No abstract

Id2 expression delineates differential checkpoints in the genetic program of CD8α+ and CD103+ dendritic cell lineagesA novel reporter mouse for the transcription factor inhibitor of DNA binding-2 (Id2) provides insight into the differentiation of DC subsets in vivo and in vitro.

About 30% of cases of the autosomal recessive immunodeficiency disorder hemophagocytic lymphohistiocytosis are believed to be caused by inactivating mutations of the perforin gene. We expressed perforin in rat basophil leukemia cells to define the basis of perforin dysfunction associated with two mutations, R225W and G429E, inherited by a compound heterozygote patient. Whereas RBL cells expressing wild-type perforin (67 kD) efficiently killed Jurkat target cells to which they were conjugated, the substitution to tryptophan at position 225 resulted in expression of a truncated ( ‫ف‬ 45 kD) form of the protein, complete loss of cytotoxicity, and failure to traffic to rat basophil leukemia secretory granules. By contrast, G429E perforin was correctly processed, stored, and released, but the rat basophil leukemia cells possessed reduced cytotoxicity. The defective function of G429E perforin mapped downstream of exocytosis and was due to its reduced ability to bind lipid membranes in a calcium-dependent manner. This study elucidates the cellular basis for perforin dysfunctions in hemophagocytic lymphohistiocytosis and provides the means for studying structure-function relationships for lymphocyte perforin.

Because CD4 ؉ T cells play a key role in aiding cellular immune responses, we wanted to assess whether increasing numbers of gene-engineered antigenrestricted CD4 ؉ T cells could enhance an antitumor response mediated by similarly gene-engineered CD8 ؉ T cells. In this study, we have used retroviral transduction to generate erbB2-reactive mouse T-cell populations composed of various proportions of CD4 ؉ and CD8 ؉ cells and then determined the antitumor reactivity of these mixtures. Gene-modified CD4 ؉ and CD8 ؉ T cells were shown to specifically secrete Tc1 (T cytotoxic-1) or Tc2 cytokines, proliferate, and lyse erbB2 ؉ tumor targets following antigen ligation in vitro. In adoptive transfer experiments using severe combined immunodeficient (scid) mice, we demonstrated that injection of equivalent numbers of antigenspecific engineered CD8 ؉ and CD4 ؉ T cells led to significant improvement in survival of mice bearing established lung metastases compared with transfer of unfractionated (largely CD8 ؉ ) engineered T cells. Transferred CD4 ؉ T cells had to be antigen-specific (not just activated) and secrete interferon ␥ (IFN-␥) to potentiate the antitumor effect. Importantly, antitumor responses in these mice correlated with localization and persistence of geneengineered T cells at the tumor site. Strikingly, mice that survived primary tumor challenge could reject a subsequent rechallenge. Overall, this study has highlighted the therapeutic potential of using combined transfer of antigen-specific gene-modified CD8 ؉ and CD4 ؉ T cells to significantly enhance T-cell adoptive transfer strategies for cancer therapy. IntroductionMuch attention has been directed at the genetic modification of T cells and their therapeutic potential in the adoptive immunotherapy of cancer. T cells engineered to express chimeric surface receptors that incorporate an extracellular single-chain antibody domain (scFv) and a transmembrane and cytoplasmic signaling domain can specifically direct antitumor immune responses in a major histocompatibility complex (MHC)-independent manner at targets normally capable of evading immune recognition. These targets also often lack important costimulatory ligands for maximal T-cell activation. The therapeutic potential of engineered T cells extends from studies demonstrating specific antigen binding and target-cell lysis in vitro 1-3 and in a range of different mouse tumor models 4-8 to their successful transfer into patients with minimal side effects in phase 1 clinical trials. 9 A number of studies have demonstrated the importance of CD4 ϩ T-cell help in elimination of infectious disease and for antitumor immunity. 10,11 CD4 ϩ T cells have been demonstrated to be critical for maintenance of CD8 ϩ T-cell numbers, their recruitment to the tumor site, 12-14 and induction of a memory response. 15,16 In patients infected with HIV, the coinfusion of genetically modified CD4 ϩ and CD8 ϩ T cells has been demonstrated to overcome the lack of T-cell persistence observed with transfusion of engineered CD8 ϩ T cel...

Key Points• Transcription factors Batf3, Id2, and Nfil3 are not essential for induced CD8␣ ϩ DC generation.• Induced CD8␣ ؉ DCs can crosspresent cellular antigens. Antiviral immunity and cross-presentation is mediated constitutively through CD8␣؉ and CD103 ؉ DCs. Development of these DC subsets is thought to require the transcription factors Irf8, Id2, Nfil3, and Batf3, although how this network is regulated is poorly defined. We addressed the nature of the differentiation blocks observed in the absence of these factors and found that although all 4 factors are required for CD103 ؉ DC development, only Irf8 is essential for CD8␣ ؉ DCs. CD8␣ ؉ DCs emerged in the absence of Id2, Nfil3 and Batf3 in short-term bone marrow reconstitution. These "induced" CD8␣ ؉ DCs exhibit several hallmarks of classic CD8␣ ؉ DCs including the expression of CD24, Tlr3, Xcr1, Clec9A, and the capacity to cross-present soluble, cell-associated antigens and viral antigens even in the absence of Batf3. Collectively, these results uncover a previously undescribed pathway by which CD8␣ ؉ DCs emerge independent of Id2, Nfil3, and Batf3, but dependent on Irf8. (Blood. 2013;121(9):1574-1583) Introduction CD103 ϩ and CD8␣ ϩ dendritic cells (DCs) are important for cross-presentation of antigens and for the induction of effector CD8 ϩ T-cell responses against pathogens. Differentiation of these 2 subsets is regulated by a common set of transcription factors including interferon regulatory factor 8 (Irf8), nuclear factor interleukin 3-regulated (Nfil3 or E4BP4) and Id2 (inhibitor of DNA binding 2). [1][2][3][4] Both populations also express the transcription factor Batf3 (basic leucine zipper transcriptional factor ATF-like 3, also known as Jun dimerization protein p21SNFT) 2,5 but display differential dependence on Batf3 for differentiation. 6,7 In initial studies, loss of Batf3 was thought to be critical for the development of CD8␣ ϩ DCs which provided an elegant explanation for the inability of Batf3 Ϫ/Ϫ mice (generated on a B6.129 background) to cross-present cell-associated or viral antigens. 5 Subsequently, these mice were discovered to also lack CD103 ϩ DCs. 2 This allowed confirmation of the hypothesis that CD103 ϩ DCs from the lung were important for capture and transport of antigen to the draining lymph node (LN) where an immune response is initiated after inflammation or infection. [8][9][10] Despite this, it has been observed that at steady-state in mice of a C57BL/6 genetic background, a population of DCs that express low levels of CD8␣ persist in spleen and LNs, and CD8␣ ϩ DC precursors can be identified in vitro in response to Flt3L stimulation. 6,7 Furthermore, CD103 expression could be induced in vitro on DCs after addition of GM-CSF even in the absence of Batf3. 6,7 Collectively, these findings raise major questions concerning the exact role of Batf3 Ϫ/Ϫ in DC differentiation and function.We now report that Batf3 intrinsically controls the development of CD103 ϩ Epcam ϩ DCs and the ability to cross-present cellassociated, but not ...

A new strategy to improve the therapeutic utility of redirected T cells for cancer involves the development of novel Ag-specific chimeric receptors capable of stimulating optimal and sustained T cell antitumor activity in vivo. Given that T cells require both primary and costimulatory signals for optimal activation and that many tumors do not express critical costimulatory ligands, modified single-chain Ab receptors have been engineered to codeliver CD28 costimulation. In this study, we have compared the antitumor potency of primary T lymphocytes expressing carcinoembryonic Ag (CEA)-reactive chimeric receptors that incorporate either TCR-ζ or CD28/TCR-ζ signaling. Although both receptor-transduced T cell effector populations demonstrated cytolysis of CEA+ tumors in vitro, T cells expressing the single-chain variable fragment of Ig (scFv)-CD28-ζ chimera had a far greater capacity to control the growth of CEA+ xenogeneic and syngeneic colon carcinomas in vivo. The observed enhanced antitumor activity of T cells expressing the scFv-CD28-ζ receptor was critically dependent on perforin and the production of IFN-γ. Overall, this study has illustrated the ability of a chimeric scFv receptor capable of harnessing the signaling machinery of both TCR-ζ and CD28 to augment T cell immunity against tumors that have lost expression of both MHC/peptide and costimulatory ligands in vivo.

Key Points• Lyl1 is required for Lmo2-induced T-cell leukemia in mice, whereas Scl is dispensable.• LYL1 is required for growth of ETP-ALL cell lines.Lmo2 is an oncogenic transcription factor that is frequently overexpressed in T-cell acute lymphoblastic leukemia (T-ALL), including early T-cell precursor ALL (ETP-ALL) cases with poor prognosis. Lmo2 must be recruited to DNA by binding to the hematopoietic basic helix-loop-helix factors Scl/Tal1 or Lyl1. However, it is unknown which of these factors can mediate the leukemic activity of Lmo2. To address this, we have generated Lmo2-transgenic mice lacking either Scl or Lyl1 in the thymus. We show that although Scl is dispensable for Lmo2-driven leukemia, Lyl1 is critical for all oncogenic functions of Lmo2, including upregulation of a stem cell-like gene signature, aberrant self-renewal of thymocytes, and subsequent generation of T-cell leukemia. Lyl1 expression is restricted to preleukemic and leukemic stem cell populations in this model, providing a molecular explanation for the stage-specific expression of the Lmo2-induced gene expression program. Moreover, LMO2 and LYL1 are coexpressed in ETP-ALL patient samples, and LYL1 is required for growth of ETP-ALL cell lines. Thus, the LMO2-LYL1 interaction is a promising therapeutic target for inhibiting self-renewing cancer stem cells in T-ALL, including poor-prognosis ETP-ALL cases. (Blood. 2013;122(12):2093-2103

We identify empirical scaling laws for the cross-entropy loss in four domains: generative image modeling, video modeling, multimodal image↔text models, and mathematical problem solving. In all cases autoregressive Transformers smoothly improve in performance as model size and compute budgets increase, following a power-law plus constant scaling law. The optimal model size also depends on the compute budget through a power-law, with exponents that are nearly universal across all data domains. The cross-entropy loss has an information theoretic interpretation as S(True) + D KL (True||Model), and the empirical scaling laws suggest a prediction for both the true data distribution's entropy and the KL divergence between the true and model distributions. With this interpretation, billion-parameter Transformers are nearly perfect models of the YFCC100M image distribution downsampled to an 8 × 8 resolution, and we can forecast the model size needed to achieve any given reducible loss (ie D KL ) in nats/image for other resolutions. We find a number of additional scaling laws in specific domains: (a) we identify a scaling relation for the mutual information between captions and images in multimodal models, and show how to answer the question "Is a picture worth a thousand words?"; (b) in the case of mathematical problem solving, we identify scaling laws for model performance when extrapolating beyond the training distribution; (c) we finetune generative image models for ImageNet classification and find smooth scaling of the classification loss and error rate, even as the generative loss levels off. Taken together, these results strengthen the case that scaling laws have important implications for neural network performance, including on downstream tasks.