It must be noted that oligoclonal expansion in a pool is strongly suggestive of a public repertoire.



View larger version (22K):
[in this window]
[in a new window]
  		FIGURE 5. BV-BJ CDR3 profiles of recurrent expansions identified in brain, PBL, and spleen. Brain, PBL, and spleen cDNA were prepared from pooled lymphocytes isolated from four CM+ mice. cDNA were amplified with BV8.1- or BV2-specific primers with a BC-specific primer. PCR products were then subjected to runoff with appropriate BJ-specific primers. Products were separated on an automated sequencer and analyzed with Immunoscope and ISEApeaks. BV02-BJ1.3, BV08.1-BJ1.5, and BV08.1-BJ2.2 are BV-BJ combinations that were found recurrently expanded in PBL of individual CM+ mice. BV02-BJ2.3 and BV08.1-BJ2.3 combinations serve as controls. Horizontal axis represents the nucleotide size, and vertical axis the fluorescence intensity, in arbitrary units.

 
The BV02-BJ1.3 expansion, common to CM+ and CM– mice, was again identified in all three organs (>70.0%). The BV08.1-BJ1.5 combination was also noted in the brain compartment (30.4%), with a slight increase compared with PBL (25.2%) and spleen (25.7%). The two BV08.1-BJ2.2 expansions were also identified in the brain (65.0%), to a similar extent to the PBL (65.7%) and spleen (62.9%) compartments. The BV08.1-BJ2.2 and BV08.1-BJ1.5 clones, identified as differentially expanded in CM+ mice by comparison with CM– mice, are thus also present in the brain, the site of pathogenesis, and could thus mediate the neuropathology of CM.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
The aim of the present study was to exhaustively characterize the TCRB repertoire during CM in B10.D2 mice infected with P. berghei ANKA clone 1.49L. To this end, we devised new global methods and tools, based on a large-scale use of the CDR3 spectatyping approach. Our results show that the TCRB repertoire is specifically altered in the PBL of CM+ mice as compared with PBL of CM– mice, whereas no difference was evidenced in the spleen. This perturbation of the TCRB repertoire is partly explained by recurrent clones that are present in CM+ and absent in CM– mice. Preliminary data indicate that these recurrent clones are also found in the brain of CM+ mice.

Analysis of the repertoire using the CDR3 spectratyping describes the entire ex vivo TCRB repertoire of a sample by up to 2400 measurements. As a high number of parameters are measured, it must be ensured that the cell quantity used is sufficient. In particular, paucity of material tends to favor stochastic PCR amplifications. We therefore carefully determined by repertoire analysis of a set of cell dilutions that the cell numbers we used were sufficient to guarantee the quality of our repertoire data (A. Six and O. Gorgette, unpublished data). Furthermore, the identification of recurrent BV-BJ CDR3 expansions shows that the repertoire modifications documented in this study are not artifacts (Table III).

The original bioinformatic tools we devised enabled us to analyze the 3450 CDR3 profiles generated in this study. All three independent multivariate statistics consistently gather in a similar manner the six experimental groups into three to four clusters. Moreover, as expected, the CTR PBL and CTR spleen groups are not separated for DA, ANOVA, and k-mean clustering (with k = 3). We demonstrated for the first time that T cell repertoire data can give diagnostic/prognostic information when analyzed by class prediction. The alteration of the BV-BC repertoire enabled the group classification of 85% of the PBL samples of CM+ mice. PBL samples of CM– mice are less correctly classified (50%). However, only 1 of 10 CM– PBL samples is erroneously classified as a CM+ PBL, indicating that the risk to predict falsely that an infected individual is developing CM is small.

