Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Brief communication: Long-term absence of Langerhans cells alters the gene expression profile of keratinocytes and dendritic epidermal T cells

  • Qingtai Su,

    Roles Formal analysis, Investigation, Methodology, Validation, Visualization

    Affiliation Baylor Scott & White Research Institute, Baylor Institute for Immunology Research, Dallas, Texas, United States of America

  • Aurélie Bouteau,

    Roles Data curation, Methodology

    Affiliations Baylor Scott & White Research Institute, Baylor Institute for Immunology Research, Dallas, Texas, United States of America, Baylor University, Institute of Biomedical Studies, Waco, Texas, United States of America

  • Jacob Cardenas,

    Roles Data curation, Formal analysis, Methodology, Software, Validation, Visualization

    Affiliation Baylor Scott & White Research Institute, Dallas, Texas, United States of America

  • Balaji Uthra,

    Roles Data curation, Formal analysis

    Affiliation Baylor Scott & White Research Institute, Dallas, Texas, United States of America

  • Yuanyaun Wang,

    Roles Methodology

    Affiliation Baylor Scott & White Research Institute, Dallas, Texas, United States of America

  • Cynthia Smitherman,

    Roles Methodology

    Affiliation Baylor Scott & White Research Institute, Dallas, Texas, United States of America

  • Jinghua Gu,

    Roles Data curation, Formal analysis, Methodology

    Affiliation Baylor Scott & White Research Institute, Dallas, Texas, United States of America

  • Botond Z. Igyártó

    Roles Conceptualization, Funding acquisition, Resources, Supervision, Writing – original draft, Writing – review & editing

    botond.igyarto@jefferson.edu

    Affiliations Baylor Scott & White Research Institute, Baylor Institute for Immunology Research, Dallas, Texas, United States of America, Thomas Jefferson University, Department of Microbiology and Immunology, Philadelphia, Pennsylvania, United States of America

Abstract

Tissue-resident and infiltrating immune cells are continuously exposed to molecules derived from the local cells that often come in form of secreted factors, such as cytokines. These factors are known to impact the immune cells’ biology. However, very little is known about whether the tissue resident immune cells in return also affect the local environment. In this study, with the help of RNA-sequencing, we show for the first time that long-term absence of epidermal resident Langerhans cells led to significant gene expression changes in the local keratinocytes and resident dendritic epidermal T cells. Thus, immune cells might play an active role in maintaining tissue homeostasis, which should be taken in consideration at data interpretation.

Introduction

The effect of tissue environment on immune cells has been widely studied. Tissue microenvironment through an unknown mechanism is capable of shaping the chromatin landscapes of macrophages, which results in tissue-specific functions of macrophages[1]. Dendritic cell (DC) populations in different tissues display tissue-specific diversity and functions[2], and thus, it is anticipated that the close communication between DCs and the tissue microenvironment might endow them with functional diversity and plasticity. It is well documented that keratinocytes (KCs) for example can regulate immune responses by affecting epidermal resident, antigen presenting Langerhans cells’ (LCs) biology through secretion of cytokines and other factors[3]. LCs are a subset of DCs that are radiation-resistant and reside in the epidermis, where they are tightly attached to the surrounding keratinocytes[4]. LCs participate in promotion of self-tolerance, anti-fungal immunity, skin immunosurveillance, and protective humoral immune responses[5]. In this study, we tested the idea whether long-term absence of an immune cell, LCs from the epithelial environment, affects the constituent KCs and the resident dendritic epidermal T cells (DETCs). Here we show, to our knowledge, first-time evidence that long-term absence of an immune cell can lead to significant changes in the local cells and to an altered tissue microenvironment.

Materials and methods

Mice

huLangerin-DTA (LC-/-) mice have been previously described[6]. All experiments were performed with 8 weeks old littermate-controlled mice. Mice were housed in microisolator cages and fed autoclaved food. The Baylor Institutional Care and Use Committee specifically approved this study.

