Evasion of Peptide, but Not Lipid Antigen Presentation, Through Pathogen-Induced Dendritic Cell Maturation

Evasion of peptide, but not lipid antigen presentation, through pathogen-induced dendritic cell maturation

David L. Hava*, Nicole van der Wel†, Nadia Cohen*, Christopher C. Dascher‡, Diane Houben†, Luis Leo´n*, Sandeep Agarwal*, Masahiko Sugita§, Maaike van Zon†, Sally C. Kent¶, Homayoun Shamsʈ, Peter J. Peters†, and Michael B. Brenner*,**

*Division of Rheumatology, Immunology, and Allergy, Brigham and Women’s Hospital, Harvard Medical School, 1 Jimmy Fund Way, Boston, MA 02115; †The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX, Amsterdam, The Netherlands; §Laboratory of Cell Regulation, Institute for Virus Research, Kyoto University, 53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan; ¶Center for Neurologic Diseases, Brigham and Women’s Hospital, 77 Avenue Louis Pasteur, Boston, MA 02115; ʈCenter for Pulmonary and Infectious Diseases Control, University of Texas Health Center at Tyler, Tyler, TX 75708; and ‡Immunology Institute, Mount Sinai School of Medicine, 1 Gustave L. Levy Place, New York, NY 10029

Contributed by Michael B. Brenner, May 13, 2008 (sent for review March 20, 2008) Dendritic cells (DC) present lipid and peptide antigens to T cells on CD1 infected hosts. These strategies include the alteration of phagosome and MHC Class II (MHCII), respectively. The relative contribution of maturation in macrophages (13). In contrast, DCs are less able to these systems during the initiation of adaptive immunity after mi- support the growth of Mtb during the initial 48 h after phagocytosis crobial infection is not characterized. MHCII molecules normally ac- (14, 15), and following this time, Mtb have the ability to escape from quire antigen and rapidly traffic from phagolysosomes to the plasma phagolysosomes to replicate in the cytosol of DC (14). In all of these membrane as part of DC maturation, whereas CD1 molecules instead instances, the intersection of Mtb with the MHCII antigen presen- continually recycle between these sites before, during, and after DC tation pathway is changed. Given that DCs are required for the maturation. We find that in Mycobacterium tuberculosis (Mtb)-in- initiation of Mtb-specific adaptive immune responses (16–18), these fected DCs, CD1 presents antigens quickly. Surprisingly, rapid DC findings raise the issue of when during the course of infection DC maturation results in early failure and delay in MHCII presentation. are able to stimulate MHCII responses. Whereas both CD1b and MHCII localize to bacterial phagosomes early Further, because MHCII and CD1b traffic differently during DC after phagocytosis, MHCII traffics from the phagosome to the plasma maturation, we sought to determine whether their antigen presen- membrane with a rapid kinetic that can precede antigen availability tation functions differently after Mtb infection of DCs, potentially and loading. Thus, rather than facilitating antigen presentation, a lack leading to the discrete induction of lipid and peptide immune of coordination in timing may allow organisms to use DC maturation responses. Strikingly, we find that lipid antigen presentation by as a mechanism of immune evasion. In contrast, CD1 antigen presen- CD1b is detected within the first day after infection of DC, whereas tation occurs in the face of Mtb infection and rapid DC maturation MHCII responses are delayed or absent for up to4dafterinfection. because a pool of CD1 molecules remains available on the phagoly- These differences correlate with the presence of CD1b, but the sosome membrane that is able to acquire lipid antigens and deliver absence of MHCII at the Mtb phagosome, and are observed despite them to the plasma membrane. the successful induction of DC maturation. We suggest that microbial pathogens, such as Mtb, can uncouple the strict sequential CD1 ͉ Mycobacterium tuberculosis ͉ MHCII ͉ immune evasion relationship between MHCII antigen loading and DC maturation by delaying the formation of MHCII–peptide complexes while still rapidly inducing MHCII to traffic to the plasma membrane. he initiation of adaptive immunity relies on pathogen detection Through the precocious induction of DC maturation, invading Tvia pattern recognition receptors, including Toll-like receptors pathogens create a window of opportunity for growth and dissem- (TLR). The recognition of microbial components in dendritic cells ination, not by blocking maturation or MHCII trafficking, but by (DC) via TLR signaling provides a link between microbial detection disordering the sequence of antigen loading and MHCII trafficking and the presentation of microbial antigens to T cells (1). DC to the plasma membrane. In contrast, CD1b trafficking is indepen- maturation sequentially links antigen uptake and processing with dent of MHCII and persists as a means of initiating adaptive MHC Class II (MHCII) loading followed by delivery of MHCII– immunity in situations where MHCII presentation is rendered antigen complexes to the cell surface. Immature DCs exhibit a high dysfunctional. endocytic rate that enables efficient antigen uptake (2). They express low surface levels of MHCII, which instead is primarily Results localized to specialized late endocytic multilamellar or multivesicu- Infected DCs Present CD1-Restricted Lipid Antigens Efficiently lar lysosomes, or MHCII compartments (MIIC) (3). Exposure to Whereas MHCII-Restricted Presentation Is Delayed. Both immature TLR ligands or inflammatory cytokines induces the efficient re- and mature DCs similarly present exogenously pulsed CD1b- distribution of MHCII to the plasma membrane where it is stably restricted lipid antigen to T cells (12). To determine whether this expressed and displays peptides to T cells (4–7). DCs also express translated into rapid presentation of lipid antigens after microbial the group 1 CD1 isoforms (CD1a, CD1b, and CD1c), which present microbial lipid antigens to T cells (8). CD1b colocalizes with MHCII in the MIIC of immature DCs (9, 10). Unlike MHCII, the Author contributions: D.L.H., N.v.d.W., C.C.D., P.J.P., and M.B.B. designed research; D.L.H., N.v.d.W., N.C., D.H., L.L., S.A., M.S., and M.v.Z. performed research; S.C.K. and H.S. contrib- trafficking of CD1b is unaffected by DC maturation, and CD1b uted new reagents/analytic tools; D.L.H., N.v.d.W., N.C., S.A., P.J.P., and M.B.B. analyzed presents antigens equally well in both immature and mature DCs data; and D.L.H., N.v.d.W., P.J.P., and M.B.B. wrote the paper. (11, 12), indicating that CD1b-restricted antigen presentation is not The authors declare no conflict of interest. subject to the same strict DC maturation program as MHCII. **To whom correspondence should be addressed. E-mail: [email protected]. Many CD1-presented microbial lipids are derived from the cell edu. wall Mycobacterium tuberculosis (Mtb) (8). Mtb evades the micro- This article contains supporting information online at www.pnas.org/cgi/content/full/