An original quantitative scoring method, OligoScore (25), was used to identify recurrent expansion of T cell clones among the 1040 BV-BJ CDR3 peaks. It should be noted that existence of perturbation in a particular BV-BC or BV-BJ profile does not imply existence of recurrent expansion within this profile because private expansions can also distort it. Surprisingly, no recurrent peaks were found at the level of BV-BC. Two explanations can be given. Recurrent peaks in the BV-BC repertoires might be below the detection limit of our scoring method. This is unlikely because OligoScore enabled detection of recurrence that were not visible by eye, even with a small number of samples (25). More likely, it might be the consequence of a buffer effect between BV-BC and BV-BJ data (32). Variation at the more precise level of BV-BJ repertoires can be averaged in the corresponding BV-BC repertoires because these repertoires are the addition of all BV-BJ repertoires. An example of this buffer effect can be visualized on Fig. 4 for BV8.1+ cells, in which modifications seen at the BV8.1–BJ1.5 and BV8.1–BJ2.2 levels are smoothed at the BV8.1-BC level. This effect is also seen on PBL sample perturbation of CM+ mice, which is 20.2 when estimated on BV-BC repertoires and 32.3 on BV-BJ repertoires. It is because BV-BJ repertoires are 12 times more precise than BV-BC and therefore give a more accurate description of the repertoire. Finally, as BV-BJ repertoires are quantitative inside a given BV gene (33, 34), it is possible to test this buffer effect by calculus. We were indeed able to reconstruct BV-BC profiles with our BV-BJ data (data not shown).

The implication of T{alpha}{beta} cells in the neuropathogenesis of malaria has been demonstrated with depletion by Abs (6, 9, 10, 11) and use of nude (8) or knockout mice (11, 12). Our present results add to these previous studies by demonstrating, ex vivo, a profound perturbation and the existence of recurrent CDR3 peaks in TCRB PBL repertoires during CM, despite the numerous P. berghei molecules that stimulate the immune system during infection. By contrast with recurrent responses against single Ags (35, 36), the recurrent response against P. berghei is modest because it was not found at the BV-BC level. This could be related to a general activation of T cells, possibly due to Plasmodium mitogens (37, 38, 39) that prevent the expansion of Ag-specific clones. The stability of BJ use observed between groups is consistent with this observation (data not shown).

Modification of the TCRB repertoire is evidenced only in the PBL of CM+ mice in contrast with the usually well-accepted idea that PBL reflects spleen. This is also in contrast with spleen being necessary for the occurrence of CM (6, 40, 41). Absence of specific alteration in the spleen of CM+ mice could be explained by the dilution of stimulated cells in the bulk of T cells present in this organ and the fact that they leave to recirculate upon activation (42).

T cell clones associated with neuropathogenesis can be of different types. First, they can be either private, specific to one individual, or recurrent, present in different individuals and sometimes designated as public clones. Second, their function might be pathogenic, protective, or regulatory. Based on our BV-BJ repertoire analysis of PBL, we have identified in this study recurrent clones associated to neuropathogenesis by assessing their presence in CM+ mice and absence in CM– mice. It is of interest to investigate the relation of these expansions with neuropathology, in particular to determine whether these expansions are also found in the brain. We and others have recently observed the accumulation of CD8+ T{alpha}{beta} lymphocytes in the brain of CM+ mice (43, 44). However, due to the small number of infiltrating lymphocytes in the brain, even when pathology occurs (43, 44), it was difficult to investigate this point at the same time in individual mice. These experiments are underway in the laboratory. In our preliminary experiment on brain lymphocytes pooled from four mice presented in this study, we show that these expanded BV-BJ combinations are also found in the brain of CM+ mice.

Five of the 10 most recurrent and differentially expressed peaks use the BV8.1 segment. Observations that support their having a pathogenic role is that depletion of BV8.1/2+ cells diminishes the incidence of CM from 90 to 40% (12). Others peaks, including the ones using BV6, BV7, and BV9 segments, have been identified in CM+ mice (Table II). This relates to the absence of CM in mice treated with a superantigen that deletes BV6, 7, 8.1, and 9 using T cells (12, 45).

We did not identify any CM–-specific recurrent peak (data not shown), which could be involved in protection, but a BV2-BJ1.3 peak is recurrently present both in CM+ and CM– mice, as judged by the high score obtained for the two groups (Table II and Fig. 4). This peak is by far the most expanded peak in CM– mice (data not shown). Mechanisms contributing to neuropathogenesis could thus be the result of a regulatory pathway between BV2- and BV8.1-expanded clones, leading to alteration of their cytokine profiles (46).

