Mini-review: Specificity and expression of CIITA, the master regulator of MHC class II genes

Authors


Abstract

The class II transactivator (CIITA) has been referred to as the "master control factor" for the expression of MHC class II (MHCII) genes. As our knowledge on the specificity and function of CIITA grows, it is becoming increasingly evident that this sobriquet is entirely justified. First, despite extensive investigations, the major target genes of CIITA remain those implicated in the presentation of antigenic peptides by MHCII molecules. Although other putative target genes have been reported, the contribution of CIITA to their expression remains indirect, controversial or comparatively minor relative to its decisive role as a regulator of MHCII and related genes. Second, the most important parameter dictating MHCII expression is by far the expression pattern of the gene encoding CIITA (MHC2TA). The vast majority of signals that activate or repress MHCII expression under physiological and pathological situations converge on one or more of the three alternative promoters that drive transcription of the MHC2TA gene. In short, with respect to its specificity and its exquisitely controlled pattern of expression, CIITA is by a long stretch the single most important transcription factor for the regulation of genes required for MHCII-restricted antigen-presentation.

Abbreviations:
CIITA:

Class II transactivator

cTEC:

Cortical TEC

LTR:

Long terminal repeat

MHCII:

MHC class II

mTEC:

Medullary TEC

pDC:

Plasmacytoid DC

RFX:

Regulatory factor X

SOCS-1:

Suppressor of cytokine signaling-1

TEC:

Thymic epithelial cell

USF-1:

Upstream regulatory factor-1

1 Introduction

MHC class II (MHCII) molecules play a pivotal role in the induction and regulation of adaptive immune responses to pathogens. They are also central to the maintenance of self-tolerance and to the breaking of this tolerance during the initiation and development of autoimmune diseases. MHCII molecules are displayed at the surface of APC where they present peptides to the TCR of CD4+ T cells. This triggers the activation and proliferation of the T cells and thus elicits an immune response directed against the antigen from which the MHCII-bound peptides were derived. MHCII molecules are also crucial for selection and maturation of CD4+ T cells in the thymus. Positive selection, which ensures the survival of T cells that carry TCR capable of recognizing self-MHC molecules, is believed to be driven by MHCII+ cortical thymic epithelial cells (cTEC) 1. On the other hand, elimination of autoreactive T cells by negative selection is driven by MHCII+ thymic DC and/or medullary thymic epithelial cells (mTEC) 13.

Constitutive expression of MHCII molecules is largely restricted to APC, namely dendritic cells (DC), B cells and macrophages. In addition, thymic epithelial cells (TEC) and activated human T cells express MHCII. In many MHCII cell types, MHCII expression can be induced by various stimuli of which IFN-γ is by far the most potent. Both constitutive and induced MHCII expression can be further modulated by additional signals. Constitutive expression in B cells and DC is for instance regulated as a function of developmental stage and can be modulated by various cytokines.IFN-γ-induced expression can also be inhibited by numerous stimuli such as TGF-β, IFN-α and IL-4.

MHCII expression is regulated mainly at the level of transcription 4, 5. The promoters of MHCII and related genes are characterized by the presence of conserved sequence elements referred to as the W (or S), X, X2 and Y boxes (Fig. 1). The X box is bound by RFX (regulatory factor X), a trimeric complex composed of RFX5 (a member of the RFX family of DNA-binding proteins), RFXANK (also called RFX-B) and RFXAP 69. The X2 box is recognized by X2BP, a complex that includes CREB 10. Finally, the trimeric NF-Y complex, composed of NF-YA, NF-YB and NF-YC, binds to the Y box 11. A number of proteins can bind the W box in vitro, including RFX 12, but none of them has been formally shown to be the functionally relevant W-box-binding protein in vivo.

All of these factors binding to the cis-regulatory elements of MHCII promoters are required for MHCII gene expression. They bind cooperatively to the promoter to form a highly stable macromolecular nucleoprotein complex referred to as the MHCII enhanceosome 13 (Fig. 1). The enhanceosome serves as a landing pad for the class II transactivator (CIITA) 13, 14. CIITA is a non-DNA-binding coactivator that serves as the master control factor for MHCII expression 4, 5.