cidal capabilities of antigen-presenting cells through the modulation 0804681105/DCSupplemental. IMMUNOLOGY of host signaling pathways resulting in long-term persistence in © 2008 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0804681105 PNAS ͉ August 12, 2008 ͉ vol. 105 ͉ no. 32 ͉ 11281–11286 Downloaded by guest on September 30, 2021 AB C restricted T cell line DN1. T cell recognition of CD1b–mycolic acid 1.5 8 complexes on the DC cell surface occurred 20 h after infection and 4 7 increased in magnitude over time as indicated by both IL-2 release 6 ␥ 3 1.0 and IFN- secretion by activated T cells (Fig. 1 A and B). In the 5 latter case, the T cell response to infected DCs is Ϸ4-fold greater 2 4 than background after 20 h and plateaus between 48 and 72 h of 0.5 3 infection (Fig. 1C). In addition, T cell recognition of the CD1b- 1 2 restricted antigen, glucose monomycolate (GMM), and the CD1c- 0 1 0 restricted antigen, mannosyl-phosphomycoketide (MPM), was an- 0 24 48 72 0 24 48 72 0 24 48 72 Time (h) Time (h) Time (h) alyzed. Similar to DN1, the CD1b-restricted T cell line, LDN5, and the CD1c-restricted T cell line, CD8–1, both recognized distinct D EFlipid antigens presented by DCs infected with Mtb 20–24 h after 6 250 4 infection but not at4hafterinfection (Fig. 1D and supporting 5 information (SI) Fig. S1). Thus, after Mtb infection, various CD1- 200 3 4 restricted lipid antigens are efficiently loaded onto CD1b and CD1c 150 molecules and presented to T cells. 3 2 To compare the kinetics of MHCII antigen presentation to lipid 2 100 antigen presentation, an MHCII-restricted T cell line (3F) specific 1 1 for the Ag85 complex, a major secreted antigen, was derived. 50 Surprisingly, DCs infected with Mtb failed to stimulate the 3F line 0 12 24 36 48 0 24 48 72 0 24 48 72 Time (h) Time (h) Time (h) (Fig. 1A); however, the same infected DCs stimulated the DN1 line (Fig. 1A). To determine whether the observed nonresponsive Fig. 1. T cell recognition of CD1- and MHCII-presented antigens by Mtb- phenotype was specific to the 3F line, a second HLA-DR-restricted infected DCs. For each T cell assay, Mtb-infected DCs were harvested at the T cell clone that recognizes a peptide derived from the secreted indicated time points, fixed with glutaraldehyde, and cultured with either antigen CFP10 was tested in similar assays. In contrast to the rapid CD1-restricted or MHCII-restricted T cells at a ratio of 1:1. (A) IL-2 release by responses noted for the CD1-restricted T cells during the first 48 h DN1 (CD1b-restricted) and 3F (MHCII-restricted) in response to Mtb infected DCs (ᮀ, uninfected DCs cultured with DN1; ■, infected DCs cultured with DN1; of infection, D3 recognition of HLA-DR–antigen complexes on the छ, uninfected DCs cultured with 3F; ࡗ, infected DCs cultured with 3F). (B) cell surface was minimal (ratio of IFN-␥ secretion by D3 in response IFN-␥ secretion by DN1 in response to Mtb infected DCs (E, uninfected DCs; F, to infected DCs was Ͻ2-fold greater than background). After 48 h, infected DCs). (C and D) IFN-␥ secretion by CD1b (DN1 and LDN5) restricted T a sharp increase in HLA-DR presentation to D3 became evident cells cultured with infected DCs was used to calculate a stimulation index (SI) (Fig. 1 E and F). A similar response curve was observed when T cell relative to uninfected DCs. (E and F) INF-␥ secretion and SI of D3 in response recognition was measured by IL-2 release (data not shown). There- to MHCII-restricted CFP10 derived peptides by Mtb infected DCs. PBMC iso- fore, during Mtb infection the presentation of several peptide lated from random donors were screened for their ability to induce an antigen-specific T cell response by D3 in the absence of an allogeneic mixed antigens by HLA-DR is delayed relative to lipid antigen presenta- lymphocyte response. DCs from such donors were infected with Mtb for up to tion by CD1. 72 h, fixed, and cultured with the D3 T cell clone (E, uninfected DCs; F, Mtb-infected DCs). Human DCs Undergo Maturation and Maintain CD1 Surface Expression After Mtb Infection. Next, we examined the effect of Mtb infection on DC maturation and CD1 surface expression by flow cytometry infection of DCs, CD1b-restricted antigen presentation by DCs as a possible explanation for the perceived delay in MHCII pre- infected with Mtb was analyzed. Immature human monocyte- sentation. Uninfected control cells exhibited the expected imma- derived DCs were infected with Mtb-expressing GFP, and at each ture DC phenotype indicated by the lack of cell surface expression time point after infection, DCs were fixed with glutaraldehyde and of CD83 and low to moderate surface levels of HLA-DR, CD80, tested for their ability to stimulate the mycolic acid-specific, CD1b- and CD86 (Fig. 2A). However, after infection of Mtb, DCs rapidly