Altogether, these results suggest that few T cell clones are implicated in the development of CM. The immunological history for each individual shapes the emergent repertoire differentially between inbred individuals (47). These variations could explain why, among a group of genetically identical mice infected with the same stabilate of parasites, only some develop CM.

In a previous study, we had observed by flow cytometry an increase of BV8.1/2+ T cells in the PBL of CM+ mice, while no expansion was seen in the spleen of CM+ mice nor in the PBL and spleen of CM– mice (12) (data not shown). This increase, if caused by the expansion of few T cell clones, should distort the bell-shaped CDR3 length distribution of BV8.1/2-BC profiles. However, the PBL BV8.1/2-BC repertoire of CM+ mice is not perturbed by comparison with CTR mice (ANOVA, p = 0.71) or CM– mice (ANOVA, p = 0.29). In addition, no modification of the BJ segment use was observed in the CM+ mice. The increase of the representation of BV8.1/2 cells in the PBL of CM+ mice thus cannot be attributed to a mono- or oligoclonal increase, and is therefore polyclonal. BV8.1/2 could be stimulated by a superantigen-like molecule, as observed in Plasmodium yoelii infection (48). Implications of such molecules in pathogenesis have been reported for infections by Toxoplasma gondii (49) and Leishmania infantum (A. Sassi, B. Largueche-Darwaz, A. Collette, A. Six, D. Laouini, P. A. Cazenave, and K. Dellagi, unpublished results).

Plasmodium parasites being constituted of a very complex pool of molecules, it can comprise different types of activities, such as mitogens, superantigens, and conventional Ags that together shape the immune response qualitatively and quantitatively. In our P. berghei ANKA-infected B10.D2 mice, one can hypothesize that among expanded BV8.1/2 T lymphocytes, Ag-specific T cell clones could have a pathogenic function. This is under investigation in our laboratory by determining their function and fine specificity.

The original method presented in this work allows the exhaustive analysis of immune repertoires. Applied to a mouse model of malaria, it demonstrates that the neuropathology induced by P. berghei ANKA is associated with a global perturbation of TCRB repertoires specifically found in PBL together with the recurrent expansion of few T cell clones.

It will be of interest to determine whether the CM+ perturbed repertoire is responsible for neuropathology. We have recently described two CM resistance loci in a wild-derived inbred strain mice (50), and are currently in the process of obtaining a double congenic B6 strain resistant to CM, but still susceptible to infection. This model will be best suited to test whether an adoptive transfer of CM+ T lymphocytes from B6 mice can alter the course of infection and shift pathology to CM in genetically CM-resistant mice.

The repertoire analysis method reported in this work can easily be transposed to human malaria because PBL are easily accessible to experiment. It is intriguing to know whether the same association between PBL perturbation and neuropathology can be found in P. falciparum malaria. Furthermore, classification experiments allowed separation of the CM+ and CM– mice, and thus provide new tools for a better understanding of the immune response during malaria in humans. We are currently testing our hypotheses by studying cohorts of malaria patients. Finally, our original approach for deciphering lymphocyte repertoires can be transposed to various pathological conditions. For instance, this methodology is used in clinical follow-ups of patients after bone marrow transplantation or vaccination. The results and approach we present provide a promising basis for the development of innovative investigation strategies in the field of immunology.


   Acknowledgments
 
We thank D. Rueff-Juy, P. Boudinot, E. Rocha, and J. Kanellopoulos for critical reading of the manuscript; G. Milon, C. Pannetier, and C. Fesel for fruitful discussions; and M. Idrissa-Boubou and O. Gorgette for their technical help.