Flow cytometry and cell sorting

Single-cell suspensions of flank skin were obtained and stained as previously described[7]. Briefly, mice were sprayed with 70% ethanol and the flank skin area shaved using Personna razor blades. The shaved area was harvested, the subcutaneous fat scraped away using forceps and ~1 cm wide strips were prepared using razor blades. The resulting skin strips were floated on Trypsin-GNK solution for ~1.5 hours in CO2 incubators. After the incubation the epidermal sheets were separated from the dermis using forceps and further incubated in DNase solution for 15 minutes in water bath with occasional vortexing. The resulting cell suspensions were then filtered through 40 μm cell strainers, washed and stained for flow cytometer. Cell suspensions were directly labeled with fluorochrome-conjugated antibodies for cell surface markers: MHC-II, CD45, langerin, c-kit and fixable Viability Dye. KCs (live/singlets, MHC-II-, CD45-, langerin-, c-kit-) and DETCs (live/singlets, MHC-II-, CD45+, langerin-, c-kit-) were sorted on flow cytometer (S1 Fig)[8]. Stringent doublets discrimination and live/dead gating were used to exclude possible contaminants and dead cells, respectively.

RNA preparation

Total RNA was isolated from cell lysates using the RNeasy Micro Kit (Qiagen) including on-column DNase digestion. Total RNA was analyzed for quantity and quality using the RNA 6000 Pico Kit (Agilent).

Sequencing library preparation

Poly-A enriched Next-Generation Sequencing (NGS) library construction was performed using the KAPA mRNA Hyper Prep Kit (KAPA Biosystems) using 50ng of input total RNA according to manufacturer’s protocol using 16 amplification cycles. Quality of the individual libraries was assessed using the High Sensitivity DNA Kit (Agilent). Individual libraries were quantitated via qPCR using the KAPA Library Quantification Kit, Universal (KAPA Biosystems) and equimolar pooled. Final pooled libraries were sequenced on an Illumina NextSeq 500 with paired-end 75 base read lengths.

Bioinformatics analysis

Raw sequencing reads assessed for quality using FastQC software[9]. The adapters were trimmed and low-quality reads (< 20) were filtered using cutadapt[10]. Reads were aligned to the mouse reference genome (GRCm38) using hisat2. Aligned SAM (Sequence Alignment Map) files were converted to BAM (Binary Alignment Map) format using samtools[11] and featureCounts[12] was used to quantify total number of counts for each gene.

RNA-seq analysis

Transcripts with low expression, i.e., count-per-million (CPM) > 1 in less than two samples, were removed from downstream analysis, leaving 14,964 transcripts. Differential gene expression (DGE) analysis was performed using DESeq2[13] and comparisons were made between LC-/- and WT within DETC and KC cell populations.

Pathway and gene ontology analysis

Two approaches to pathway and Gene Ontology (GO) analysis were used[14]. The Database for Annotation, Visualization and Integrated Discovery (DAVID)[15] was used for functional annotation of significantly regulated genes based on false discovery rate (FDR) < .05 and fold change (FC) cut-off of 1.5 for each comparison. Additionally, a fast implementation of pre-ranked Gene Set Enrichment Analysis (FGSEA) using the fgsea R package[16,17] was performed on KEGG (Kyoto Encyclopedia of Genes and Genomes) and GO gene sets obtained from the Molecular Signatures Database v6.2 (MSigDB)[18].

RNA-seq data visualization

Counts were normalized using the median-of-ratios method[19] and log2 transformed for data visualization. Principal component analysis (PCA) and hierarchical clustering were performed using the R. The transcripts of all heatmaps were hierarchically clustered using Euclidean distance and complete linkage function. Heatmaps were plotted using the NMF (Non-negative Matrix Factorization) package[20], while PCA and volcano plots were made using ggplot2[21].

Results

Long-term absence of LCs leads to gene expression changes in KCs and DETCs

To determine the possible effect of the absence of LCs on the cells of the epidermis, we took advantage of the huLangerin-DTA (diphtheria toxin subunit A) mice (hereafter LC-/-), which lack LCs starting from birth[6]. Thus, for these mice, KCs and DETCs develop, differentiate, and function in the absence of differentiated LCs. Epidermal cells suspensions were generated from a cohort of LC-/- mice, along with littermate wild type (WT) controls (Fig 1A). After staining with specific markers, the KCs and DETCs were sorted using flow cytometer (S1 Fig) and RNA-sequencing performed. Unsupervised PCA of the expression data revealed that KCs and DETCs, which developed in the absence of LCs, clearly clustered away from their WT counterparts (Fig 1B). We identified 1220 up- and 537 downregulated genes in KCs, while in DETCs, we identified 880 up- and 214 downregulated genes using a false discovery rate (FDR) <0.05 (Fig 1C). Out of the upregulated genes, 348 (19.9%) were common between KCs and DETCs, while 22 genes (3.02%) were commonly downregulated (Fig 1C). Next, we performed hierarchical clustering of differentially expressed genes with at least 2-fold change and plotted heatmaps to show the distinct patterns of up- and downregulated genes in KCs and DETCs (Fig 1D).

thumbnail
Fig 1. Absence of LCs leads to gene expression changes in KCs and DETCs.

(A). Experimental flow. KCs and DETCs were flow sorted from LC-deficient (LC-/-) and littermate WT controls and RNA-seq. performed. The resulting data were then subjected to bioinformatic analyses. (B). Principal component analysis of the RNA-seq. data. Each dot represents a separate animal. (C). The overlaps between the genes that were up- (top) or downregulated (bottom) in the absence of LCs in KCs and DETCs are presented in forms of Venn diagrams; FDR<0.05. (D). Heatmap presentation of the genes that showed two-fold changes between LC-/- and WT mice. KCs (left) and DETC (right). FDR<0.05.

https://doi.org/10.1371/journal.pone.0223397.g001

We used color-coded volcano plots to better capture and visualize the common and cell specific changes in gene expression (Fig 2A; for DGEs please see S1 and S2 Files). We observed that in the absence of LCs nerve growth factor (Ngf) was the most highly upregulated gene in both KCs and DETCs. NGF is part of the neurotrophin family and is involved in the differentiation and survival of neuronal cells[22], which suggest that LCs might directly or indirectly regulate nerve homeostasis in the epidermis[23].

thumbnail
Fig 2. LCs have common and cell specific effects on KCs’ and DETCs’ biology.

(A). Common and cell specific gene expression changes are presented in form of color-coded volcano plots. KCs (left) and DETCs (right). (B). Flow cytometry confirmation of the RNA-seq. data on protein levels. Each dot represents a separate mouse. Two tailed Student’s t-test. *p<0.05, ***p<0.001. (C). The overlap between up- or downregulated regulated KEGG pathways in KCs and DETCs from LC-/- mice are presented in forms of Venn diagrams. (D). Selected KEGG pathways altered in KCs and DETCs in the absence of LCs are depicted based on normalized enrichment scores (NES). FDR<0.05.

https://doi.org/10.1371/journal.pone.0223397.g002

Out of highly published, skin-relevant molecules, thymic stromal lymphopoietin (Tslp) was upregulated in the KCs (trend in DETCs) in the absence of LCs, which confirms the findings recently published by Lee et al.[24]. TSLP is a known regulator of the Th2 responses and it is also needed for mast cell homeostasis[25,26]. Dysregulated TSLP production by KCs, in the absence of LCs, could have contributed to altered IgE levels[27] and increased mast cell numbers observed in the LC-/- mice (FVB background). The KCs, among others, showed downregulation of the MHC-II pathway genes H2-Eb1 (Fig 2A and 2B), Cd74 (invariant chain; also downregulated in DETCs;), and Cyp17a1, a member of the P450 cytochrome family involved in carcinogen metabolism[28,29]. MHC-II and CD74 are highly expressed in LCs, and in light of our recent observation that LCs is involved in bidirectional mRNA exchange with KCs and DETCs[8], it is plausible that the observed decrease of the MHC-II pathway genes in KCs and DETCs might be due to the absence of LCs as a source of these mRNAs rather than changes in transcription.

Lack of LCs might affect DETCs’ biology and homeostasis

We discovered that DETCs downregulated the Th17 pathway associated molecules, including Rorc transcription factor, Il17rb receptor, and Il17a and Il17f cytokines (Fig 2A and 2B). The IL-17 pathway plays an important role in the DETC’s innate immune function to fight bacterial infections[30]. More interestingly, we observed that DETCs showed lower expression of the γ/δ TCRs (Trdv4 and Tcrg-V5) and upregulation of TCR alpha chains (Trav16 and Trav13-4-dv7). Transcription factors and other molecules that regulate the development, differentiation, and homeostasis of DETCs, such as Sox13, Blk, and Il7r[31], were also downregulated. Thus, these data suggest that LCs might directly or indirectly regulate DETCs’ biology and homeostasis, and could contribute to maintain their identity/fitness in the epidermis/periphery in the absence of the thymic environment. Contrary to our findings, a previous report presented data supporting that LCs are not required for DETCs’ homeostasis and function[32]. In this manuscript the authors focused on whether the absence of LCs affects DETC’s density, steady state and activation markers, cytotoxic activity and reaction to skin injury. The manuscript reported no changes in the limited number of markers studied nor defect in DETCs’ cytotoxic activity. Our data confirm their findings on mRNA levels (please see DGE supplementary data). However, they did report a significant increase in the size of activated cell region in the LC deficient mice 48 hours post skin injury. Thus, these data collectively support that LCs can specifically affect certain aspects of DETC’s biology.

The absence of LCs affected a variety of different cellular pathways in KCs and DETCs

To gain a broader picture about the effect of the absence of LCs on KCs and DETCs, we performed KEGG pathway analysis on the expression data. We present data of significantly altered pathways using FDR < 0.05. We observed significant overlap of pathways upregulated by KCs and DETCs, but very minimal overlap of downregulated pathways (Fig 2C). The commonly upregulated pathways included different forms of cell adhesions (focal, adherent and tight)-, ribosome and RNA biogenesis-, autophagy-, bacterial invasion/infection-, MAPK- and ErbB signaling pathways (Fig 2D). Alterations in adhesion molecules and the ErbB signaling pathway in KCs, in the absence of LCs, were recently reported[24,33]. The downregulated pathways showed considerably less overlap between these two cell types and included some of the amino acid degradation pathways (Fig 2D). KCs showed distinct dysregulation (mostly downregulation) of various metabolic pathways (sugar, protein, fatty acids, hormones, drug, xenobiotics etc.), while DETCs presented with alterations in TGF-β-, Hippo-, oxidative phosphorylation-, citrate cycle-, lipid metabolism-, Staphylococcus aureus infection- etc. pathways (Fig 2D). Thus, these data suggest that LCs might have common and cell-specific effects on KCs’ and DETCs’ biology.

Discussion

Here we bring experimental evidence that long-term absence of LCs leads to gene expression changes in KCs and DETCs. The significant changes discovered by pathway analysis also suggest that KCs’ and DETCs’ biology and hemostasis are likely affected. Further studies are indeed needed to confirm the observed changes and their consequences. It will also be important to determine which LC-derived factors play role in the epidermal homeostasis. Our preliminary IPA Upstream Regulator Analysis identified a list of potential regulators, including cytokines, growth factors, membrane proteins and enzymes (please see S3 and S4 Files), known or anticipated to be expressed by LCs.

It will be critical to determine whether acute depletion of LCs would lead to similar changes in the local environment, and if yes, how long the LCs have to be depleted for, before their absence is noticed. These studies probably should be carried out side-by-side using co-housed (or intercrossed) huLangerin-DTR and muLangerin-DTR mice treated or not with DT. Performing the experiments this way, should help to minimalize the possible effect of colony-related microbiome differences, that alone, could lead to gene expression changes in the local cells. Furthermore, these experiments could also help to reconcile the discrepancies reported -especially fueled by a single study[34]- using these two inducible systems[35]. Reconstitution of irradiated LC-/- mice with WT bone marrow or hu/muLangerin-DTR injected with DT and left to recover the depleted LCs, could also test whether the repopulating LCs can substitute for yolk sac/fetal liver-derived LCs[36] and normalize the epidermal homeostasis.

To our knowledge we show for the first time that long-term absence of an immune cell can lead to significant changes in the local cells and to altered tissue environment. The effect of local cells on resident immune cells is very much appreciated by the immunologist, however, our findings support the idea that the resident immune cells are not mere passive receivers, but rather play an active and indispensable role in maintaining tissue homeostasis. It is expected that our findings will not be limited to LCs or LC-derived factors, and that the long-term absence of other DCs, DC molecules or immune cells would lead to specific changes in the local environment. Thus, studies using constitutive immune-cell knockouts, including LC-/- mice, in which the immunological changes and outcomes were directly attributed to the absence of a specific immune cell, might have to be reassessed[5,6,3739].

Acknowledgments

We thank the animal facility, flow core and genomics core for their help and support.

References

  1. 1. Lavin Y, Winter D, Blecher-Gonen R, David E, Keren-Shaul H, Merad M, et al. Tissue-resident macrophage enhancer landscapes are shaped by the local microenvironment. Cell. 2014. pmid:25480296
  2. 2. Pakalniškytė D, Schraml BU. Tissue-Specific Diversity and Functions of Conventional Dendritic Cells. Advances in Immunology. 2017. pmid:28413024
  3. 3. Clayton K, Vallejo AF, Davies J, Sirvent S, Polak ME. Langerhans cells-programmed by the epidermis. Frontiers in Immunology. 2017. pmid:29238347
  4. 4. Romani N, Holzmann S, Tripp CH, Koch F, Stoitzner P. Langerhans cells—Dendritic cells of the epidermis. APMIS. 2003. pmid:12974775
  5. 5. Kaplan DH. Ontogeny and function of murine epidermal Langerhans cells. Nature Immunology. 2017. pmid:28926543
  6. 6. Kaplan DH, Jenison MC, Saeland S, Shlomchik WD, Shlomchik MJ. Epidermal Langerhans cell-deficient mice develop enhanced contact hypersensitivity. Immunity. 2005. pmid:16356859
  7. 7. Kashem SW, Kaplan DH. Isolation of Murine Skin Resident and Migratory Dendritic Cells via Enzymatic Digestion. Curr Protoc Immunol. 2018. pmid:30040218
  8. 8. Su Q, Igyártó BZ. Keratinocytes Share Gene Expression Fingerprint with Epidermal Langerhans Cells via mRNA Transfer. J Invest Dermatol. 2019. pmid:31129057
  9. 9. Andrews S, Krueger F, Seconds-Pichon A, Biggins F, Wingett S. FastQC. A quality control tool for high throughput sequence data. Babraham Bioinformatics. Babraham Inst. 2015.
  10. 10. Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet.journal. 2011.
  11. 11. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009. pmid:19505943
  12. 12. Liao Y, Smyth GK, Shi W. FeatureCounts: An efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics. 2014. pmid:24227677
  13. 13. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014. pmid:25516281
  14. 14. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. Gene ontology: Tool for the unification of biology. Nature Genetics. 2000. pmid:10802651
  15. 15. Huang DW, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc. 2009. pmid:19131956
  16. 16. Sergushichev AA. An algorithm for fast preranked gene set enrichment analysis using cumulative statistic calculation. bioRxiv. 2016.
  17. 17. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005. pmid:16199517
  18. 18. Liberzon A, Subramanian A, Pinchback R, Thorvaldsdóttir H, Tamayo P, Mesirov JP. Molecular signatures database (MSigDB) 3.0. Bioinformatics. 2011. pmid:21546393
  19. 19. Anders S, Huber W. Differential expression analysis for sequence count data. Genome Biol. 2010. pmid:20979621
  20. 20. Gaujoux R, Seoighe C. A flexible R package for nonnegative matrix factorization. BMC Bioinformatics. 2010. pmid:20598126