The enhanceosome components are expressed more or less ubiquitously and thus fail to account for either the cell-type-specificity or IFN-γ-inducibility of MHCII expression. In contrast, CIITA exhibits a cell-type-specific, cytokine-inducible and differentiation-stage-specific pattern of expression that precisely parallels that of MHCII genes 4, 5 (Fig. 1). Thus, MHCII+ cells such as DC, B cells, macrophages and TEC express CIITA. In addition, CIITA expression is up-regulated by IFN-γ in most other cell types and this induced expression is down-regulated by numerous other stimuli. In all of these situations it has been firmly established that it is indeed the expression of CIITA that is responsible for driving the activation of MHCII genes. The regulation of CIITA expression occurs primarily at the level of transcription of the MHC2TA gene.

Numerous original articles have addressed the structure and mode of action of CIITA and this subject has been discussed in recent reviews 4, 5, 15, 16. Here we will instead concentrate on the specificity of CIITA, the activation of its expression in different cell types and its silencing under physiological and pathological conditions.

Figure 1.

 Activation and silencing of the MHC2TA gene. Positive and negative influences on CIITA expression are indicated in green and red, respectively. DC, B cells, macrophages, activated human T cells and TEC express CIITA constitutively. The expression of CIITA is repressed during terminal differentiation of B cells into plasma cells and of immature DC into mature DC. CIITA expression is induced by IFN-γ in most other cell types and this is repressed by many factors including cytokines (TGF-β, IL-4, IL-10 and IL-1), statins, negative feedback regulators (SOCS-1 and NO) and a growing number of pathogens, including Chlamydia, Mycobacterium BCG and MTB, Toxoplasma, varicella zoster virus, cytomegalovirus, parainfluenza virus type 3 and HIV. IFN-γ-induced CIITA expression is also repressed in fetal trophoblasts. Both constitutive and IFN-γ-induced CIITA expression are frequently silenced in tumor cell lines of various origins.

2 The specificity of CIITA

MHCII and related genes are undoubtedly the most important target genes of CIITA. In addition to the genes encoding classical MHCII molecules (HLA-DR, -DP and -DQ), CIITA activates the expression of several genes encoding accessory proteins required for MHCII-restricted antigen-presentation, namely the invariant chain (Ii), HLA-DM and HLA-DO 1720. It also contributes, albeit to a lesser degree, to classical and non-classical MHCI expression 4, 16, 21, 22. The expression of multiple genes involved in antigen presentation is thus controlled either completely or in part by CIITA. This clearly remains the primary function of CIITA. However, a series of recent papers have suggested that CIITA may also be implicated in other functions within and outside the immune system (Table 1).

IL-4, a Th2-type cytokine, was found to be aberrantly up-regulated in Th1 cells derived from CIITA knockout mice 23. In addition, CIITA was reported to be expressed in wild-type CD4+ T cell preparations under Th1 but not Th2 conditions 23, 24. Finally, transfection experiments suggested that CIITA might suppress IL-4 expression in Th1 cells by competing with NF-AT, a key IL-4 gene transcription factor, for binding to the general coactivator CBP 25. These results led to the model that CIITA could be a Th1-specific factor that functions as a repressor for IL-4 expression. However, our own studies do not support this model. We found that endogenous CIITA expression is not regulated differentially during Th1 and Th2 differentiation in human or mouse T cells 26. Furthermore, ectopic expression of CIITA in T cells does not repress IL-4 expression in Th2 cells, but results instead in a strong Th2 bias during CD4+ T cell activation 26. The explanation for the discrepancy between our results and the earlier studies remains a matter of debate.

As with the IL-4 gene, FasL expression was shown to be increased in CIITA knockout mice 24. Overexpression of CIITA in T cells was also found to lead to repression of the FasL gene 27. As in the case of the IL-4 gene, the mechanism was proposed to involve a competition between CIITA and NF-AT for binding to CBP 27.

The collagen α2 (I) gene has been reported to be repressed by CIITA 28, 29. One mechanism proposed for this repression again involves the sequestration of CBP by CIITA 28. The same mechanism may also be implicated in down-regulation of the thymidine kinase and cyclin D1 genes 28.

The promoter in the long terminal repeat (LTR) of the HIV provirus has attracted attention as a novel target of transcriptional regulation by CIITA. Expression of CIITA was reported to increase LTR promoter activity and HIV replication in fibroblast and T cell lines 30, 31. This was speculated to be relevant because CIITA is expressed in activated human T cells and macrophages, both of which are primary targets of HIV infection. However, although TGF-β is well known to inhibit the expression of CIITA, it instead stimulates HIV replication 32. Furthermore, in other studies HIV infection and LTR activity were found to be reduced in CIITA+ cell lines 33, 34. Considering these conflicting findings it is presently not clear exactly how CIITA affects transcription driven by the HIV promoter in vivo under physiological conditions.

It has recently been reported that the semaphorin receptor plexin-A1 is expressed abundantly in mature mouse DC and that this expression is dependent on CIITA 35. This observation is particularly interesting because plexin-A1 expression was found to enhance the ability of APC to promote T cell stimulation. The mechanism through which CIITA drives plexin-A1 expression remains obscure. The promoter of Plxna1, the gene coding for Plexin-A1, does not contain a characteristic W–X–X2–Y module 35, indicating that CIITA may be recruited by a mechanism distinct from that operating at MHCII genes. Alternatively, it has not been excluded that activation of the Plxna1 promoter by CIITA could be indirect. The latter could explain why Plxna1 expression is strongly increased after DC maturation 35 whereas the CIITA gene is in fact turned off during this process 36.

A recent microarray study has compared the gene expression profiles of the CIITA+ B cell line Raji and its CIITA counterpart RJ2.2.5 37. Over 40 genes whose expression appears to be modulated by CIITA were identified in this study. These genes have diverse functions, some of which could have an impact on antigen-processing, intracellular signaling or cell proliferation. However, compared with the MHCII genes, the changes in their expression levels between Raji and RJ2.2.5 were quite modest 37. Moreover, none of these genes was found to contain the characteristic W–X–X2–Y module and physical recruitment of CIITA to the regulatory regions of these genes has not been demonstrated. It thus remains to be demonstrated that they are indeed bona fide targets of CIITA.

To determine whether the potential new target genes of CIITA are relevant in vivo, we have examined the expression of several of them in mice that express CIITA ubiquitously and in mice lacking CIITA in defined cell types (unpublished data). Mice expressing CIITA ubiquitously 26 do not display a significant change in expression of any of the putative novel target genes tested, whereas the MHCII and related genes are strongly up-regulated in different organs including the liver, lung and kidney. Furthermore, in contrast to the MHCII and related genes, expression of the proposed novel target genes was found to be independent of CIITA, as no reduction in their expression was observed in B cells or IFN-γ-induced fibroblasts from mice lacking CIITA in these cells (pIII+IV knockout mice; see below). The importance of CIITA for expression of the putative new target genes is thus not evident in vivo.

In conclusion, MHCII and related genes remain the major known targets of CIITA. There is growing evidence that CIITA may influence the expression of additional genes in certain cell types and under certain experimental conditions. However, it is not clear whether these genes are indeed bona fide targets of CIITA. In several cases CIITA may affect them via an indirect mechanism, for instance by sequestering another factor such as CBP. In other cases the molecular mechanisms involved remain unknown. Finally both the microarray analysis and our in vivo experiments in transgenic and knockout mouse models indicate that the contribution of CIITA to the expression of many of these genes is relatively minor at best.

Table 1. Previously known and novel targets of CIITA
Target geneFunctionExpressionCommentsEffect of CIITAFold effecta)MechanismRef.
  1. a) The fold effect indicated is based on the assays described in foonotes b) to k).

  2. b) Absence of HLA-DR, -DP, -DQ mRNA and protein in CIITA deficient cells.

  3. c) HLA-DO mRNA levels in Raji compared with RJ2.2.5.

  4. d) MHCI promoter activity (reporter gene assay) in transiently transfected HeLa cells.

  5. e) IL-4 production (ELISA) in CD4+ T cells from CIITA knockout mice.

  6. f) FasL mRNA and surface expression in T cells from CIITA knockout mice, and in transfected Jurkat cells and T cell hybridoma.

  7. g) Collagen α2 (I) mRNA in IFN-γ-induced 2fTGH cells compared with G3A cells (defective in IFN-γ-induced CIITA expression).

  8. h) HIV expression (p24) and LTR promoter activity in Jurkat T cells stably transfected with CIITA and infected with HIV.

  9. i) HIV-LTR activity in 293T, HL3T1 and Jurkat cell lines cotransfected with CIITA and Tat.

  10. j) Plexin-A1 mRNA and protein in BM-DC from CIITA knockout mice.

  11. k) Microarray study comparing gene expression in Raji and RJ2.2.5 cell lines.

HLA-DR HLA-DP HLA-DQ Classical MHCII molecules, presentation of peptides to TCR of CD4+ T cellsAPC, TEC, IFN-γ-induced cells Expression abolished in CIITA-deficient bare lymphocyte syndrome (BLS) patients in complementation group A Activation >1000 ×b)Activation by binding to the MHCII enhanceosome 14
HLA-DM HLA-DO Ii chain Non-classical MHCII molecules, accessory proteins required for MHCII-restricted antigen presentation APC, TEC, IFN-γ-induced cells Expression reduced in CIITA-deficient BLS patients in complementation group A Activation 10 ×c)Activation by binding to the MHCII enhanceosome [17−20]
MHCI Presentation of peptides to TCR of CD8+ T cells Ubiquitous MHCI promoters contain W−X−X2−Y motif, expression reduced in BLS patients Activation 5 – 8 ×d)Activation bybinding to the MHCII enhanceosome 21, 22
IL-4 Th2 cytokine Th2 cells IL-4 expression increased in CIITA knockout mice Repression 6 ×e)Sequestration of CBP 23, 25
FasL Mediator of apoptosis T cells, immune-privileged sites FasL expression increased in CIITA knockout mice Repression 5 ×f)Sequestration of CBP 24, 27
Collagen α2(I) Extracellular matrix protein Fibroblasts osteoblasts odontoblasts Collagen α2 (I) promoter activity inhibited in CIITA+ cells Repression 5 ×g)Sequestration of CBP [28]
HIV (LTR) Proviral promoter HIV-infected cells HIV transcription enhanced in CIITA+ cells Activation 6 ×h)Unknown 30, 31
HIV (LTR) Proviral promoter HIV-infected cells HIV transcription decreased in CIITA+ cells Repression 5 ×i)Inhibition of viral transactivator Tat 33, 34
Plexin-A1 Semaphorin receptor, contributes to DC-mediated T cell stimulation Mature DC Reduction of Plexin-A1 expression in DC from CIITA knockout mice Activation 20 ×j)Unknown 35
Others Various functions Various Microarray analysis comparing wild-type and CIITA-deficient B cells Activation or repression 2 – 10 ×k)Unknown 37

3 Activation of CIITA expression

3.1 The MHC2TA gene is controlled by differential promoter activities

Expression of the gene encoding CIITA (MHC2TA) is controlled by four different promoters (pI to pIV) 38 (Fig. 2). Three of these promoters are highly conserved between the human and mouse genes (pI, pIII and pIV). However, pII has only been found in the human gene. It displays only very low transcriptional activity and its significance remains unknown. The different promoters do not share any sequence homology and are not co-regulated. They are distributed over a large (>12 kb) genomic region. Each promoter precedes a distinct first exon that is spliced alternatively to the shared downstream exons. This leads to the production of three types of transcripts (type I, type III and type IV) possessing different 5′ ends 38 (Fig. 2).

The shared second exon contains a translation initiation codon that can be used in all three types of transcript to give rise to a 1106 amino acid protein. However, the first exons of the type I and type III transcripts each contain an additional in-frame translation initiation codon. Usage of these alternative initiation codons leads to the synthesis of protein isoforms of 1207 and 1130 amino acids, respectively. The three CIITA isoforms have apparent molecular weights of 132 kDa, 124 kDa and 121 kDa (Fig. 2). All three protein variants exist in vivo36, 39.

Cell-type-specific and modulated expression of the MHC2TA gene are controlled by the differential activities of the three promoters. It is thus the sophisticated transcriptional control of the MHC2TA gene that dictates the cell-type-specific and inducible expression of MHCII genes. The specificities of the MHC2TA promoters were initially determined by examining their usage and activity in cell lines and primary cells in vitro. More recently, two new mouse strains (pIV and pIII+IV knockout mice) have allowed us to define more precisely the function of each promoter in vivo. In pIV knockout mice, pIV is deleted but transcription from pI and pIII is unaffected 40. In pIII+IV knockout mice, pIII and pIV have been excised and only pI-driven CIITA expression remains intact (unpublished data). In the following sections, we will discuss our current view of the differential usage of the CIITA promoters (Fig. 2).

Figure 2.

 Expression of the MHC2TA gene is controlled by three independent promoters having different functions: pI is active in cells of myeloid origin (conventional DC and IFN-γ-activated macrophages); pIII is used in cells of lymphoid origin (all B cell subsets, activated human T cells and pDC); pIV is essential for responsiveness to IFN-γ in non-BM-derived cells (endothelia, epithelia, fibroblasts and astrocytes) and is expressed constitutively in cTEC and mTEC. The three types of CIITA mRNA (white bars; types I, III and IV) initiated at pI, pIII and pIV encode three different protein isoforms (black bars; 121, 124 and 132 kDa). These proteins differ only at their N-terminal ends. Colored bars represent exons. The boundary between the alternative first exons and the shared downstream exons is indicated by a vertical line. The positions of translation initiation codons are indicated.

3.2 CIITA usage among different cell populations

3.2.1 IFN-γ-stimulated cells

CIITA pIV is induced by IFN-γ in most cell types 38, 4145. A 300 bp promoter-proximal region is sufficient for the IFN-γ response 38, 41, 43. This region contains a GAS element, an E box and an IRF-1-binding site, all three of which are required for induction in most cell types 38, 41, 42, 44, 46. The first two are bound cooperatively by STAT-1 and upstream regulatory factor-1 (USF-1) 41, 42, 44. STAT-1 is activated and translocated to the nucleus by the classical IFN-γ signal transduction pathway. USF-1 is a constitutively expressed member of the basic helix-loop-helix / leucine zipper family. The IRF-1-binding site of pIV is occupied by IRF-1. The synthesis of IRF-1 itself is induced by IFN-γ. This dependence on IRF-1 explains the delayed kinetics of CIITA induction by IFN-γ 47.

pIV knockout mice exhibit a highly selective abrogation of IFN-γ-induced MHCII expression on a wide variety of cells of non-hematopoietic origin, including endothelia, epithelia, astrocytes and fibroblasts 40. This provides the formal proof that pIV is indispensable for IFN-γ-inducible MHCII expression in non-professional APC.

In human fibrosarcoma and glioblastoma cell lines, a 5′ flanking sequence situated approximately 5 kb upstream of the transcription initiation site of pIII has been reported to confer IFN-γ responsiveness 42. The functional relevance of this sequence is supported by the presence of a DNaseI hypersensitive site at this position in vivo43. The putative regulatory region acts as a STAT-1-dependent and IRF-1-independent enhancer 43. In contrast to these findings in human cells, pIII does not seem to be inducible by IFN-γ in primary rat astrocytes 48 or in mouse macrophages 40, 49. Moreover, pIII is not sufficient for driving IFN-γ-induced CIITA and MHCII expression in non-BM-derived cells of the pIV knockout mice 40. There may thus be a species-specific difference in the IFN-γ-responsiveness of pIII.

3.2.2 Macrophages

The first analysis of CIITA promoter usage in macrophages was based on the examination of monocyte/macrophage cell lines. IFN-γ-induced human THP-1 cells were for instance found to expressCIITA type IV transcripts 38. However, it is now clear that pIV is not the most important inducible promoter in macrophages. IFN-γ-induced macrophages of both the pIV and pIII+IV knockout mice express MHCII molecules at normal levels, indicating that pIV is in fact not essential in these cells (40 and unpublished data).

Instead, the key promoter for sustaining CIITA and MHCII expression in IFN-γ-induced macrophages is pI rather than pIV. This is in sharp contrast to non-BM-derived cells, which remain MHCII in the pIV and pIII+IV knockout mice (see above), indicating that they are strictly dependent on pIV. In wild-type macrophages, both CIITA type I and type IV transcripts are induced by IFN-γ at early time points 49. However, at later time points, type IV mRNA declines and CIITA expression is dominated by type I mRNA, which remains elevated for long periods oftime 40, 49. Induction of pIV is thus only transient while that of pI is sustained. It remains unknown how IFN-γ activates pI. No IFN-γ-responsive sequences have been identified in the vicinity of pI. It is thus possible that IFN-γ affects pI indirectly as a consequence of macrophage activation. The answer to this question awaits further dissection of the regulatory mechanisms controlling pI.

3.2.3 TEC

MHCII molecules are constitutively expressed on cTEC and drive the positive selection of CD4+ T cells 1. Unexpectedly, cTEC of pIV knockout mice are MHCII40, 50. This loss of MHCII expression results in the abrogation of positive selection of CD4+ T cells in the thymus 40, 50. CD4+ T cell counts in the thymus and in the periphery of pIV knockout mice are reduced as strongly as in MHCII knockout mice. Thus pIV is absolutely essential for constitutive expression of CIITA in cTEC. It is also required in mTEC 50. The molecular mechanism mediating pIV activation in TEC is not known. It must be independent of the IFN-γ signaling pathway, because knockout mice lacking key components of this pathway, such as the IFN-γ receptor, STAT-1 and IRF-1, have normal MHCII expression on cTEC and unperturbed positive selection of CD4+ T cells 50.

3.2.4 B cells

Transcription of the MHC2TA gene in B cells is initiated from pIII. This was already evident from early studies using B cell lines 38, 51. Our recent analysis of the pIII+IV knockout mice has provided the final proof that pIII is indeed essential for expression of the MHC2TA gene in all B cells in vivo, including B-1 and B-2 cells in the spleen, thymus, blood and peritoneum (unpublished data). A 320 bp promoter-proximal regulatory region of pIII is sufficient for the B-cell-specific activity of pIII 38, 51. This region contains five sequence elements that have been shown by genomic footprinting experiments to be occupied in vivo in B cells 52. At least two of these elements — activation response element (ARE)-1 and ARE-2 — are critical for proper activity 52. In addition, the 5′ untranslated region of pIII seems to function as an important regulatory region in B cells 53.

3.2.5 T cells

Activated human T cells express MHCII molecules 54, 55. This expression is regulated by CIITA induced from pIII 55, 56. Mouse T cells have also been shown to express low levels of CIITA after activation 23, 24, 26, but this expression is not sufficient to induce the presence of MHCII molecules at the cell surface 26, 57. As discussed before, in the section on the specificity of CIITA, there is a controversy concerning the differential expression of CIITA in Th1 and Th2 cells. Although others have proposed that CIITA is expressed in Th1 but not Th2 cells 23, 24, we have found no difference in CIITA expression between Th1 and Th2 cells either in humans or mice 26. Certain differences have been observed between the molecular regulation of pIII in B cells and human T cells 55, 56. These reflect a difference between the pathways leading to constitutive expression in B cells and induced expression in activated T cells.

3.2.6 Melanoma cells

Some melanoma cells display an unusual constitutive expression of MHCII molecules. This aberrant pattern of expression is due to the constitutive activation of MHC2TA pIII 58, 59. It has been proposed that the 5′ flanking sequence conferring IFN-γ responsiveness to pIII in human fibrosarcoma and glioblastoma cell lines 42 is implicated in constitutive expression of pIII in these melanoma cells 58.

3.2.7 DC

In DC, pI is the promoter used predominantly for CIITA expression. CIITA type I transcripts always represent a preponderant fraction in various DC preparations including ex vivo mousesplenic and thymic DC, mouse BM-derived DC, long-term mouse DC cultures and human monocyte-derived DC (36, 38, 40, 60 and unpublished data). However, CIITA type III transcripts are also found in significant amounts in human monocyte-derived DC 36. Surprisingly, the recently discovered plasmacytoid DC (pDC) or interferon-producing cells 61 are completely devoid of CIITA, and thus MHCII expression, in mice lacking pIII (unpublished data). In agreement with this observation we have found that the expression of CIITA in pDC, unlike all other conventional DC subsets, is controlled by pIII rather than pI. Thus, pDC can be uncoupled from conventional DC in terms of their molecular regulation of CIITA and MHCII expression. This is consistent with the notion that pDC may be more closely related to the lymphoid cell lineage 62.

In summary, the pIV and pIII+IV knockout mice have provided definitive evidence that differential CIITA promoter usage does indeed play an important physiological role (Fig. 2). Cells of myeloid origin (conventional DC and macrophages) rely mainly on pI for constitutive or IFN-γ-induced CIITA expression. In cells of lymphoid origin (B cells, T cells and pDC) CIITA expression is driven almost exclusively by pIII. Finally, pIV is indispensable for IFN-γ-activated expression in non-professional APC and for expression in TEC. The three MHC2TA promoters are independent of each other and there appears to be no cross-talk between them. This point is demonstrated by the fact that individual promoters can be excised from the genome without affecting the specificity or activity of those that are retained.

4 Silencing of CIITA expression

4.1 Repression of the MHC2A gene is important in physiological and pathological situations

It has become increasingly evident over the past few years that repression of the CIITA gene plays a key role in the down-regulation of MHCII expression in various physiological and pathological situations. First, the down-regulation of MHCII expression is observed in normal cells and tissues and this is likely to be important for homeostasis of the immune system, for the regulation of immune responses and for avoiding autoimmunity. Second, there is growing evidence that tumor cells may escape recognition and elimination by the host immune response by silencing MHCII expression. Finally, several pathogens have acquired mechanisms to evade immune surveillance by inhibiting MHCII expression. As a general rule, the inhibition of MHCII expression in these situations is mediated by repression of the MHC2TA gene (Fig. 1).

4.2 Down-regulation of MHCII expression in normal cells

4.2.1 CIITA silencing during DC maturation

In response to a variety of stimuli, such as infections by bacteria or viruses, immature DC are induced to undergo profound changes in their morphology and function. Changes in the synthesis, peptide-loading and cellular localization of MHCII molecules represent key aspects of this maturation process. The density of MHCII molecules expressed at the cell surface is increased as a result of changes in the intracellular localization and stability of pre-existing MHCII proteins. In contrast, de novo synthesis of MHCII molecules is shut down. This reduction in MHCII synthesis during DC maturation is a consequence of a rapid transcriptional inactivation of the MHC2TA gene 36. This is mediated by a global repression mechanism implicating histone deacetylation over a large domain spanning the entire MHC2TA regulatory region 36.

4.2.2 CIITA silencing in plasma cells

Expression of the MHC2TA gene is actively silenced during terminal differentiation of B cells into plasma cells 63, 64. The sequence elements of pIII that are occupied in normal B cells are completely bare in plasma cells 52, 63. The human positive regulatory domain I binding factor-1 (PRDI-BF1) 65 and its mouse homologue B-lymphocyte-induced maturation protein-1 (Blimp-1) 66 have been proposed to play a crucial role in the repression of pIII in plasma cells 67, 68. PRDI-BF1/Blimp-1 is up-regulated when B cells differentiate into plasma cells 69. It has also been shown to drive, at least partially, the final differentiation of B cells into plasma cells if expressed ectopically in BCL1 lymphoma cells 69. PRDI-BF1/Blimp-1 can bind in vitro to a sequence situated in the promoter-proximal region of pIII 67, 68. However, no occupation of this site is seen in in vivo footprint experiments performed with plasma cells 52. The mechanism by which PRDI-BF1/Blimp-1 silences the MHC2TA gene thus remains to be clarified.

4.2.3 Inhibition of IFN-γ-induced CIITA expression

IFN-γ-activated expression of CIITA can be suppressed by a number of different stimuli including TGF-β, IL-1, IL-4 and IL-10 45, 7072. TGF-β markedly attenuates IFN-γ-induced CIITA expression. The inhibitory mechanism involves inhibition of MHC2TA transcription 7072. Surprisingly, TGF-β does not affect IFN-γ-induced phosphorylation of JAK-1, JAK-2 or STAT-1. Nor does it interfere with binding of STAT-1, USF-1 or IRF-1 to pIV of the MHC2TA gene 73, 74. Moreover, TGF-β even inhibits basal non-induced expression levels of pIV 42, 75. Finally, the activity of the putative IFN-γ response element situated upstream of pIII is also inhibited by TGF-β 43. Dong et al. have reported that Smad-3 is essential for the inhibitory effect of TGF-β on CIITA expression 74. IL-1, IL-4 and IL-10 have also been shown to exert an inhibitory effect on CIITA transcription in human astrocytes (IL-1) 45 and in mouse microglia (IL-4 and IL-10) 72. The role of IL-4-mediated CIITA inhibition seems to be cell-type-dependent.

IFN-γ-induced gene activation is generally a transient event. Suppressor of cytokine signaling-1 (SOCS-1) has been shown to be induced by IFN-γ and this protein negatively regulates the IFN-γ signal transduction pathway by binding to JAK-2 and inhibiting its kinase activity 76, 77. SOCS-1 can thus also suppress IFN-γ-activated expression of pIV of the MHC2TA gene 46. Similar to SOCS-1, nitric oxide — which is produced by macrophages upon IFN-γ stimulation — may act as a feedback inhibitor of MHCII synthesis by inhibiting INF-γ-induced CIITA expression 78.

Statins (HMG-CoA reductase inhibitors), which are well known for their cholesterol-lowering effect, have been shown to exhibit a number of anti-inflammatory properties. Among other effects, they repress IFN-γ-induced MHCII expression by inhibiting activation of the MHC2TA gene 79, 80. Statins may thus be of potential interest as a treatment in clinical situations where repression of MHCII-dependent T cell activation is desired. Such situations include immunosuppression following organ transplantation or autoimmune diseases such as multiple sclerosis and rheumatoid arthritis 7981.

4.2.4 CIITA silencing in trophoblasts

Fetal trophoblasts lack expression of MHCII molecules, both constitutively and after exposure to IFN-γ. The absence of MHCII molecules on trophoblasts is thought to play a critical role inpreventing rejection of the fetus by the maternal immune system. The inability of trophoblasts to express MHCII genes is primarily due to the lack of CIITA expression 82, 83. It has been shown that pIV is hypermethylated at CpG dinucleotides in trophoblast cell lines and primary trophoblasts 84, 85. This has been shown toblock the activation of pIV by inhibiting the binding of STAT-1 and IRF-1 as well as the ensuing chromatin remodeling 86. Recently, an additional intriguing mechanism has been proposed to be implicated in MHCII silencing in trophoblasts 87. It involves a trophoblast-derived non-coding RNA that is able to suppress IFN-γ-induced CIITA expression throughan inhibitory domain in pIV.

4.3 CIITA silencing in tumor cells

MHC molecules play a pivotal role in presenting tumor-derived antigens and hence in activating and regulating antitumor immune responses 88, 89. Consequently, one strategy employed by malignant cells for evading recognition and elimination by the immune system involves the loss or down-regulation of MHC expression 88, 89. The partial or complete loss of MHCI expression is frequently observed because cytotoxic CD8+ T cells, which recognize antigenic peptides presented by MHCI molecules, constitute the primary effector cells mediating tumor rejection 90. However, efficient and long-lasting antitumor immunity requires help provided by CD4+ cells during both the priming and effector phases of the antitumor response 91, 92. The loss of MHCII expression on malignant cells is thus also a commonly observed escape strategy.

The loss of constitutive MHCII expression is observed in tumor cells of hematopoietic origin, particularly in B and T cell malignancies 9395. Moreover, theinability to induce MHCII expression in response to IFN-γ is often associated with tumor cells of non-hematopoietic origin 58, 95100. Thereis growing evidence that this inability to express MHCII results from epigenetic silencing of the MHC2TA gene 95102. The regulatory regions of the MHC2TA gene have been found to be hypermethylated at CpG dinucleotides in MHCII T cell leukemias, B cell lymphomas and various tumor cells that are unable to express MHCII upon exposure to IFN-γ, including teratocarcinoma, choriocarcinoma, neuroblastoma, erythroleukemia and small cell lung cancer 95, 97, 100102. Histone deacetylation rather than DNA hypermethylation has been implicated in silencing of MHC2TA expression in several squamous cell carcinoma cell lines 103.

4.4 Repression of CIITA expression by pathogens

Pathogens have developed a wide variety of strategies to escape immune surveillance by their hosts 104, 105. In order to inhibit the establishment of a protective immune response, several bacteria and viruses down-regulate MHCII expression and thus prevent the activation of specific CD4+ T cells. They achieve this by interfering with the function or expression of CIITA.

The intracellular bacterium Chlamydia down-regulates CIITA expression by inducing the degradation of USF-1, a ubiquitous factor that is required for the activation of pIV of the MHC2TA gene 106. Infections with Mycobacterium bovis bacillus Calmette-Guérin (BCG), Mycobacterium tuberculosis (MTB) or Toxoplasma gondii also down-regulate CIITA expression, but the precise mechanisms that are involved remain poorly defined 107109. The MTB 19-kDa lipoprotein inhibits IFN-γ-induced MHCII expression in macrophages by preventing the induction of CIITA type I and type IV mRNA. Interestingly, it seems to exert its negative effect on IFN-γ signaling by inhibiting IRF-1, independently of SOCS-1 and STAT-1 108.

Varizella zoster virus, human cytomegalovirus (CMV) and human parainfluenza virus type 3 (HPIV3) also inhibit IFN-γ-induced CIITA expression 110113. Varizella zoster virus does so by interfering with the IFN-γ signaling pathway. It blocks STAT-1α and JAK2 expression and therefore transcription of the downstream IRF-1 and CIITA genes 110. Infection with CMV can direct JAK-1 to the proteasome for degradation or affect the IFN-γ signaling pathway downstream of STAT phosphorylation and nuclear translocation 111, 112. The inhibitory effect of HPIV3 on IFN-γ-induced CIITA expression is due to a defect downstream of STAT-1 activation, but the precise mechanism remains unclear 113. Finally, as mentioned earlier, the HIV Tat protein can inhibit MHCII expression by interfering with the function of CIITA in HIV-infected or Tat-transfected fibroblasts and T cell lines 34, 114. Tat does this by competing with CIITA for binding to cyclin T1, a component of the transcriptional elongation complex P-TEFb 34, 114.

Taken together, these findings demonstrate that many pathogens have acquired a means to target the IFN-γ signal transduction pathway. Among other consequences, this results in the inhibition of IFN-γ-induced CIITA and MHCII expression, which may favor the escape from immune surveillance and thus facilitate the establishment of persistence. It will be of great interest to dissect in greater detail the strategies used by pathogens to interfere with CIITA and MHCII expression. This may contribute to the design of new approaches for fighting these pathogens.

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

Ancillary