   Footnotes
 
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1 A.C., S.B., and M.E.F. were recipients of a fellowship from the Ministère de la Recherche and Université Pierre et Marie Curie. S.B. also received a fellowship from the Fondation pour la Recherche Médicale. This work was supported by Grant PRFMMIP from the Ministère de la Recherche and a grant from the European Community, INCO-DC IC18CT980363. Back

2 Current address: Texcell SA, campus de l’Institut Pasteur, 25–28, rue du Dr. Roux, 75015 Paris, France. Back

3 Address correspondence and reprint requests to Dr. Adrien Six, Unité d’Immunophysiopathologie Infectieuse, Institut Pasteur, 25, rue du Dr Roux, 75015 Paris, France. E-mail address: adrien.six{at}pasteur.fr Back

4 Abbreviations used in this paper: CM, cerebral malaria; ANKA, Anvers/Kasapa; CTR, control; DA, discriminant analysis; MANOVA, multiple ANOVA; PCA, principal components analysis. Back

Received for publication August 23, 2002. Accepted for publication July 27, 2004.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. World Health Organization. 2000. Severe falciparum malaria. Trans. R. Soc. Trop. Med. Hyg. 94:S1.
  2. Plebanski, M., A. V. Hill. 2000. The immunology of malaria infection. Curr. Opin. Immunol. 12:437.[Medline]
  3. Hill, A. V. S., C. E. M. Allsopp, D. K. Kiatkowsky, N. M. Anstey, P. Twumasi, P. A. Rowe, S. Benettt, D. Brewster, A. J. Mc Michael, B. M. Greenwood. 1991. Common West African HLA antigens are associated with protection from severe malaria. Nature 352:595.[Medline]
  4. Lou, J., R. Lucas, G. E. Grau. 2001. Pathogenesis of cerebral malaria: recent experimental data and possible applications for humans. Clin. Microbiol. Rev. 14:810.[Abstract/Free Full Text]
  5. Grau, G. E., D. Gretener, P. H. Lambert. 1987. Prevention of murine cerebral malaria by low-dose cyclosporin A. Immunology 61:521.[Medline]
  6. Curfs, J. H., T. P. Schetters, C. C. Hermsen, C. R. Jerusalem, A. A. van Zon, W. M. Eling. 1989. Immunological aspects of cerebral lesions in murine malaria. Clin. Exp. Immunol. 75:136.[Medline]
  7. Eckwalanga, M., M. Marussig, M. D. Tavares, J. C. Bouanga, E. Hulier, J. H. Pavlovitch, P. Minoprio, D. Portnoi, L. Renia, D. Mazier. 1994. Murine AIDS protects mice against experimental cerebral malaria: down-regulation by interleukin 10 of a T-helper type 1 CD4+ cell-mediated pathology. Proc. Natl. Acad. Sci. USA 91:8097.[Abstract/Free Full Text]
  8. Finley, R. W., L. J. Mackey, P. H. Lambert. 1982. Virulent P. berghei malaria: prolonged survival and decreased cerebral pathology in cell-dependent nude mice. J. Immunol. 129:2213.[Abstract]
  9. Grau, G. E., P. F. Piguet, H. D. Engers, J. A. Louis, P. Vassalli, P. H. Lambert. 1986. L3T4+ T lymphocytes play a major role in the pathogenesis of murine cerebral malaria. J. Immunol. 137:2348.[Abstract]
  10. Hermsen, C., T. Vandewiel, E. Mommers, R. Sauerwein, W. Eling. 1997. Depletion of CD4+ or CD8+ T-cells prevents Plasmodium berghei induced cerebral malaria in end-stage disease. Parasitology 114:7.
  11. Yanez, D. M., J. Batchelder, H. C. van der Heyde, D. D. Manning, W. P. Weidanz. 1999. {gamma}{delta} T-cell function in pathogenesis of cerebral malaria in mice infected with Plasmodium berghei ANKA. Infect. Immun. 67:446.[Abstract/Free Full Text]
  12. Boubou, M. I., A. Collette, D. Voegtlé, D. Mazier, P.-A. Cazenave, S. Pied. 1999. T cell response in malaria pathogenesis: selective increase in T cells carrying the TCR V{beta}8 during experimental cerebral malaria. Int. Immunol. 11:1553.[Abstract/Free Full Text]
  13. Louis, J., H. Himmelrich, C. Parralopez, F. Tacchini-Cottier, P. Launois. 1998. Regulation of protective immunity against Leishmania major in mice. Curr. Opin. Immunol. 10:459.[Medline]
  14. Denkers, E. Y., R. T. Gazzinelli. 1998. Regulation and function of T-cell-mediated immunity during Toxoplasma gondii infection. Clin. Microbiol. Rev. 11:569.[Abstract/Free Full Text]