  21. 21. Ginestet C. ggplot2: Elegant Graphics for Data Analysis. J R Stat Soc Ser A (Statistics Soc. 2011.
  22. 22. Skaper SD. Nerve growth factor: a neuroimmune crosstalk mediator for all seasons. Immunology. 2017. pmid:28112808
  23. 23. Boulais N, Misery L. The epidermis: A sensory tissue. European Journal of Dermatology. 2008. pmid:18424369
  24. 24. Lee HJ, Kim TG, Kim SH, Park JY, Lee M, Lee JW, et al. Epidermal Barrier Function Is Impaired in Langerhans Cell-Depleted Mice. J Invest Dermatol. 2019. pmid:30448384
  25. 25. Han NR, Oh HA, Nam SY, Moon PD, Kim DW, Kim HM, et al. TSLP induces mast cell development and aggravates allergic reactions through the activation of MDM2 and STAT6. J Invest Dermatol. 2014. pmid:24751726
  26. 26. Leyva-Castillo JM, Hener P, Michea P, Karasuyama H, Chan S, Soumelis V, et al. Skin thymic stromal lymphopoietin initiates Th2 responses through an orchestrated immune cascade. Nat Commun. 2013. pmid:24284909
  27. 27. Nakajima S, Igyártó BZ, Honda T, Egawa G, Otsuka A, Hara-Chikuma M, et al. Langerhans cells are critical in epicutaneous sensitization with protein antigen via thymic stromal lymphopoietin receptor signaling. J Allergy Clin Immunol. 2012. pmid:22385635
  28. 28. Lewis JM, Bürgler CD, Freudzon M, Golubets K, Gibson JF, Filler RB, et al. Langerhans Cells Facilitate UVB-Induced Epidermal Carcinogenesis. J Invest Dermatol. 2015. pmid:26053049
  29. 29. Modi BG, Neustadter J, Binda E, Lewis J, Filler RB, Roberts SJ, et al. Langerhans cells facilitate epithelial DNA damage and squamous cell carcinoma. Science (80-). 2012. pmid:22223807
  30. 30. Cho JS, Pietras EM, Garcia NC, Ramos RI, Farzam DM, Monroe HR, et al. IL-17 is essential for host defense against cutaneous Staphylococcus aureus infection in mice. J Clin Invest. 2010. pmid:20364087
  31. 31. Kang J, Malhotra N. Transcription Factor Networks Directing the Development, Function, and Evolution of Innate Lymphoid Effectors. Annu Rev Immunol. 2015. pmid:25650177
  32. 32. Taveirne S, De Colvenaer V, Van Den Broeck T, Van Ammel E, Bennett CL, Taghon T, et al. Langerhans cells are not required for epidermal V 3 T cell homeostasis and function. J Leukoc Biol. 2011. pmid:21486908
  33. 33. Shipman WD, Chyou S, Ramanathan A, Izmirly PM, Sharma S, Pannellini T, et al. A protective Langerhans cell keratinocyte axis that is dysfunctional in photosensitivity. Sci Transl Med. 2018. pmid:30111646
  34. 34. Bobr A, Olvera-Gomez I, Igyarto BZ, Haley KM, Hogquist KA, Kaplan DH. Acute Ablation of Langerhans Cells Enhances Skin Immune Responses. J Immunol. 2010. pmid:20855870
  35. 35. Clausen BE, Stoitzner P. Functional specialization of skin dendritic cell subsets in regulating T cell responses. Front Immunol. 2015. pmid:26557117
  36. 36. Hoeffel G, Wang Y, Greter M, See P, Teo P, Malleret B, et al. Adult Langerhans cells derive predominantly from embryonic fetal liver monocytes with a minor contribution of yolk sac-derived macrophages. J Exp Med. 2012. pmid:22565823
  37. 37. Obhrai JS, Oberbarnscheidt M, Zhang N, Mueller DL, Shlomchik WD, Lakkis FG, et al. Langerhans cells are not required for efficient skin graft rejection. J Invest Dermatol. 2008. pmid:18337832
  38. 38. Igyarto BZ, Jenison MC, Dudda JC, Roers A, Müller W, Koni PA, et al. Langerhans Cells Suppress Contact Hypersensitivity Responses Via Cognate CD4 Interaction and Langerhans Cell-Derived IL-10. J Immunol. 2009. pmid:19801524
  39. 39. Igyártó BZ, Haley K, Ortner D, Bobr A, Gerami-Nejad M, Edelson BT, et al. Skin-Resident Murine Dendritic Cell Subsets Promote Distinct and Opposing Antigen-Specific T Helper Cell Responses. Immunity. 2011. pmid:21782478