HLA-DR CD80 CD83 CD86 A 100 100 100 100

80 80 80 80

60 60 60 60

40 40 40 40

20 20 20 20

0 0 0 0 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 PE CD1aPE CD1bPE CD1cPE B 100 100 100

80 80 80

60 60 60

40 40 40

20 20 20

0 0 0 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 PE PE PE

Fig. 2. Mtb-infected DCs undergo maturation and maintain stable expression of CD1 surface levels. (A) Flow cytometric analysis of DC maturation on infected (shaded histogram) or uninfected control DCs (open histogram) 20 h after infection with Mtb-GFP. Infected DCs were analyzed by gating on the GFPϩ population. (B) Flow cytometric analysis of CD1 surface expression on uninfected (dashed line histogram) and infected DCs (shaded histogram) 48 h after infection with Mtb.

11282 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0804681105 Hava et al. Downloaded by guest on September 30, 2021 Fig. 3. Differential localization and trafﬁcking of CD1b and MHCII in DCs infected with Mtb. (A) Confocal microscopy images of CD1b (red) and MHCII (blue) localization to the Mtb phagosome at 4 h, 20 h, and 48 h after infection. Inset depicts a zoomed image of the bacterial phagosome. Inset images from top to bottom are H37Rv-gfp (green), CD1b (red), MHCII (blue), and the merged image of all three channels. The arrow in each image depicts the location of the Mtb phagosome. Immunogold electron microscopy images of Mtb infected DCs stained for CD1b (B) at 1 h after infection and MHCII (C) at 4 h and 48 h after infection. Mitochondria (M), lysosomes (L), the plasma membrane (PM), and Mtb are labeled. Note the localization of gold particles in the phagosomal membrane at 4h, whereas they are then absent and appear on the plasma membrane at 48 h. (D) The labeling density of CD1b and MHCII on phagosomal membranes was determined by counting the number of gold particles per phagosome and calculating the fold change in labeling density over time (F, CD1b; E, MHCII). The percentage of phagosomes that stained positively for CD1b ranged from 28 to 52% and for MHCII 76 to 20%.