  15. Urban, B. C., D. J. P. Ferguson, A. Pain, N. Willcox, M. Plebanski, J. M. Austyn, D. J. Roberts. 1999. Plasmodium falciparum-infected erythrocytes modulate the maturation of dendritic cells. Nature 400:73.[Medline]
  16. Reina-San-Martin, B., A. Cosson, P. Minoprio. 2000. Lymphocyte polyclonal activation: a pitfall for vaccine design against infectious agents. Parasitol. Today 16:62.[Medline]
  17. Davis, M. M., P. J. Bjorkman. 1988. T-cell antigen receptor genes and T-cell recognition. Nature 334:395.[Medline]
  18. Pannetier, C., M. Cochet, S. Darche, A. Casrouge, M. Zöller, P. Kourilsky. 1993. The sizes of the CDR3 hypervariable regions of the murine T-cell receptor {beta} chains vary as a function of the recombined germ-line segments. Proc. Natl. Acad. Sci. USA 90:4319.[Abstract/Free Full Text]
  19. Pannetier, C., J. Even, P. Kourilsky. 1995. T-cell diversity and clonal expansion in normal and clinical samples. Immunol. Today 16:176.[Medline]
  20. Amani, V., M. I. Boubou, S. Pied, M. Marussig, D. Walliker, D. Mazier, L. Renia. 1998. Cloned lines of Plasmodium berghei ANKA differ in their abilities to induce experimental cerebral malaria. Infect. Immun. 66:4093.[Abstract/Free Full Text]
  21. Pannetier, C., J. Even, P. Kourilsky. 1997. The Immunoscope technique for analysis of TCR repertoire. The Human Antigen T Cell Receptor: Selected Protocols and Applications 287. J. R. Oksenberg, Austin.
  22. Chou, H. S., S. J. Anderson, M. C. Louie, S. A. Godambe, M. R. Pozzi, M. A. Behlke, K. Huppi, D. Y. Loh. 1987. Tandem linkage and unusual RNA splicing of the T-cell receptor {beta}-chain variable-region genes. Proc. Natl. Acad. Sci. USA 84:1992.[Abstract/Free Full Text]
  23. Wade, T., J. Bill, P. C. Marrack, E. Palmer, J. W. Kappler. 1988. Molecular basis for the nonexpression of V{beta}17 in some strains of mice. J. Immunol. 141:2165.[Abstract]
  24. Louie, M. C., C. A. Nelson, D. Y. Loh. 1989. Identification and characterization of new murine T cell receptor {beta} chain variable region (V{beta}) genes. J. Exp. Med. 170:1987.[Abstract/Free Full Text]
  25. Collette, A., A. Six. 2002. ISEApeaks: an Excel platform for GeneScan and Immunoscope data retrieval, management and analysis. Bioinformatics 18:329.[Abstract/Free Full Text]
  26. Collette, A., P.-A. Cazenave, S. Pied, A. Six. 2003. New methods and software tools for high throughput CDR3 spectratyping: application to T lymphocyte repertoire modifications during experimental malaria. J. Immunol. Methods 278:105.[Medline]
  27. Gorochov, G., A. U. Neumann, A. Kereveur, C. Parizot, T. S. Li, C. Katlama, M. Karmochkine, G. Raguin, B. Autran, P. Debré. 1998. Perturbation of CD4+ and CD8+ T-cell repertoires during progression to AIDS and regulation of the CD4+ repertoire during antiviral therapy. Nat. Med. 4:215.[Medline]
  28. Han, M., L. Harrison, P. Kehn, K. Stevenson, J. Currier, M. A. Robinson. 1999. Invariant or highly conserved TCR{alpha} are expressed on double-negative (CD3+CD4–CD8–) and CD8+ T cells. J. Immunol. 163:301.[Abstract/Free Full Text]
  29. Rencher, A. C.. 1995. Methods of Multivariate Analysis J. Wiley, New York.
  30. Spanakis, E., D. Brouty-Boye. 1997. Discrimination of fibroblast subtypes by multivariate analysis of gene expression. Int. J. Cancer 71:402.[Medline]
  31. Tavazoie, S., J. D. Hughes, M. J. Campbell, R. J. Cho, G. M. Church. 1999. Systematic determination of genetic network architecture. Nat. Genet. 22:281.[Medline]
  32. Boudinot, P., S. Boubekeur, A. Benmansour. 2001. Rhabdovirus infection induces public and private T cell responses in teleost fish. J. Immunol. 167:6202.[Abstract/Free Full Text]
  33. Musette, P., A. Galelli, P. Truffa-Bachi, W. Peumans, P. Kourilsky, G. Gachelin. 1996. The J{beta} segment of the T cell receptor contributes to the V{beta}-specific T cell expansion caused by staphylococcal enterotoxin B and Urtica dioica superantigens. Eur. J. Immunol. 26:618.[Medline]
