Collimation border with U-Net segmentation on chest radiographs compared to radiologists.

INTRODUCTION: Chest Radiography (CXR) is a common radiographic procedure. Radiation exposure to patients should be kept as low as reasonably achievable (ALARA), and monitored continuously as part of quality assurance (QA) programs. One of the most effective dose reduction tools is proper collimation practice. The purpose of this study is to determine whether a U-Net convolutional neural networks (U-CNN) can be trained to automatically segment the lungs and calculate an optimized collimation border on a limited CXR dataset. METHODS: 662 CXRs with manual lung segmentations were obtained from an open-source dataset. These were used to train and validate three different U-CNNs for automatic lung segmentation and optimal collimation. The U-CNN dimensions were 128 × 128, 256 × 256, and 512 × 512 pixels and validated with five-fold cross validation. The U-CNN with the highest area under the curve (AUC) was tested externally, using a dataset of 50 CXRs. Dice scores (DS) were used to compare U-CNN segmentations with manual segmentations by three radiographers and two junior radiologists. RESULTS: DS for the three U-CNN dimensions with segmentation of the lungs ranged from 0.93 to 0.96, respectively. DS of the collimation border for each U-CNN was 0.95 compared to the ground truth labels. DS for lung segmentation and collimation border between the junior radiologists was 0.97 and 0.97. One radiographer differed significantly from the U-CNN (p = 0.016). CONCLUSION: We demonstrated that a U-CNN could reliably segment the lungs and suggest a collimation border with great accuracy compared to junior radiologists. This algorithm has the potential to automate collimation auditing of CXRs. IMPLICATIONS FOR PRACTICE: Creating an automatic segmentation model of the lungs can produce a collimation border, which can be used in CXR QA programs.

Cladribine modifies functional properties of microglia.

Cladribine (CdA), an oral prodrug approved for the treatment of relapsing multiple sclerosis, selectively depletes lymphocytes. CdA passes the blood-brain barrier, suggesting a potential effect on central nervous system (CNS) resident cells. We examined if CdA modifies the phenotype and function of naive and activated primary mouse microglia, when applied in the concentrations 0·1-1 µM that putatively overlap human cerebrospinal fluid (CSF) concentrations. Primary microglia cultures without stimulation or in the presence of proinflammatory lipopolysaccharide (LPS) or anti-inflammatory interleukin (IL)-4 were treated with different concentrations of CdA for 24 h. Viability was assessed by MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay. Phagocytotic ability and morphology were examined by flow cytometry and random migration using IncuCyte Zoom and TrackMate. Change in gene expression was examined by quantitative polymerase chain reaction (qPCR) and protein secretion by Meso Scale Discovery. We found that LPS and IL-4 up-regulated deoxycytidine kinase (DCK) expression. Only activated microglia were affected by CdA, and this was unrelated to viability. CdA 0·1-1 µM significantly reduced granularity, phagocytotic ability and random migration of activated microglia. CdA 10 µM increased the IL-4-induced gene expression of arginase 1 (Arg1) and LPS-induced expression of IL-1ß, tumor necrosis factor (TNF), inducible nitric oxide synthase (iNOS) and Arg1, but protein secretion remained unaffected. CdA 10 µM potentiated the increased expression of anti-inflammatory TNF receptor 2 (TNF-R2) but not TNF-R1 induced by LPS. This suggests that microglia acquire a less activated phenotype when treated with 0·1-1 µM CdA that putatively overlaps human CSF concentrations. This may be related to the up-regulated gene expression of DCK upon activation, and suggests a potential alternative mechanism of CdA with direct effect on CNS resident cells.

Associations between levels of insulin-like growth factor 1 and sinusoidal obstruction syndrome after allogeneic haematopoietic stem cell transplantation.

Allogeneic myeloablative haematopoietic stem cell transplantation (HSCT) is challenged by severe adverse events, as cytotoxic effects of the conditioning may result in systemic inflammation, leaky epithelial barriers and organ toxicities, contributing to treatment-related morbidity and mortality. We hypothesised that insulin-like growth factor-1 (IGF-1), a mediator of growth and proliferation of various tissues, may attenuate chemotherapy-induced tissue damage after HSCT. We prospectively measured plasma levels of IGF-1 and its binding protein 3 (IGFBP-3) in 41 patients undergoing myeloablative HSCT. IGF-1 and IGFBP-3 levels were inversely correlated with C-reactive protein and interleukin-6 levels post HSCT. In multivariate analyses, low levels of IGF-1 and IGFBP-3 before conditioning were associated with increased risk of developing sinusoidal obstruction syndrome (SOS; OR=5.00 per 1 SDS decrease in IGF-1 (95% CI: 1.45-16.67), P=0.011 and OR=5.00 (1.37-20.00), P=0.015, respectively). Furthermore, low pre-transplant levels of IGF-1 and IGFBP-3 were associated with increased fluid retention during the first 21 days post transplant (OR=7.69 (95% CI: 1.59-33.33), P=0.012, and OR=2.94 (1.03-8.33), P=0.045). These data suggest that high levels of IGF-1 and IGFBP-3 may have a protective effect against fluid retention and SOS, possibly by attenuating systemic inflammation, and may prove useful as predictive biomarkers of SOS.

Analysis of miR-146a and miR-142-3p as Potential Markers of Freshly Isolated or In Vitro-Expanded Human Treg cells.

Regulatory CD4+ T cells (Tregs) are pivotal for prevention of autoimmunity. The use of Tregs is therefore of increasing interest in in vitro drug screening assays as well as for a cytotherapy per se against autoimmune disorders. For both purposes, in vitro expansion of peripheral blood Tregs is necessary and there is an increasing need to identify novel markers that can discriminate natural thymic-derived Tregs (tTregs) from other T cell subsets, and ideally, such markers should be stably expressed during in vitro expansion procedures. We screened for novel miRNAs differentially expressed in tTregs and identified miR-146a and 142-3p as possible candidates. We analysed freshly isolated naïve and activated tTregs and non-Treg subsets after or prior to in vitro expansion. We observed a tTreg-specific profile of these miRNAs together with FOXP3 and Helios in freshly isolated tTregs, but observed a decline in the same markers in activated tTregs as opposed to naïve tTregs. In vitro-expanded Tregs could be identified based on FOXP3 expression, but with loss of a discriminate profile for miRNA candidates and a decline in FOXP3 when activated tTregs were expanded. Our data demonstrate miR-146a and 142-3p as potential miRNA markers for discrimination between non-Treg cells and tTregs, but these miRNAs are not stable markers for in vitro-expanded Treg cells. In addition, the loss of FOXP3 in expansion of activated tTregs has implication for in vitro use of this cell subset in immunopharmacological assays and cytotherapy as FOXP3 is pivotal for suppressive function.

TL1A regulates TCRγδ+ intraepithelial lymphocytes and gut microbial composition.

TL1A is a proinflammatory cytokine, which is prevalent in the gut. High TL1A concentrations are present in patients with inflammatory bowel disease (IBD) and in IBD mouse models. However, the role of TL1A during steady-state conditions is relatively unknown. Here, we used TL1A knockout (KO) mice to analyse the impact of TL1A on the intestinal immune system and gut microbiota. The TL1A KO mice showed reduced amounts of small intestinal intraepithelial TCRγδ(+) and CD8(+) T cells, and reduced expression of the activating receptor NKG2D. Moreover, the TL1A KO mice had significantly reduced body weight and visceral adipose tissue deposits, as well as lower levels of leptin and CXCL1, compared with wild-type mice. Analysis of the gut microbial composition of TL1A KO mice revealed a reduction of caecal Clostridial cluster IV, a change in the Firmicutes/Bacteroidetes ratio in caecum and less Lactobacillus spp. in the mucosal ileum. Our results show that TL1A deficiency impacts on the gut microbial composition and the mucosal immune system, especially the intraepithelial TCRγδ(+) T-cell subset, and that TL1A is involved in the establishment of adipose tissue. This research contributes to a broader understanding of TL1A inhibition, which is increasingly considered for treatment of IBD.

Induction of Bcl-xL-specific cytotoxic T lymphocytes in mice.

The induction of active immunity against tumour-associated antigens to prevent relapse of cancer is a promising approach but has so far shown only low efficacy. This low efficacy may in part be due to clonal escape of tumour cell variants by the downregulation of antigen expression or inflammation-induced dedifferentiation. Identification of novel tumour-associated antigens that at the same time are essential for continued tumour cell survival is thus critical for the development of active cancer vaccinations. At the same time, identification of novel endogenous murine tumour antigens will help improve preclinical development of cancer immunotherapy. The anti-apoptotic protein Bcl-xL has been suggested to be such an essential tumour antigen, but the lack of well-defined murine epitopes have delayed preclinical studies of Bcl-xL-targeting cancer vaccines. Here, we report the identification of two novel murine tumour-associated epitopes TAYQSFEQV and AFFSFGGAL derived from mouse Bcl-xL. Dendritic cell (DC)-based vaccination induced CD8(+) T cells capable of producing IFN-γ upon restimulation with these epitopes. Thus, our data may benefit the design of future immunotherapy strategies by providing a preclinical model for cancer vaccination with an endogenous tumour antigen that can be combined with other cancer treatments.

Early detection of cancer in the general population: a blinded case-control study of p53 autoantibodies in colorectal cancer.