mature as indicated by the up-regulation of HLA-DR, CD80, localization to CD1b at both 4 h and 20 h after infection. The CD86, and CD83 as early as 20 h after infection. DC maturation striking localization of CD1b, LAMP1, and CD63 at all time points coincided with the redistribution of HLA-DR from the endocytic was detected as a ‘‘ring’’ that marked the phagosomal membrane. system to the plasma membrane, which is evident by confocal Importantly, the rapid and sustained colocalization correlated with microscopy staining of infected DCs (Fig. S2) and confirmed in the early presentation of CD1b antigens to T cells. subsequent experiments by electron microscopy (Fig. 3C). In To confirm the localization of CD1b on phagosomal membranes, contrast to HLA-DR, the localization of CD1b is unchanged and immunogold-labeled transmission electron microscopy was carried remains localized to the endocytic system in infected DCs (Fig. S2) out. CD1b discretely localized to the membrane of the Mtb as well as on the cell surface (Fig. 2B). phagosome within just a few hours after phagocytosis and was stably Immature DCs uniformly express high levels of CD1b and CD1c present for up to 48 h after infection (Fig. 3 B and D and Fig. S4). and exhibit a bimodal distribution of CD1a expression. Infection Delivery of CD1b to the phagosome by phagolysosome fusion was with Mtb for 48 h did not change the surface levels of CD1b and confirmed by detecting an accumulation of multivesicular lyso- CD1c but slightly increased the percentage of CD1alo cells (Fig. somes in the immediate area surrounding the phagosome (Fig. 3B) 2B). To extend these observations, the expression ratios of CD1a, and by observing LAMP1ϩ lysosomes directly fusing with the -b, and -c and HLA-DR surface levels on Mtb infected and phagosome (Fig. S5). Similar to CD1b, MHCII was present on the uninfected DCs isolated from multiple donors were determined. phagosomal membrane soon after phagocytosis (Fig. 3C). How- This analysis revealed that in the majority of donors, CD1 expres- ever, at later time points after infection, MHCII was redistributed sion was unchanged by Mtb infection, whereas HLA-DR expression to the plasma membrane (Fig. 3C and Fig. S4), consistent with the was consistently increased (Fig. S2). Collectively, these results occurrence of DC maturation. This change in MHCII localization demonstrate that DCs infected with Mtb mature phenotypically, coincided with a 9.4-fold reduction in the density of MHCII on the resulting in increased surface expression of MHCII and costimu- Mtb phagosome over the 48-h infection period (Fig. 3D). In latory molecules, but show little change in the constitutive expres- contrast, the density of CD1b on Mtb phagosomes remained stable sion pattern of CD1 molecules. for the first 20 h and then slightly increased between 20 h and 48 h (Fig. 3D), reflecting a more stable steady-state localization to CD1b and MHCII Differentially Localize to Mtb Phagosomes. In unin- bacterial phagosomes than that observed for MHCII during DC fected cells, CD1b is present in the same compartment as Lamp1, maturation. CD63, and HLA-DR in MIICs (9). Here, we examined the local- We next used inert polystyrene beads to induce phagosomes in ization of CD1b and MHCII in lysosomes and phagosomes con- DC and found that unlike Mtb phagosomes, they were positive for taining Mtb. Early (4 h) after infection, both CD1b and HLA-DR CD1b and MHCII at both 4 h and 20 h after phagocytosis (Fig. 4 were present on the phagosomal membrane; however, they did not and Fig. S6, arrowheads), suggesting that bacterial components fully colocalize with one another (Fig. 3A). Interestingly, at later must be present to induce the differential localization of CD1b and time points (20 h and 48 h), CD1b and HLA-DR differentially MHCII at phagosomes. To test this, we examined the localization localized to the phagosome: CD1b abundantly localized to the of CD1b and MHCII to phagosomes formed with polystyrene phagosome membrane, whereas HLA-DR staining on the phago- beads coated with the TLR4 agonist LPS. Similar to cells infected

somal membrane was less pronounced and in many cases not with Mtb, LPS-bead phagosomes stained positively for CD1b and IMMUNOLOGY detected (Fig. 3A). LAMP1 and CD63 (Fig. S3) exhibited a similar MHCII at4hafteruptake; however, although they remained

Hava et al. PNAS ͉ August 12, 2008 ͉ vol. 105 ͉ no. 32 ͉ 11283 Downloaded by guest on September 30, 2021 Fig. 4. Localization of CD1b and MHCII to phagosomes formed with 3-␮M polystyrene beads or LPS-coated polystyrene beads. DCs were harvested 20 h after the administration of beads and stained for MHCII (green) or CD1b (red). The arrow denotes the phagosome shown in the inset images.

CD1b-positive, they became MHCII-negative after 20 h (Fig. 4 and 20). In infected DCs, LAM was strictly detected in the Mtb Fig. S6, arrowheads). This correlated precisely with the relocaliza- phagosome, where it was found on the bacterial cell surface (Fig. tion of MHCII to the plasma membrane (Fig. 4; note green labeling 6A). A similar localization was observed in cells infected with of cell surface at 20 h). Similar localizations were observed when heat-killed bacteria (Fig. S7), suggesting that innate properties of phagosomes were formed with zymosan, which contains TLR2/6 the DC phagosome restrict lipid antigen traffic. Next, we examined agonists (Fig. S6). Thus, as TLR ligands expressed by pathogens the intracellular localization of CD1b in the same infected DCs. As regulate DC maturation, they differentially stimulate MHCII traf- observed in Fig. 3, CD1b localized to the phagosome of both live ficking from phagosome to the plasma membrane, whereas CD1b and dead bacteria in a pattern indicative of localization to the does not undergo a similar maturation-induced redistribution. phagosomal membrane (Fig. 6A;seeFig. S7). To further support the hypothesis that CD1b antigens may be Bacterial Gene Expression Is Repressed Early After Phagocytosis. loaded in phagolysosomes, the localization of the CD1 accessory Given the transient intersection of MHCII with bacterial phago- protein, Saposin C, in infected DCs was examined. Saposins are somes, the formation of peptide–MHCII complexes would be endosomal lipid transfer proteins and Saposin C has been impli- governed by the availability of processed peptide antigens during cated in the loading and transfer of lipids into CD1b (21). Like other this time. As such, the gene expression patterns of the genes lysosomal proteins, Saposin C intensely stained the Mtb phago- encoding the Ag85 complex (fbpA and fbpB) and ESAT6, which is somes at both 4 h and 20 h after infection where it colocalized with cotranscribed with CFP10 (19), were determined by quantitative CD1b (Fig. 6B and see Fig. S7). Additionally, 25–30% of Mtb PCR. The expression level for each gene in infected DC cultures phagosomes stained positively with Lysotracker-Red at 4 h, 20 h, over time was determined relative to that observed in bacteria grown in T cell media without DC. Whereas fbpA expression was moderately reduced4hafterinfection (1.5-fold), ESAT6 and fbpB ESAT6 FbpA FbpB expression were sharply reduced, 9.5- and 16.9-fold, respectively 1.5 (Fig. 5). After this initial drop, the expression of each gene steadily 1.0 0.5 increased over the remaining 3 d; remarkably, the induction of 0.0 ESAT6 expression paralleled the occurrence of MHCII-restricted -0.5 T cell recognition (Fig. 1). Thus, the early and substantial repression -1.0 of Mtb gene expression restricts the potential pool of peptide -1.5 antigens, which likely contributes to the delayed or absent MHCII 4 244872 4 244872 4 244872 antigen presentation. Time (h)