  34. Pannetier, C., S. Delassus, S. Darche, C. Saucier, P. Kourilsky. 1993. Quantitative titration of nucleic acids by enzymatic amplification reactions run to saturation. Nucleic Acids Res. 21:577.[Abstract/Free Full Text]
  35. Levraud, J. P., C. Pannetier, P. Langlade-Demoyen, V. Brichard, P. Kourilsky. 1996. Recurrent T cell receptor rearrangements in the cytotoxic T lymphocyte response in vivo against the p815 murine tumor. J. Exp. Med. 183:439.[Abstract/Free Full Text]
  36. Faure, M., S. Calbo, J. Kanellopoulos, A. M. Drapier, P. A. Cazenave, D. Rueff-Juy. 1999. Tolerance to maternal immunoglobulins: resilience of the specific T cell repertoire in spite of long-lasting perturbations. J. Immunol. 163:6511.[Abstract/Free Full Text]
  37. Ballet, J. J., P. Druilhe, M. A. Querleux, C. Schmitt, M. Agrapart. 1981. Parasite-derived mitogenic activity for human T cells in Plasmodium falciparum continuous cultures. Infect. Immun. 33:758.[Abstract/Free Full Text]
  38. Riley, E. M., G. Anderson, L. N. Otoo, S. Jepsen, B. M. Greenwood. 1988. Cellular immune responses to Plasmodium falciparum antigens in Gambian children during and after acute attack of falciparum malaria. Clin. Exp. Immunol. 73:17.[Medline]
  39. Ho, M., H. K. Webster, B. Green, S. Looareesuwan, S. Kongchareon, N. J. White. 1988. Defective production of and response to IL-2 in acute human falciparum malaria. J. Immunol. 141:2755.[Abstract]
  40. Mercado, T. I.. 1973. Plasmodium berghei: inhibition by splenectomy of a paralyzing syndrome in infected rats. Exp. Parasitol. 34:142.[Medline]
  41. Hermsen, C. C., E. Mommers, T. van de Wiel, R. W. Sauerwein, W. M. Eling. 1998. Convulsions due to increased permeability of the blood-brain barrier in experimental cerebral malaria can be prevented by splenectomy or anti-T cell treatment. J. Infect. Dis. 178:1225.[Medline]
  42. Mackay, C. R., U. H. von Andrian. 2001. Memory T cells: local heroes in the struggle for immunity. Science 291:2323.[Free Full Text]
  43. Bagot, S., F. Nogueira, A. Collette, V. do Rosario, F. Lemonier, P.-A. Cazenave, S. Pied. 2004. Comparative study of brain CD8+ T cells induced by sporozoites and those induced by blood-stage Plasmodium berghei ANKA involved in the development of cerebral malaria. Infect. Immun. 72:2817.[Abstract/Free Full Text]
  44. Belnoue, E., M. Kayibanda, A. M. Vigario, J.-C. Deschemin, N. Van Rooijen, M. Viguier, G. Snounou, L. Renia. 2002. On the pathogenic role of brain-sequestered {alpha}{beta} CD8+ T cells in experimental cerebral malaria. J. Immunol. 169:6369.[Abstract/Free Full Text]
  45. Gorgette, O., A. Existe, M. Idrissa-Boubou, S. Bagot, J.-L. Guénet, D. Mazier, P.-A. Cazenave, S. Pied. 2002. Deletion of T cells bearing V{beta}8.1 TCR following MTV-7 integration confers resistance to murine cerebral malaria. Infect. Immun. 70:3701.[Abstract/Free Full Text]
  46. De Kossodo, S., G. E. Grau. 1993. Profiles of cytokine production in relation with susceptibility to cerebral malaria. J. Immunol. 151:4811.[Abstract]
  47. Bousso, P., A. Casrouge, J. D. Altman, M. Haury, J. Kanellopoulos, J. P. Abastado, P. Kourilsky. 1998. Individual variations in the murine T cell response to a specific peptide reflect variability in naive repertoires. Immunity 9:169.[Medline]
  48. Pied, S., D. Voegtlé, M. Marussig, L. Rénia, F. Miltgen, D. Mazier, P.-A. Cazenave. 1997. Evidence for a superantigenic activity during murine malaria infection. Int. Immunol. 9:17.[Abstract/Free Full Text]
  49. Subauste, C. S., F. Fuh, R. D. Malefyt, J. S. Remington. 1998. {alpha}{beta} T cell response to Toxoplasma gondii in previously unexposed individuals. J. Immunol. 160:3403.[Abstract/Free Full Text]
  50. Bagot, S., S. Campino, C. Penha-Gonçalves, S. Pied, P.-A. Cazenave, D. Holmberg. 2002. Identification of two cerebral malaria resistance loci using an inbred wild derived mouse strain. Proc. Natl. Acad. Sci. USA 99:9919.[Abstract/Free Full Text]
  51. Kabat, E. A., T. T. Wu, H. M. Perry, L. S. Gottesman, C. Foeller. 1991. Sequences of Proteins of Immunological Interest Bethesda, MD.



