点击显示 收起
Department of Immunology and Microarray Core Facilities, Max Planck Institute for Infection Biology, Berlin, Institute for Medical Biometry and Statistics, University of Lübeck, Lübeck
Asklepios Center for Respiratory Medicine and Thoracic Surgery, Munich-Gauting, Germany
Ras-associated small GTPases (Rabs) are specific regulators of intracellular vesicle trafficking. Interference with host cell vesicular transport is a hallmark of many intracellular pathogens, including the notable example Mycobacterium tuberculosis. We performed, by quantitative polymerase chain reaction, gene-expression analyses for selected Rab molecules in peripheral-blood mononuclear cells from patients with tuberculosis (TB) and healthy control subjects, to identify candidate genes that are critically involved in the host immune response. Comparison revealed significant differences in the expression of genes for Rab13, Rab24, and Rab33A. Rab33A gene expression was down-regulated in patients with TB and was predominantly expressed in CD8+ T cells. We excluded possible influences of differences in T cell percentages between the 2 study groups, demonstrating that Rab33A gene expression changes on the single-cell level. In vitro, Rab33A RNA expression was induced in T cells on activation and by dendritic cells infected with M. tuberculosis. Our findings identify Rab33A as a T cell regulatory molecule in TB and suggest its involvement in disease processes.
Tuberculosis (TB) remains a major health threat, with 9 million new cases and 2 million deaths occurring annually worldwide [1]. The situation has worsened with increasing incidences of multidrug-resistant TB, which amount to >300,000 new cases annually worldwide. However, the vast majority of the estimated 2 billion individuals infected with the main etiologic agent, Mycobacterium tuberculosis, are capable of controlling the pathogen, thus preventing disease outbreak. Protective immunity against M. tuberculosis involves a broad repertoire of innate and adaptive immune response mechanisms [2]. CD4+ as well as CD8+ T cells are essential mediators of protection against facultative intracellular pathogens such as M. tuberculosis [3, 4]. CD8+ T cells contribute to the antimycobacterial host response in manifold fashion, including (1) cytotoxicity mediated via a granule-dependent exocytosis pathway; (2) cytotoxicity mediated via Fas/Fas ligand interaction; (3) direct microbicidal activity; and (4) release of cytokines [5]. Among relevant cytokines, T cellderived interferon (IFN) is central to the protective host immune response, both in mice and in humans [6, 7]. IFN- is expressed on antigen-specific activation and leads to maturation of effector T cell functions and activation of infected macrophages. Macrophages play a dual role in the antimycobacterial host response: on the one hand, they restrict the growth of engulfed mycobacteria; on the other hand, they shield the pathogen from being extinguished [8]. M. tuberculosis resides in phagosomal vesicles, where complex interactions between pathogen and host have evolved. A crucial element of mycobacterial survival involves the blockade of phagosomal maturation, which normally leads to the death of the engulfed microbes [9]. In this process, the recruitment of regulatory enzymes to the phagosomal membrane is disrupted by M. tuberculosis. Ras-associated small GTPase (Rab) 5 and Rab7 are essential mediators of these fusion events; Rab5 accumulates in the membranes of M. tuberculosiscontaining phagosomes, whereas recruitment of Rab7 is inhibited [9]. Rab molecules play decisive roles in the regulation of intracellular trafficking in various cell types and tissues [10]. In T cells, for example, Rab27A controls the exocytosis of lysosomes as an important step in the release of cytotoxic granules as part of antimicrobial T cell effector functions. More than 60 Rab family members presumably exist in humans, and some of them play vital roles in distinct diseases [11]; nonetheless, the function of the majority of them remains unknown [10]. Broader implications for the crucial role played by regulatory GTPases against TB have arisen from mouse studies, in which LRG-47, a member of the IFN-inducible p47 GTPase family, has been shown to be essential for protection against TB [12].
On the basis of differential gene expression in microarray analyses of peripheral-blood mononuclear cells (PBMCs) from patients with TB and healthy control subjects, we selected 13 Rab candidates. Quantitative PCR (qPCR) analyses revealed 3 differentially expressed Rab candidates, which were further defined with respect to their cellular distribution. Moreover, the function of Rab33A was tentatively characterized in a model of M. tuberculosis infection.
PARTICIPANTS, MATERIALS, AND METHODS
Patients and control subjects.
Patients with TB were recruited at the Asklepios Center for Respiratory Medicine and Thoracic Surgery. Diagnosis was based on chest radiography, and laboratory confirmation was based on mycobacterial culture analysis. Healthy control subjects without a history of clinical TB were recruited from the laboratory staff at the Max Planck Institute for Infection Biology and the hospital staff at the Asklepios Center.
Altogether, 53 patients with TB and 47 control subjects (table 1) were included in the present study. Because of limitations in specimen quantity, subgroups were randomly chosen for different experiments. All donors provided written, informed consent. This study was approved by the local ethics committees (ek.205-18.1) and was conducted in accordance with the Helsinki Declaration.
Preparation of PBMCs and magnetic cell sorting.
Forty milliliters of heparinized peripheral venous blood was obtained from each patient with TB and control subject, and PBMCs were isolated on Ficoll gradients (Biochrom). For RNA isolation, cells were immediately mixed with TRIzol reagent (Invitrogen) and were frozen at -80°C until the RNA was extracted in accordance with the manufacturer's instructions. RNA content, purity, and integrity were determined by use of an Agilent 2100 bioanalyzer (Agilent Technologies). For flow cytometry and in vitro assays, cells were frozen in RPMI 1640 containing 10% fetal calf serum (FCS; Invitrogen) and 10% dimethyl sulfoxide (Sigma-Aldrich) in liquid nitrogen. Freshly isolated immune cell populations were enriched by magnetic cell sorting with microbead-labeled monoclonal antibodies against CD3, CD4, CD8, or CD14 (Miltenyi Biotec) and positive-selection or depletion columns (LS and AS separation columns; Miltenyi Biotec), in accordance with the manufacturer's instructions.
Real-time qPCR analysis.
RNA was reverse-transcribed to cDNA as described elsewhere [13]. SYBR Green (Applied Biosystems) uptake in double-stranded DNA was measured by use of an ABI PRISM 7000 thermocycler (Applied Biosystems), in accordance with the manufacturer's instructions. We designed primer pairs by use of ABI PRISM Primer Express software (version 2.0.0; Applied Biosystems). Primer sequences are summarized in table 2. With the exception of the in vitro stimulation assays, for which -glucuronidase was selected as the internal standard, glyceraldehyde-3-phosphate dehydrogenase was used as the internal control.
Flow cytometry.
Flow cytometry was performed to analyze the efficiency of magnetic cell sorting and to determine the percentages of T cells in the PBMC populations of 24 of the patients with TB and 20 of the control subjects. Percentages are based on gated PBMCs; cell debris was excluded on the basis of size and granularity. Antibodies against CD4-allophycocyanin (APC; clone RPA-T4; BD Pharmingen), CD3fluorescein isothiocyanate (clone UCHT-1; BD Pharmingen), CD8peridinin chlorophyll protein (clone SK1; BD Pharmingen), and CD14-APC-phycoerythrin (clone M5E2; BD Pharmingen) were used to stain immune cell population markers. The staining procedures used were the same as those described elsewhere [13]. A FACSCalibur flow cytometer and CellQuest software (BD Biosciences) were used for analyses. T cell percentages were determined by use of FCS Express software (version 2; De Novo Software).
In vitro stimulation assays.
For in vitro T cell activation, PBMCs were incubated with either 10 g/mL phytohemagglutinin (PHA) or 10 L/mL CD3/CD28 Dynabeads (Dynal Biotech) for 4, 24, and 72 h in RPMI 1640 (Invitrogen) containing 5% male human AB serum (Sigma-Aldrich), 1% penicillin, and 1% streptomycin (Invitrogen). For stimulation of PBMCs, M. tuberculosis H37Rv straininfected dendritic cells (DCs) and CD14+ monocytic cells were isolated, and 1 × 106 cells/well (12-well plate; Greiner Bio-One) were incubated for 5 days with 200 U/mL recombinant human granulocyte-macrophage colony-stimulating factor (Sigma-Aldrich) and 5 ng/mL recombinant human interleukin (IL)4 (Biosource International) in RPMI 1640 containing 5% human serum. For infection of DCs with M. tuberculosis, 5 × 106 log phase mycobacteria in RPMI 1640 containing 10% FCS were added to DCs and were incubated for 3 h at 37°C in 5% CO2. Afterward, the supernatant was discarded, and 5 × 106 PBMCs, prestimulated PBMCs, or selected CD4+ or CD8+ T cells in RPMI 1640 containing 5% FCS were added. Prestimulation of PBMCs was achieved by incubation with 10 g/mL M. tuberculosis protein extract for 7 days at 37°C in 5% CO2; after 2 days, 20 U/mL IL-2 (Strathmann Biotech) was added to the culture. M. tuberculosis H37Rv strain was cultured for 3 days in 7H11 medium (BD Biosciences). For DC infection, mycobacteria were resuspended and counted. M. tuberculosis proteins were extracted by use of Fast Protein Blue Matrix tubes (Qbiogene), in accordance with the manufacturer's instructions.
Confounding of Rab33A gene expression.
Rab33A gene expression, as determined by qPCR, was analyzed for confounding of the differences in T cell percentages between the 2 study groups. To adjust for confounding, we used the following linear model:
where y is the measured gene expression, is the overall mean (i.e., average of all expression data), p is the percentage of T cells in the respective PBMC population as a confounding factor, G is a binary factor (control subjects are encoded as G = 0, whereas patients with TB are encoded as G = 1), p × G is the interaction effect of study group and T cell percentage, and is the residual.
A significant interaction effect implies that there are differences in single-cell gene-expression levels in the specific cells under consideration between the patients with TB and the control subjects. In this case, measured gene-expression signals are a function of both the quantity of the specific cells in the respective PBMC population as well as their single-cell gene-expression levels.
Statistical analysis of T cell percentages and gene expression.
Differences in T cell percentages, as measured by flow cytometry, and in Rab candidate gene expression were assessed by the Mann-Whitney U test. All reported P values are nominal and are not adjusted for multiple testing. P < .05 was considered to be statistically significant.
RESULTS
qPCR analyses of Rab molecules in PBMCs.
In initial microarray experiments, molecules involved in intracellular trafficking processesnotably Rab family membersshowed differential expression between PBMCs from the patients with TB and those from the control subjects (authors' unpublished data). To follow up on these results, we chose 13 Rab candidates that showed a tendency toward differential gene expression for further evaluation. PBMCs from 23 of the patients with TB and 29 of the control subjects were analyzed by qPCR for gene expression of the Rab candidates. The genes for 3 of themRab33A, Rab24, and Rab13were differentially expressed between the 2 study groups (P < .001, P = .005, and P = .02, respectively) (figure 1). In the patients with TB, Rab33A was down-regulated, whereas Rab24 and Rab13 were up-regulated.
Preferential Rab33A gene expression in CD8+ T cells.
Because relevant PBMC subpopulations for these Rab candidates are ill defined, we determined the levels of gene expression in CD14+ monocytes and CD3+ T cells. Comparison with total gene expression in all PBMCs revealed preferential Rab33A gene expression in T cells and moderately increased Rab24 gene expression in monocytes (figure 2A). Both monocytes and T cells equally expressed the Rab13 gene, at a level slightly above total gene expression in all PBMCs. To further define the Rab33A geneexpressing T cell subpopulation, we both enriched and depleted CD4+, CD8+, and total CD3+ T cells (figure 2B). As expected, enrichment of CD3+ T cells increased Rab33A RNA levels, whereas enrichment of CD4+ T cells had hardly any effect. In contrast, enrichment of CD8+ T cells caused a 5-fold increase in Rab33A gene expression, compared with total gene expression in all PBMCs. These results suggested that Rab33A might be a novel CD8+ T cell factor that is relevant to adaptive immunity against TB; we therefore focused our study on Rab33A, to clarify the possible role that this molecule might play.
Absence of an effect of sex on Rab33A gene expression.
Rab33A is an X chromosomally coded gene located at Xq26, a region that has been found to be associated with susceptibility to TB in an African population [14]. Accordingly, we analyzed the possible effect of sex on Rab33A gene expression by comparing groups of participants categorized by disease status and sex. No difference between the male and the female control subjects was detected, but the male patients with TB, compared with the control subjects of either sex, had significantly down-regulated Rab33A gene expression (P = .0001, for comparison with the female control subjects; P = .003, for comparison with the male control subjects) (figure 3). The differences between the female patients with TB and the male and the female control subjects were less pronounced but still reached significant levels (P = .03, for comparison with the female control subjects; P = .01, for comparison with the male control subjects) (figure 3). These results demonstrate the disease-specific decrease in Rab33A gene expression and render an effect of sex improbable.
Induction of Rab33A gene expression in vitro on T cell activation and by M. tuberculosisinfected DCs.
Limited information on the function of Rab33A is available [15]. Formerly known as S10, Rab33A was discovered during the early 1990s and was found to be expressed in the human T cell line Jurkat and the monocyte cell line U937 [15]. Rab33A fulfills all of the criteria for small GTPases and, therefore, likely plays a role in the regulation of intracellular vesicle transport or signaling. As a first step toward characterization of the function of Rab33A, we stimulated PBMCs in vitro with the T cell mitogen PHA or CD3/CD28 Dynabeads for various periods of time. Afterward, RNA levels for Rab33A and IFN- were analyzed by qPCR. As was expected, IFN- RNA levels were strongly increased after PHA or T cell receptor (TCR) stimulation; compared with the levels in unstimulated control wells, IFN- RNA levels were increased 2425-fold after only 4 h and then were increased 211212-fold after 24 h (figure 5A, bottom graph). Rab33A showed no difference in expression after 4 h but did show a 2121.5-fold increase after 24 h and a 23-fold increase after 72 h (figure 5A, top graph). Because Rab27A is an important regulator of the release of cytotoxic granules by T cells, we determined its expression under the same conditions and found a pattern similar to that for Rab33A (data not shown). These results demonstrate that, analogous to Rab27A, Rab33A is induced on TCR stimulation. The impact of M. tuberculosis infection on Rab33A gene expression was determined in vitro by culturing different immune cell populations with monocyte-derived DCs that had been pulsed with M. tuberculosis. PBMCs, PBMCs prestimulated with M. tuberculosis protein, CD4+ T cells, and CD8+ T cells were cocultured with infected DCs for 48 h. Rab33A and IFN- gene expression was strongly induced in all tested cell populations (figure 5B). Interestingly, IFN- gene expression after stimulation with infected DCs was least pronounced in CD8+ T cells (28-fold less than that in PBMCs and 24-fold less than that in CD4+ T cells) (figure 5B, bottom graph). In contrast, Rab33A gene expression in CD8+ T cells after stimulation with infected DCs was similar to that in the total PBMC population and was elevated when compared with that in CD4+ T cells (figure 5B, top graph). In vitro restimulation with M. tuberculosis protein increased IFN- and Rab33A gene expression, but no effect on Rab33A gene expression was observed after stimulation with infected DCs. Therefore, Rab33A gene expression is induced in T cells after in vitro stimulation with M. tuberculosis.
Finally, we compared Rab33A gene expression in PBMCs from the patients with TB and the control subjects after in vitro T cell stimulation. PBMC samples were stimulated for 48 h with CD3/28 Dynabeads. We detected no significant difference between the 2 study groups, although median Rab33A gene expression was slightly higher in the control subjects (P = .64) (figure 6). Interestinglyand consistent with our M. tuberculosisinfected DC experiments (figure 5B)no correlation between Rab33A gene expression and the amount of IFN- RNA induced by stimulation was detected (data not shown).
DISCUSSION
Here we describe the involvement of regulatory molecules of the Rab family in the host immune response against TB. Three moleculesRab33A, Rab24, and Rab13showed differential expression in PBMCs from patients with TB versus healthy control subjects. Gene expression analyses of PBMC subpopulations revealed that Rab33A is a novel CD8+ T cell factor with significantly reduced gene expression in patients with TB. Rab33A gene expression was induced in vitro on PHA or TCR stimulation and by coculture with M. tuberculosisinfected DCs.
Strong evidence supports a decisive role for CD8+ T cells in the defense against TB, both in animal models [16, 17] and in human TB [18]. In humans, killing of M. tuberculosis in macrophages by CD8+ T cells via a perforin/granulysinmediated process has been described [19]. At present, Rab27A is the best-defined Rab family member related to cytotoxic CD8+ T cell functions (reviewed in [20, 21]). Knowledge of the impact of Rab27A derives from a study in patients with Griscelli syndrome [22] as well as a study in mice [23]. In these studies, a functional mutation in the Rab27A gene impaired the release of secretory lysosomes by melanocytes and CD8+ T cells. The dramatic effect of Rab27A failure on CD8+ T cell function demonstrates that this Rab family member plays a decisive role in acquired immunity. Our own results for Rab27A imply an induction pattern comparable to that for Rab33A (data not shown). It is tempting to speculate that Rab33A plays a similar role in antigen-specific CD8+ T cell responses against TB. The gene encoding Rab33A is located on the X chromosome (Xq26) and is close to a region that has been found to be critical to susceptibility to TB in an African population [14, 24]. Our data demonstrate that, on TCR stimulation, PBMCs from patients with TB express Rab33A at a level comparable to that of PBMCs from control subjects. Nevertheless, a specific impact of genetic variations in Rab33A on susceptibility to TB has yet to be determined. Although further experiments are, likewise, needed to determine the precise biological role played by Rab33A in CD8+ T cell responses, our results demonstrate that this molecule qualifies as a candidate correlate of resistance against and susceptibility to TB.
In the present study, Rab24, which was preferentially expressed in monocytes, was up-regulated in patients with TB. Rab24 may participate in fusion events between the endoplasmatic reticulum/cis-Golgi compartment and lysosomes that control autophagy processes [25, 26]. Induction of autophagy changes the distribution of Rab24, leading to its accumulation in autophagic vesicles [26]. Very recently, an important role for autophagy in the host immune response against M. tuberculosis infection has been observed [27]. This study revealed that autophagy abrogates the blockage of phagosome maturation, leading to increased M. tuberculosis elimination within macrophages. It remains to be determined whether autophagy processes are prevalent in PBMCs from patients with TB.
Our third candidate, up-regulated in PBMCs from patients with TB, was Rab13. Rab13 is preferentially expressed in epithelial cells, which are crucially involved in the regulation of functional tight junction assembly [28, 29]. Recently, the role played by Rab13 in this process was further clarifiedMorimoto et al. demonstrated that Rab13 specifically regulates continuous endocytic recycling in epithelial cell lines [30]. In this study, a dominant active mutant (Rab13 Q67L) selectively inhibited the postendocytic recycling of occludin but not that of transferrin receptors. Endosomal recycling processesand especially the transferrin-mediated iron metabolismcontribute to interactions between mycobacteria and host macrophages. In this way, M. tuberculosis improves its intracellular habitat [31]. Rab5 and Rab7 are key, well-defined regulators of phagosome maturation, a process that is targeted by M. tuberculosis [9]. It remains to be established whether Rab13 plays a similar role in the early endosome recycling of immune cells.
In summary, our data emphasize the critical involvement of Rab family members in the host response against TB. Our data have also identified Rab33A as a novel CD8+ T cell factor that is likely involved in resistance against and susceptibility to the disease.
References
1. World Health Organization. TB as a public health problem. Available at: http://w3.whosea.org/tb/publichealth.htm. Last accessed 15 August 2005. First citation in article
2. Flynn JL, Chan J. Immunology of tuberculosis. Annu Rev Immunol 2001; 19:93129. First citation in article
3. Murray PJ. Defining the requirements for immunological control of mycobacterial infections. Trends Microbiol 1999; 7:36672. First citation in article
4. Stenger S, Modlin RL. T cell mediated immunity to Mycobacterium tuberculosis. Curr Opin Microbiol 1999; 2:8993. First citation in article
5. Lazarevic V, Flynn J. CD8+ T cells in tuberculosis. Am J Respir Crit Care Med 2002; 166:111621. First citation in article
6. Cooper AM, Dalton DK, Stewart TA, Griffin JP, Russell DG, Orme IM. Disseminated tuberculosis in interferon gamma gene-disrupted mice. J Exp Med 1993; 178:22437. First citation in article
7. Jouanguy E, Altare F, Lamhamedi S, et al. Interferon-gamma-receptor deficiency in an infant with fatal bacille Calmette-Guerin infection. N Engl J Med 1996; 335:195661. First citation in article
8. Kaufmann SH. How can immunology contribute to the control of tuberculosis Nat Rev Immunol 2001; 1:2030. First citation in article
9. Via LE, Deretic D, Ulmer RJ, Hibler NS, Huber LA, Deretic V. Arrest of mycobacterial phagosome maturation is caused by a block in vesicle fusion between stages controlled by rab5 and rab7. J Biol Chem 1997; 272:1332631. First citation in article
10. Zerial M, McBride H. Rab proteins as membrane organizers. Nat Rev Mol Cell Biol 2001; 2:10717. First citation in article
11. Stein MP, Dong J, Wandinger-Ness A. Rab proteins and endocytic trafficking: potential targets for therapeutic intervention. Adv Drug Deliv Rev 2003; 55:142137. First citation in article
12. MacMicking JD, Taylor GA, McKinney JD. Immune control of tuberculosis by IFN-gamma-inducible LRG-47. Science 2003; 302:6549. First citation in article
13. Jacobsen M, Schweer D, Ziegler A, et al. A point mutation in PTPRC is associated with the development of multiple sclerosis. Nat Genet 2000; 26:4959. First citation in article
14. Bellamy R, Beyers N, McAdam KP, et al. Genetic susceptibility to tuberculosis in Africans: a genome-wide scan. Proc Natl Acad Sci USA 2000; 97:80059. First citation in article
15. Koda T, Kakinuma M. Molecular cloning of a cDNA encoding a novel small GTP-binding protein. FEBS Lett 1993; 328:214. First citation in article
16. Sousa AO, Mazzaccaro RJ, Russell RG, et al. Relative contributions of distinct MHC class I-dependent cell populations in protection to tuberculosis infection in mice. Proc Natl Acad Sci USA 2000; 97:42048. First citation in article
17. Flynn JL, Goldstein MM, Triebold KJ, Koller B, Bloom BR. Major histocompatibility complex class I-restricted T cells are required for resistance to Mycobacterium tuberculosis infection. Proc Natl Acad Sci USA 1992; 89:120137. First citation in article
18. Stenger S, Mazzaccaro RJ, Uyemura K, et al. Differential effects of cytolytic T cell subsets on intracellular infection. Science 1997; 276:16847. First citation in article
19. Stenger S, Hanson DA, Teitelbaum R, et al. An antimicrobial activity of cytolytic T cells mediated by granulysin. Science 1998; 282:1215. First citation in article
20. Trambas CM, Griffiths GM. Delivering the kiss of death. Nat Immunol 2003; 4:399403. First citation in article
21. Stinchcombe J, Bossi G, Griffiths GM. Linking albinism and immunity: the secrets of secretory lysosomes. Science 2004; 305:559. First citation in article
22. Menasche G, Pastural E, Feldmann J, et al. Mutations in RAB27A cause Griscelli syndrome associated with haemophagocytic syndrome. Nat Genet 2000; 25:1736. First citation in article
23. Wilson SM, Yip R, Swing DA, et al. A mutation in Rab27a causes the vesicle transport defects observed in ashen mice. Proc Natl Acad Sci USA 2000; 97:79338. First citation in article
24. Bellamy R, Hill AV. Genetic susceptibility to mycobacteria and other infectious pathogens in humans. Curr Opin Immunol 1998; 10:4837. First citation in article
25. Olkkonen VM, Dupree P, Killisch I, Lutcke A, Zerial M, Simons K. Molecular cloning and subcellular localization of three GTP-binding proteins of the rab subfamily. J Cell Sci 1993; 106:124961. First citation in article
26. Munafo DB, Colombo MI. Induction of autophagy causes dramatic changes in the subcellular distribution of GFP-Rab24. Traffic 2002; 3:47282. First citation in article
27. Gutierrez MG, Master SS, Singh SB, Taylor GA, Colombo MI, Deretic V. Autophagy is a defense mechanism inhibiting BCG and Mycobacterium tuberculosis survival in infected macrophages. Cell 2004; 119:75366. First citation in article
28. Marzesco AM, Dunia I, Pandjaitan R, et al. The small GTPase Rab13 regulates assembly of functional tight junctions in epithelial cells. Mol Biol Cell 2002; 13:181931. First citation in article
29. Sheth B, Fontaine JJ, Ponza E, et al. Differentiation of the epithelial apical junctional complex during mouse preimplantation development: a role for rab13 in the early maturation of the tight junction. Mech Dev 2000; 97:93104. First citation in article
30. Morimoto S, Nishimura N, Terai T, et al. Rab13 mediates the continuous endocytic recycling of occludin to the cell surface. J Biol Chem 2004; 280:22208. First citation in article
31. Clemens DL, Horwitz MA. The Mycobacterium tuberculosis phagosome interacts with early endosomes and is accessible to exogenously administered transferrin. J Exp Med 1996; 184:134955. First citation in article
32. Bennett L, Palucka AK, Arce E, et al. Interferon and granulopoiesis signatures in systemic lupus erythematosus blood. J Exp Med 2003; 197:71123. First citation in article
33. Bomprezzi R, Ringner M, Kim S, et al. Gene expression profile in multiple sclerosis patients and healthy controls: identifying pathways relevant to disease. Hum Mol Genet 2003; 12:21919. First citation in article
34. Bullinger L, Dohner K, Bair E, et al. Use of gene-expression profiling to identify prognostic subclasses in adult acute myeloid leukemia. N Engl J Med 2004; 350:160516. First citation in article
35. Twine NC, Stover JA, Marshall B, et al. Disease-associated expression profiles in peripheral blood mononuclear cells from patients with advanced renal cell carcinoma. Cancer Res 2003; 63:606975. First citation in article