CD1b and Bacterial Lipids Colocalize Rapidly After Infection. The Fig. 5. Bacterial gene expression in intracellular Mtb. The gene expression ability of CD1b to quickly present lipid antigens after Mtb infection patterns of ESAT6, fbpA, and fbpB were analyzed after infection of DCs by and the localization of CD1b to Mtb phagosomes led us to examine qPCR using the comparative Ct method. The expression of each gene was the intracellular localization of Mtb lipids by confocal microscopy. normalized to the bacterial 16S RNA expression in each sample. The expression levels in the infected culture were then normalized to the expression levels in Mtb-infected DCs were stained with monoclonal antibody against RNA samples prepared from bacteria grown in complete T cell media. Each CD1b and antisera against lipoarabinomannan (LAM), a Mtb cell expression ratio was logarithmically transformed and plotted as the fold wall glycolipid involved in the manipulation of Mtb phagosome difference between infected cultures and cultures grown in T cell media. Bars maturation in macrophages and also a CD1b-presented antigen (13, depict the mean and standard deviation from samples run in duplicate.

11284 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0804681105 Hava et al. Downloaded by guest on September 30, 2021 the induction of lysosomal proteolysis (23), enabling the sequential degradation of invariant chain and the generation of peptides to be loaded on MHCII. Moreover, Ii proteolysis and the formation of MHCII–peptide complexes in microbial phagosomes is regulated in a TLR-dependent, phagosome-specific manner leading to MHCII antigen presentation to T cells (1). Our data suggest a model in which the tight coupling of MHCII antigen processing and loading with DC maturation may provide an opportunity for intracellular microbial pathogens to delay the initiation of MHCII-restricted T cell responses. This may occur through the rapid induction of MHCII traffic from intracellular sites to the plasma membrane (by an abundance of TLR ligands), before the synthesis of antigens in microbial phagosomes. This Fig. 6. Colocalization of CD1b, Mtb lipids, and Saposin C in phagolysosomes. model is supported by the finding that expression of several (A) Detection of LAM in the bacterial phagosome. Confocal microscopy im- bacterial antigens is significantly repressed during the initial inter- ages of Mtb (green) infected DCs stained for CD1b (blue) and LAM (red) after action between Mtb and DCs. In addition to the regulated traffic 18 h of infection. Inset images from top to bottom are Mtb-gfp (green), LAM of MHCII molecules and the formation of MHCII–peptide com- (red), CD1b (blue), and the merged image of all three channels. (B) Saposin C plexes, several features of DCs and the localization of MHCII in the localizes to the Mtb phagosome. Confocal microscopy image of Mtb infected bacterial phagosome may account for the delay in MHCII presen- DCs after 24 h of infection stained for CD1b (red) and Saposin C (blue). tation. Particularly significant is the relatively slow rate of lysosomal proteolysis exhibited by DCs compared to macrophages (24). and 48 h after infection. Thus, the localization of CD1b, CD1b- Although this delay enables DCs to sequester antigens for pro- presented antigens, and antigen loading accessory molecules to the longed presentation to T cells, the uncoupling of MHCII trafficking Mtb phagosome during the initial hours after phagocytosis corre- from the generation of antigenic peptides could enable pathogens lates with the early presentation of lipid antigens to T cells, to shorten the window during which antigens can be processed and implicating the phagolysosome as the probable antigen-loading loaded onto MHCII. compartment for CD1b. The regulation of CD1b trafficking during DC maturation does not follow a program analogous to that of MHCII. Instead, CD1b Discussion recycles from the plasma membrane through the endocytic system The question of how MHCII and CD1b antigen-presenting mole- in both immature and mature DCs (11, 12). The dichotomy of CD1b cules may function differently led us to compare lipid and peptide from MHCII in this regard permits CD1b to readily and continu- antigen presentation during a microbial infection. We find that after ously sample and present lipid antigens to T cells regardless of the Mtb infection of DCs, lipid antigen presentation by CD1b occurs DC maturation state. To this end, the rapid and intense colocal- earlier and is uncompromised compared with peptide antigen ization of CD1b on the membrane of Mtb phagosomes with CD1b presentation by MHCII. CD1b–lipid antigen complexes were de- accessory molecules such as Saposin C is noteworthy. Furthermore, tected on the cell surface within the first 24 h after infection, and a significant percentage of the Mtb phagosomes in DCs are the magnitude of the CD1b-restricted T cell response increased acidified, which is a crucial parameter for the loading of antigens with time. In contrast, the appearance of MHCII–antigen com- with long acyl chains into the CD1b antigen binding groove (10, 25). plexes that could be recognized by T cells was delayed, and in some Collectively, the rapid intersection of CD1b on acidified phago- cases, absent for up to 72 h after infection. This delay was observed somes with CD1b accessory molecules suggests that CD1–antigen despite the rapid maturation of infected DCs, and in fact, likely complexes form at this site. Interestingly, the localization of CD1b results from the discoordination of antigen availability and DC to the Mtb phagosome is a general phenomenon for phagosomes maturation, a notion that is supported by the delayed induction of because it does not require live Mtb or bacterial products. Rather, Mtb protein gene expression after infection. The difference be- it constitutes a default pathway. This is supported by the similar tween CD1b and MHCII presentation also correlated with the localization of CD1b to inert polystyrene bead phagosomes and intersection of CD1b and MHCII with Mtb after phagocytosis. excludes the likelihood that specific receptors direct cargo to Confocal and immunogold electron microscopy experiments CD1bϩ compartments. Additionally, TLR ligands in the phago- showed that both CD1b and MHCII localize to the phagosomal some have differential effects on CD1b and MHCII traffic, allowing membrane within hours after phagocytosis. However, after 20 h of the induction of precocious DC maturation by bacterial pathogens infection, CD1b still strikingly has access to the phagosome but not and resultant evasion of MHCII presentation. The rapid T cell MHCII. Instead, MHCII had moved predominantly to the plasma response to lipid antigens suggests that they are available in membrane consistent with the rapid maturation of infected DCs. bacterial cell walls early in infection. Ultimately, strong MHCII These differences were recapitulated in TLR ligand coated- peptide presentation occurs; however, this appears to require polystyrene bead phagosomes. Thus, TLR ligands induce MHCII longer periods of intracellular infection in DC and/or acquisition of trafficking to the plasma membrane, but the same pattern does not antigens by immature bystander DCs. These findings underscore occur for CD1b. We suggest that by the time peptide antigens that not only do CD1 molecules provide T cell-mediated immunity secreted by Mtb become available for processing and loading in the to a distinct chemical universe of microbial antigens, but they also phagolysosome, the majority of MHCII molecules already may provide a fundamentally separate pathway of intracellular traffick- have moved to the cell surface therefore limiting the pool of ing that expands the opportunity for cell-mediated immunity during intracellular MHCII molecules available for peptide loading. microbial infection. The trafficking of MHCII molecules and the formation of Materials and Methods MHCII–peptide complexes are tightly coupled to the maturation state of DCs. In immature DCs, MHCII is localized to the MIIC Monocyte-Derived DCs and T Cell Clones. Monocyte-derived DCs were derived as previously described (14). T cell lines were maintained in RPMI medium 1640 where it is found on internal membranes associated with invariant supplemented with 2 nM IL-2 and have been previously described: CD1b- chain (Ii). This localization is essential for peptide loading onto restricted T cell lines specific for mycolic acid (DN1) (26) and glucose monomyco- MHCII mediated by HLA-DM, which does not occur efficiently on late (GMM) (LDN5) (27), and the CD1c-restricted mannosyl-phosphomycoketide the limiting membrane of bacterial phagosomes (22). DC matura- (MPM)-specific T cell line CD8–1 (28). To derive the antigen 85 (Ag85)-specific T IMMUNOLOGY tion induces the formation of MHCII–peptide complexes (4–7) and cell line 3F, peripheral blood mononuclear cells (PBMC) were stimulated with 20

Hava et al. PNAS ͉ August 12, 2008 ͉ vol. 105 ͉ no. 32 ͉ 11285 Downloaded by guest on September 30, 2021 ␮g/ml purified Ag85 (Colorado State University, Fort Collins, CO) in RPMI medium the Quanitect Reverse Transcription kit with genomic DNA wipeout (Qiagen). An 1640 containing 5% human male AB serum (Omega Scientific), 10 mM Hepes aliquot of each reverse transcription reaction was added to PCRs containing buffer, 2 mM L-glutamine, and 100 units/100 ␮g/ml penicillin/streptomycin (Cam- Brilliant SYBR green master mix (Stratagene) and gene specific primers. The brex). After 5 d, 20 units/ml recombinant human IL-2 (Teceleukin; National Cancer sequences of the ESAT6, fbpA, and fbpB oligos have been published (30, 31), and Institute) was added to wells. Cryopreserved autologous PBMC were thawed on for 16S RNA primer sequences were 5Ј-gcgatacgggcagactagag-3Јand 5Јaaggaag- day 14, irradiated (5,000 rads), and stimulated with antigen for 2 h, washed, gaaacccacacct-3Ј. Data were collected and analyzed by using the Mx3000P qPCR added to cell lines, and after 2 d, T cells were expanded with IL-2. After two rounds system (Stratagene). ESAT6, fbpA, and fbpB expression from each infected cul- of Ag85 stimulation, Ag85-specific lines were determined by a split well assay by ture was normalized to 16S RNA expression in the same sample. Changes in measuring [H3]thymidine incorporation and IFN-␥cytokine secretion by ELISA (BD ESAT6, fbpA, and fbpB expression over time were expressed relative to the Biosciences). The CFP10-specific T cell clone was derived as previously described expression levels observed in bacterial cultures grown in complete T cell media for (29). the same period.

␮ Bacterial Strains and In Vitro Infections. The Mtb strain used in these studies was Confocal Microscopy. DCs were plated on fibronectin (20 g/ml)-coated glass coverslips and fixed with 2% formaldehyde. When used, Lysotracker Red D-99 H37Rv expressing green-fluorescent protein (GFP) and infections were per- (200 nM; Molecular Probes) was added for the final 30 minutes of incubation. formed as described (14). Polystyrene beads (3 ␮M; Polysciences, Inc.) were coated Cells were permeabilized with 0.2% Saponin and labeled with the following with LPS (100 ␮g/ml; Salmonella typhimurium)for1hat37°C or incubated for the antibodies: H4A3 (LAMP1; BD Biosciences), H5C6 (CD63; BD Biosciences), BCD1b3, same time period in PBS. Beads were washed extensively in PBS before use. IVA12, rabbit anti-LAM antisera (Daniel Clemens, University of California, Los Alexa488-conjugated Zymosan (Molecular Probes) was diluted in RPMI medium Angeles), and rabbit anti-Saponin C antisera (Greg Grabowski, Cincinnati Chil- 1640 and administered to cells at an multiplicity of infection of 1. dren’s Hospital Medical Center, Cincinnati, OH). Primary antibodies were detected with Alexa546-conjugated (Molecular Probes) or Cy5-conjugated (Jackson T Cell Assays. At each time point, DCs were washed with PBS, fixed with 0.04% Immunochemicals) antibodies. For CD1b and MHCII colocalization studies, fixed glutaraldehyde, and cultured with T cells at a 1:1 ratio. Supernatants were cells were stained with BCD1b3 and detected with goat anti-mouse Alexa546- assayed for IFN-␥ after 24 h by sandwich ELISA (Pierce Endogen). T cell prolifer- conjugated antibodies. Labeled cells were blocked with unconjugated mouse ation was assessed after3dofDC-T cell culture by measuring [3H]thymidine IgG1 antibody and subsequently labeled with either Alexa647-conjugated P3 uptake for the final 24 h of culture, and IL-2 release was assayed on HT-2 cells as (control) or Alexa647-conjugated IVA12. Confocal microscopy images were ac- described (10). quired on a Nikon C-1 confocal microscope by using EZ C1 software. Images were processed and analyzed in Adobe Photoshop CS version 8.0 by applying the Flow Cytometry. Flow cytometry was performed by using mouse monoclonal unsharp mask filter to each image. antibodies as previously described (12): P3 (IgG1 control), IgG2a control (Sigma), 10H3.9 (anti-CD1a), BCD1b3 (anti-CD1b), F10/21A3 (anti-CD1c), IVA12 (anti- Electron Microscopy. Cells were fixed and processed for cryosectioning as previ- HLADR), and anti-CD80, -CD83, and -CD86 (BD Biosciences). Primary antibodies ously described (11) and stained with the following antibodies: 10H3.9 (CD1a), were detected by using phycoerythrin-conjugated goat anti-mouse F(abЈ)2 frag- BCD1b3 (CD1b), SG520 (HLA-DR, -DP, and -DQ), H4A3 (Lamp1), and H68.4 (Trans- ments (Biosource). Data were acquired on a FACSort flow cytometer (Becton ferrin receptor; Zymed). Dickinson) and analyzed by using FlowJo (version 6.3.2). ACKNOWLEDGMENTS. Mycobacterial antigens and materials were provided RNA Isolation and Quantitative PCR (qPCR). Infected DCs or Mtb grown in by Colorado State University through the ‘‘Tuberculosis Vaccine Testing and Research Materials’’ National Institutes of Health, National Institute of Allergy complete T cell media were harvested, suspended in TRI reagent (Sigma), and and Infectious Diseases Contract (HHSN266200400091C). This work was sup- subjected to bead beating with 0.1-mm zirconium/silica beads to lyse Mtb. Lysates ported by National Institutes of Health Grants R01 AI 028973-20 and R01 AI were extracted with chloroform, and RNA was precipitated with isopropanol. 063428-04 (to M.B.B.). D.L.H. is a Damon Runyan Fellow supported by the Reverse transcription reactions were performed with 1 ␮g of total RNA by using Damon Runyan Cancer Research Foundation (DRG-1814-04).

1. Blander JM, Medzhitov R (2006) Toll-dependent selection of microbial antigens for 17. Tian T, Woodworth J, Skold M, Behar SM (2005) In vivo depletion of CD11cϩ cells delays presentation by dendritic cells. Nature 440:808–812. the CD4ϩ T cell response to Mycobacterium tuberculosis and exacerbates the outcome 2. Sallusto F, Cella M, Danieli C, Lanzavecchia A (1995) Dendritic cells use macropinocy- of infection. J Immunol 175:3268–3272. tosis and the mannose receptor to concentrate macromolecules in the major histo- 18. Jiao X, et al. (2002) Dendritic cells are host cells for mycobacteria in vivo that trigger compatibility complex class II compartment: Downregulation by cytokines and bacte- innate and acquired immunity. J Immunol 168:1294–1301. rial products. J Exp Med 182:389–400. 19. Berthet FX, Rasmussen PB, Rosenkrands I, Andersen P, Gicquel B (1998) A Mycobacte- 3. Peters PJ, Neefjes JJ, Oorschot V, Ploegh HL, Geuze HJ (1991) Segregation of MHC class rium tuberculosis operon encoding ESAT-6 and a novel low-molecular-mass culture II molecules from MHC class I molecules in the Golgi complex for transport to lysosomal filtrate protein (CFP-10). Microbiology 144(Pt 11):3195–3203. compartments. Nature 349:669–676. 20. Sieling PA, et al. (1995) CD1-restricted T cell recognition of microbial lipoglycan 4. Cella M, Engering A, Pinet V, Pieters J, Lanzavecchia A (1997) Inflammatory stimuli antigens. Science 269:227–230. induce accumulation of MHC class II complexes on dendritic cells. Nature 388:782–787. 21. Winau F, et al. (2004) Saposin C is required for lipid presentation by human CD1b. Nat 5. Turley SJ, et al. (2000) Transport of peptide-MHC class II complexes in developing dendritic cells. Science 288:522–527. Immunol 5:169–174. 6. Pierre P, et al. (1997) Developmental regulation of MHC class II transport in mouse 22. Zwart W, et al. (2005) Spatial separation of HLA-DM/HLA-DR interactions within MIIC dendritic cells. Nature 388:787–792. and phagosome-induced immune escape. Immunity 22:221–233. 7. Inaba K, et al. (2000) The formation of immunogenic major histocompatibility complex 23. Trombetta ES, Ebersold M, Garrett W, Pypaert M, Mellman I (2003) Activation of class II-peptide ligands in lysosomal compartments of dendritic cells is regulated by lysosomal function during dendritic cell maturation. Science 299:1400–1403. inflammatory stimuli. J Exp Med 191:927–936. 24. Delamarre L, Pack M, Chang H, Mellman I, Trombetta ES (2005) Differential lysosomal 8. Brigl M, Brenner MB (2004) CD1: Antigen presentation and T cell function. Annu Rev proteolysis in antigen-presenting cells determines antigen fate. Science 307:1630–1634. Immunol 22:817–890. 25. Cheng TY, et al. (2006) Role of lipid trimming and CD1 groove size in cellular antigen 9. Sugita M, et al. (1996) Cytoplasmic tail-dependent localization of CD1b antigen- presentation. EMBO J 25:2989–2999. presenting molecules to MIICs. Science 273:349–352. 26. Beckman EM, et al. (1994) Recognition of a lipid antigen by CD1-restricted alpha betaϩ 10. Sugita M, et al. (1999) Separate pathways for antigen presentation by CD1 molecules. T cells. Nature 372:691–694. Immunity 11:743–752. 27. Moody DB, et al. (1997) Structural requirements for glycolipid antigen recognition by 11. van der Wel NN, et al. (2003) CD1 and major histocompatibility complex II molecules CD1b-restricted T cells. Science 278:283–286. follow a different course during dendritic cell maturation. Mol Biol Cell 14:3378–3388. 28. Matsunaga I, et al. (2004) Mycobacterium tuberculosis pks12 produces a novel 12. Cao X, et al. (2002) CD1 molecules efficiently present antigen in immature dendritic polyketide presented by CD1c to T cells. J Exp Med 200:1559–1569. cells and traffic independently of MHC class II during dendritic cell maturation. 29. Shams H, et al. (2004) Characterization of a Mycobacterium tuberculosis peptide that J Immunol 169:4770–4777. is recognized by human CD4ϩ and CD8ϩ T cells in the context of multiple HLA alleles. 13. Vergne I, Chua J, Singh SB, Deretic V (2004) Cell biology of mycobacterium tuberculosis J Immunol 173:1966–1977. phagosome. Annu Rev Cell Dev Biol 20:367–394. 14. van der Wel N, et al. (2007) M. tuberculosis and M. leprae translocate from the 30. Shi L, North R, Gennaro ML (2004) Effect of growth state on transcription levels of phagolysosome to the cytosol in myeloid cells. Cell 129:1287–1298. genes encoding major secreted antigens of Mycobacterium tuberculosis in the mouse 15. Tailleux L, et al. (2003) Constrained intracellular survival of Mycobacterium tubercu- lung. Infect Immun 72:2420–2424. losis in human dendritic cells. J Immunol 170:1939–1948. 31. Shi L, Jung YJ, Tyagi S, Gennaro ML, North RJ (2003) Expression of Th1-mediated 16. Marino S, et al. (2004) Dendritic cell trafficking and antigen presentation in the human immunity in mouse lungs induces a Mycobacterium tuberculosis transcription pattern immune response to Mycobacterium tuberculosis. J Immunol 173:494–506. characteristic of nonreplicating persistence. Proc Natl Acad Sci USA 100:241–246.

11286 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0804681105 Hava et al. Downloaded by guest on September 30, 2021