This article has been cited by other articles:


Home page
 J. Immunol.Home page
C. Blache, S. Adriouch, S. Calbo, L. Drouot, S. Dulauroy, C. Arnoult, S. Le Corre, A. Six, M. Seman, and O. Boyer
Cutting Edge: CD4-Independent Development of Functional FoxP3+ Regulatory T Cells
J. Immunol., October 1, 2009; 183(7): 4182 - 4186.
[Abstract] [Full Text] [PDF]

Home page
 J. Immunol.Home page
D. Bernard, A. Six, L. Rigottier-Gois, S. Messiaen, S. Chilmonczyk, E. Quillet, P. Boudinot, and A. Benmansour
Phenotypic and Functional Similarity of Gut Intraepithelial and Systemic T Cells in a Teleost Fish
J. Immunol., April 1, 2006; 176(7): 3942 - 3949.
[Abstract] [Full Text] [PDF]

Home page
 Proc. Natl. Acad. Sci. USAHome page
B. Franke-Fayard, C. J. Janse, M. Cunha-Rodrigues, J. Ramesar, P. Buscher, I. Que, C. Lowik, P. J. Voshol, M. A. M. den Boer, S. G. van Duinen, et al.
From The Cover: Murine malaria parasite sequestration: CD36 is the major receptor, but cerebral pathology is unlinked to sequestration
PNAS, August 9, 2005; 102(32): 11468 - 11473.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Collette, A.
Right arrow Articles by Pied, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Collette, A.
Right arrow Articles by Pied, S.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS