Nathalie Labrecque

An essential function of the mitogen‐activated protein kinase Erk2 in mouse trophoblast development

The closely related mitogen‐activated protein kinase isoforms extracellular signal‐regulated kinase 1 (ERK1) and ERK2 have been implicated in the control of cell proliferation, differentiation and survival. However, the specific in vivo functions of the two ERK isoforms remain to be analysed. Here, we show that disruption of the Erk2 locus leads to embryonic lethality early in mouse development after the implantation stage. Erk2 mutant embryos fail to form the ectoplacental cone and extra‐embryonic ectoderm, which give rise to mature trophoblast derivatives in the fetus. Analysis of chimeric embryos showed that Erk2 functions in a cell‐autonomous manner during the development of extra‐embryonic cell lineages. We also found that both Erk2 and Erk1 are widely expressed throughout early‐stage embryos. The inability of Erk1 to compensate for Erk2 function suggests a specific function for Erk2 in normal trophoblast development in the mouse, probably in regulating the proliferation of polar trophectoderm cells.

How Much TCR Does a T Cell Need

Kinetic features of TCR:MHC/peptide interactions dictate their outcome in vitro, some important parameters of which include the number of molecules engaged and the duration of engagement. We explored the in vivo significance of these findings in transgenic mice expressing TCRs in a quantitatively and temporally controlled manner. As anticipated, reduced TCR levels resulted in attenuated reactivity, but response thresholds were substantially lower than expected-at as low as 1/20th the normal TCR numbers and with no indication of phenotypic skewing at suboptimal levels. We also studied survival of T lymphocytes stripped of their TCRs. Unlike B cells, T cells lacking antigen receptors did not die precipitously; instead, populations decayed gradually, just as previously reported in the absence of MHC molecules.

Notch signaling regulates PD-1 expression during CD8(+) T-cell activation.

Programmed cell death 1 (PD‐1) is an inhibitory receptor involved in T‐cell activation, tolerance and exhaustion. Little is known on how the expression of PD‐1 is controlled during T‐cell activation. Recent studies demonstrated that NFATc1 and IRF9 regulate Pdcd1 (PD‐1) transcription and that T‐bet acts as a transcriptional repressor. In this study, we have investigated the role of the Notch signaling pathway in PD‐1 regulation. Using specific inhibitors of the Notch signaling pathway, we showed decreased PD‐1 expression and inhibition of Pdcd1 transcription by activated CD8+ T cells. Chromatin immunoprecipitation further showed occupancy of the Pdcd1 promoter with RBPJk and Notch1 intracellular domain at RBPJk‐binding sites. Our results identify the Notch signaling pathway as an important regulator of PD‐1 expression by activated CD8+ T cells.

Memory T-lymphocyte survival does not require T-cell receptor expression

The factors controlling memory T (Tm)-cell longevity are still poorly defined, and their identification is pivotal to the design of a vaccine conferring long-term protection against infection. Tm cells have the ability to survive in the absence of the T-cell receptor (TCR)–MHC interaction. This does not exclude a possible role for TCR-intrinsic ligand-independent constitutive signaling in Tm-cell homeostasis. Using a unique TCR tetracycline-inducible expression system, we show that the ablation of TCR expression, which abrogates any possible signaling via the TCR, did not influence the survival and self-renewal of antigen-specific CD8+ Tm cells even when they have to compete with endogenous T cells for survival factors. Moreover, CD8+ Tm-cell functionality was not altered even on prolonged maintenance in the absence of TCR-MHC interactions. Furthermore, our results show that a subset of CD4+ Tm cells can survive in the absence of TCR expression in nonlymphopenic hosts.

Toxic MHC Class II β Chains

MHC class II molecules (Aα:Aβ and Eα:Eβ in mice) present peptides to CD4+ T lymphocytes, promoting their selection in the thymus and survival and expansion in the periphery. Whether they play an equivalent role in the life history of B lymphocytes has been a long-standing question. TCR engagement by class II molecules on B cells triggers T cells to deliver cognate help; conversely, engagement of class II molecules on B cells transmits signals internally. Thus, it has long been hypothesized that class II molecules might be necessary for the differentiation or long-term survival of B cells.In a recent report, Rolink and coworkers (Rolink et al. 1999xRolink, A.G, Brocker, T, Bluethmann, H, Kosco-Vilbois, M.H, Anderson, J, and Melchers, F. Immunity. 1999; 10: 619–628Abstract | Full Text | Full Text PDF | PubMed | Scopus (44)See all ReferencesRolink et al. 1999) performed a thorough analysis of B cell populations in a line of class II–deficient mice generated by inactivation of the Aα gene in a strain harboring a natural mutation of the Eα gene. They convincingly demonstrated a 3- to 4-fold reduction in the number of mature B cells, and kinetic experiments indicated that this diminution was due to a decreased lifespan. On the other hand, immature populations seemed normal. This defect was intrinsic to the B cells and could be complemented by an Eα transgene, provided that it was expressed in B cells. These data were taken as evidence that class II molecules are important for the long-term survival of mature B cells.This interpretation immediately raised the question of why B cell pools are normal in other lines of class II–deficient mice, in particular those generated by mutation of the Aβ gene (2xCosgrove, D, Gray, D, Dierich, A, Kaufman, J, Lemeur, M, Benoist, C, and Mathis, D. Cell. 1991; 66: 1051–1066Abstract | Full Text PDF | PubMed | Scopus (641)See all References, 7xMarkowitz, J.S, Rogers, P.R, Grusby, M.J, Parker, D.C, and Glimcher, L.H. J. Immunol. 1993; 150: 1223–1233PubMedSee all References, 9xRolink, A.G, Brocker, T, Bluethmann, H, Kosco-Vilbois, M.H, Anderson, J, and Melchers, F. Immunity. 1999; 10: 619–628Abstract | Full Text | Full Text PDF | PubMed | Scopus (44)See all References). One reason put forth was that the Aβ-null mice might not really be class II–deficient, because the residual Aα chain might participate in heterodimers with the Eβ chain—although such mixed-isotype molecules have never been observed in these mice, even in highly sensitive analyses (confirmed by Rolink et al.). Alternatively, Aα chains might associate with a hitherto unrecognized molecule and these complexes transmit B cell survival signals.We have recently generated a third type of class II–deficient mouse line whose phenotype casts doubt on the proposed interpretation of the deficiency in mature B cells. The new line was produced by cre-mediated deletion of the entire MHC class II locus (Madsen et al. 1999xMadsen, L, Labrecque, N, Engberg, J, Dierich, A, Svejgaard, A, Benoist, C, Mathis, D, and Fugger, L. Proc. Natl. Acad. Sci. USA. 1999; 96: 10338–10343Crossref | PubMed | Scopus (173)See all ReferencesMadsen et al. 1999). Thus, none of the classical class II chains—Aα, Aβ, Eβ, or Eα—are synthesized in these mice. Yet, as demonstrated in Table 1Table 1, they have quite normal B cell numbers, and the partitioning between immature and mature compartments was unaffected (data not shown). Thus, the interpretation proposed by Rolink et al. is no longer tenable.Table 1B Cells in Mice Lacking All Class II GenesNumber of B Cells (B220+) per SpleenAgeWild TypeKnockout5 weeks761065 weeks68878 weeks70429 weeks5668We would like to propose an alternative explanation, taking into account other engineered mice reported to have similar “low B cell” phenotypes. The first such case was a set of high-copy-number transgenic lines overexpressing Aβ molecules (4xGilfillan, S, Aiso, S, Michie, S.A, and McDevitt, H.O. Proc. Natl. Acad. Sci. USA. 1990; 87: 7319–7323Crossref | PubMedSee all References, 11xSinger, S.M, Umetsu, D.T, and McDevitt, H.O. Proc. Natl. Acad. Sci. USA. 1996; 93: 2947–2952Crossref | PubMed | Scopus (14)See all References). In these mice, mature B cell numbers were drastically reduced, correlating with transgene copy number. This abnormality was observed in Aβ- but not Aα-overexpressing transgenic mice (Gilfillan et al. 1990xGilfillan, S, Aiso, S, Michie, S.A, and McDevitt, H.O. Proc. Natl. Acad. Sci. USA. 1990; 87: 7319–7323Crossref | PubMedSee all ReferencesGilfillan et al. 1990; M. LeMeur et al., unpublished data). Diminished numbers of mature B cells were also found in invariant chain–deficient animals (Shachar et al., 1996xShachar, I and Flavell, R.A. Science. 1996; 274: 106–108Crossref | PubMed | Scopus (90)See all ReferencesShachar et al., 1996; Kenty et al. 1998xKenty, G, Martin, W.D, van Kaer, L, and Bikoff, E.K. J. Immunol. 1998; 160: 606–614PubMedSee all ReferencesKenty et al. 1998).As illustrated in Table 2Table 2, the strains that present these B cell anomalies are those in which Aβ chains are unpaired or poorly paired, either because their normal pairing partner is lost (Aα knockout), because they are overexpressed (Aβ transgenics), or because the invariant chain does not perform its normal chaperoning function, facilitating proper formation of class II complexes (Germain and Margulies 1993xGermain, R.N and Margulies, D.H. Annu. Rev. Immunol. 1993; 11: 403–450Crossref | PubMedSee all ReferencesGermain and Margulies 1993). In contrast, engineered strains in which there are unpaired Aα chains (Aβ knockout, Aα-overexpressing transgenics) or in which there is a balanced reduction of all class II chains (whole class II knockout, CIITA knockout) do not exhibit this phenotype.Table 2Phenotypes in Class II Engineered MiceNormal B CellsReduced B CellsClass II–deficient (Aβ knockout) Class II–deficient (complete deletion)Class II–deficient (Aα knockout)CIITA-deficient (all class II reduced)Invariant chain deficientOverexpressed Aα in transgenicsOverexpressed Aβ in transgenicsThe implication is thus that unpaired or mispaired Aβ chains are toxic to B cells (an alternative explanation also raised by Rolink et al.). This may be because they associate with and sequester limiting components of the endoplasmic reticulum or transport vesicles (BiP, other chaperonins); more prosaically, Aβ chains may aggregate and “gum up” transport pathways for protein secretion or surface expression (Bonnerot et al. 1994xBonnerot, C, Marks, M.S, Cosson, P, Robertson, E.J, Bikoff, E.K, Germain, R.N, and Bonifacino, J.S. EMBO J. 1994; 13: 934–944PubMedSee all ReferencesBonnerot et al. 1994). There are precedents for such effects, for example, transgenic systems in which MHC molecules are overexpressed in pancreatic islet β cells (reviewed in Parham 1988xParham, P. Nature. 1988; 333: 500–503Crossref | PubMedSee all ReferencesParham 1988).In summary, the compiled analysis indicates that MHC class II and accessory molecules are not necessary per se for B cell differentiation or survival, but rather that mispaired β chains are toxic to mature B cells.

Development and Function of Innate Polyclonal TCRαβ+ CD8+ Thymocytes

Innate CD8 T cells are found in mutant mouse models, but whether they are produced in a normal thymus remains controversial. Using the RAG2p-GFP mouse model, we found that ∼10% of TCRαβ+ CD4−CD8+ thymocytes were innate polyclonal T cells (GFP+CD44hi). Relative to conventional T cells, innate CD8 thymocytes displayed increased cell surface amounts of B7-H1, CD2, CD5, CD38, IL-2Rβ, and IL-4Rα and downmodulation of TCRβ. Moreover, they overexpressed several transcripts, including T-bet, Id3, Klf2, and, most of all, Eomes. Innate CD8 thymocytes were positively selected, mainly by nonhematopoietic MHCIa+ cells. They rapidly produced high levels of IFN-γ upon stimulation and readily proliferated in response to IL-2 and IL-4. Furthermore, low numbers of innate CD8 thymocytes were sufficient to help conventional CD8 T cells expand and secrete cytokine following Ag recognition. This helper effect depended on CD44-mediated interactions between innate and conventional CD8 T cells. We concluded that innate TCRαβ+ CD8 T cells represent a sizeable proportion of normal thymocytes whose development and function differ in many ways from those of conventional CD8 T cells.

IL-6 production by dendritic cells is dispensable for CD8+ memory T-cell generation.

Following activation, naïve CD8+ T cells will differentiate into effectors that differ in their ability to survive: some will persist as memory cells while the majority will die by apoptosis. Signals given by antigen-presenting cells (APCs) at the time of priming modulate this differential outcome. We have recently shown that, in opposition to dendritic cell (DC), CD40-activated B-(CD40-B) cell vaccination fails to efficiently produce CD8+ memory T cells. Understanding why CD40-B-cell vaccination does not lead to the generation of functional long-lived memory cells is essential to define the signals that should be provided to naïve T cells by APCs. Here we show that CD40-B cells produce very low amount of IL-6 when compared to DCs. However, supplementation with IL-6 during CD40-B-cell vaccination did not improve memory generation. Furthermore, IL-6-deficient DCs maintained the capacity to promote the formation of functional CD8+ effectors and memory cells. Our results suggest that in APC vaccination models, IL-6 provided by the APCs is dispensable for proper CD8+ T-cell memory generation.

Human invariant chain isoform p35 restores thymic selection and antigen presentation in CD74-deficient mice.

The invariant chain (Ii) has pleiotropic functions and is a key factor in antigen presentation. Ii associates with major histocompatibility complex class II molecules in the endoplasmic reticulum (ER) and targets the complex in the endocytic pathway to allow antigenic peptide loading. The human Iip35 isoform includes a cytoplasmic extension containing a di‐arginine motif causing ER retention. This minor isoform does not exist in mice and its function in humans has not been thoroughly investigated. We have recently generated transgenic mice expressing Iip35 and these were crossed with Ii‐deficient mice to generate animals (Tgp35/mIiKO) expressing exclusively the human isoform. In these mice, we show that Iip35 is expressed in antigen presenting cells and is inducible by interferon gamma (IFN‐γ). Despite the low constitutive expression of the protein and some minor differences in the Vβ repertoire of Tgp35/mIiKO mice, Iip35 restored thymic selection of CD4+ T cells and of invariant natural killer T cells. In vitro functional assays using purified primary macrophages treated with IFN‐γ showed that Iip35 allows presentation of an Ii‐dependent ovalbumin T‐cell epitope. Altogether, our results suggest that Iip35 is functional and does not require co‐expression of other isoforms for antigen presentation.

Lowering TCR expression on naive CD8+ T cells does not affect memory T-cell differentiation.

The generation of long‐lived memory T (Tm) cells is critical for the success of vaccination, but the factors controlling their differentiation are still poorly defined. We examined the hypothesis that the level of T‐cell receptor (TCR) engagement contributed to memory CD8+ T‐cell generation. By manipulating TCR expression levels on murine, naive ovalbumin (OVA)‐specific CD8+ T cells, we showed that the expansion of antigen (Ag)‐specific CD8+ T cells is minimally affected by the level of TCR expression. Indeed, naive CD8+ T cells expressing as little as a 1000 TCRs (30‐fold less) show only a 2.5‐fold reduction in the number of effectors generated. Furthermore, the TCR expression levels influenced neither the acquisition of effector functions nor the generation of functional Tm cells. Our data indicate that during an in vivo immune response, a threshold in the number of TCRs engaged by naive CD8+ T cells is required for full T‐cell expansion but not for their differentiation into effector and Tm cells.