BACKGROUND: Recent reports from cancer screening trials in high-risk populations suggest that autoantibodies can be detected before clinical diagnosis. However, there is minimal data on the role of autoantibody signatures in cancer screening in the general population. METHODS: Informative p53 peptides were identified in sera from patients with colorectal cancer using an autoantibody microarray with 15-mer overlapping peptides covering the complete p53 sequence. The selected peptides were evaluated in a blinded case-control study using stored serum from the multimodal arm of the United Kingdom Collaborative Trial of Ovarian Cancer Screening where women gave annual blood samples. Cases were postmenopausal women who developed colorectal cancer following recruitment, with 2 or more serum samples preceding diagnosis. Controls were age-matched women with no history of cancer. RESULTS: The 50 640 women randomised to the multimodal group were followed up for a median of 6.8 (inter-quartile range 5.9-8.4) years. Colorectal cancer notification was received in 101 women with serial samples of whom 97 (297 samples) had given consent for secondary studies. They were matched 1 : 1 with 97 controls (296 serial samples). The four most informative peptides identified 25.8% of colorectal cancer patients with a specificity of 95%. The median lead time was 1.4 (range 0.12-3.8) years before clinical diagnosis. CONCLUSION: Our findings suggest that in the general population, autoantibody signatures are detectable during preclinical disease and may be of value in cancer screening. In colorectal cancer screening in particular, where the current need is to improve compliance, it suggests that p53 autoantibodies may contribute towards risk stratification.

Experimental methods and modeling techniques for description of cell population heterogeneity.

With the continuous development, in the last decades, of analytical techniques providing complex information at single cell level, the study of cell heterogeneity has been the focus of several research projects within analytical biotechnology. Nonetheless, the complex interplay between environmental changes and cellular responses is yet not fully understood, and the integration of this new knowledge into the strategies for design, operation and control of bioprocesses is far from being an established reality. Indeed, the impact of cell heterogeneity on productivity of large scale cultivations is acknowledged but seldom accounted for. In order to include population heterogeneity mechanisms in the development of novel bioprocess control strategies, a reliable mathematical description of such phenomena has to be developed. With this review, we search to summarize the potential of currently available methods for monitoring cell population heterogeneity as well as model frameworks suitable for describing dynamic heterogeneous cell populations. We will furthermore underline the highly important coordination between experimental and modeling efforts necessary to attain a reliable quantitative description of cell heterogeneity, which is a necessity if such models are to contribute to the development of improved control of bioprocesses.

A comparative study of transfection methods for RNA interference in bone marrow-derived murine dendritic cells.

Selective gene silencing using RNA interference (RNAi) has been shown to be an efficient method for manipulation of cellular functions. In this study, we compare three previously established methods for transfection of murine bone marrow-derived DC (BM-DC). We tested the efficacy of electroporation with the Mouse Nucleofector kit((R)) from Amaxa Biosystems and lipid-based transfection methods using transfection reagents from Santa Cruz Biotechnology or Genlantis. To analyse the transfection efficacy we used FITC-conjugated siRNA as a positive control together with CD80 and CD86 specific siRNA. We show that electroporation using the Mouse Nucleofector kit((R)) from Amaxa Biosystems was not an efficient method to transfect BM-DC with siRNA in our hands. Transfection with Santa Cruz Biotechnology reagents resulted in up to 59% FITC-siRNA positive cells, but did not result in effective silencing of CD80 surface expression. In contrast, the most effective method was the lipid-based method using the siRNA transfection reagent GeneSilencer((R)) from Genlantis. This protocol resulted in up to 92% FITC-siRNA positive cells after 4 h which declined to 62% and 59% 24 and 48 h post-transfection, respectively. The transfected BM-DC remained CD11c positive, expressed high MHC class II and intermediate CD40 and were functional as APC. In conclusion, this protocol was effective for manipulation of murine BM-DC function through the use of specific siRNA and such methods can be important for the future study of DC-T cell interactions.

Immune responses in patients with metastatic renal cell carcinoma treated with dendritic cells pulsed with tumor lysate.

Patients with metastatic renal cell carcinoma (mRCC) have a limited life expectancy but still a subset of these patients develop immune and clinical responses after immunotherapy including dendritic cell (DC) vaccination. In a recently published phase I/II trials, fourteen HLA-A2 negative patients with progressive mRCC were vaccinated with autologous DC pulsed with allogeneic tumour lysate. Low-dose IL-2 administered subcutaneously was given concomitantly. In this study, we analysed lysate specific proliferation of PBMCs from these patients together with the TH1/TH2 balance of the responding T cells. Also, serum concentrations of IL-10, IL-12, IL-15, IL-17 and IL-18 from these patients and additional thirteen HLA-A2 positive mRCC patients treated with autologous DC pulsed with survivin and telomerase peptides were analysed during vaccination to identify systemic immune responses and potential response biomarkers. In HLA-A2 negative mRCC patients a spontaneous predominance of TH1 secreting tumour lysate specific T cells was observed prior to vaccination in patients attaining stable disease (SD) during treatment whereas patients with continued progressive disease (PD) had a mixed TH1/TH2 response. The TH1/TH2 balance was unchanged during vaccination also when tumour lysate specific T cell responses increased. An increase in IL-12, IL-17 and IL-18 serum concentrations was observed during vaccination but no difference between patients with SD and PD was observed. IL-10 or IL-15 was not measurable in serum.

Genome-wide gene expression profiling of SCID mice with T-cell-mediated Colitis.

Inflammatory bowel disease (IBD) is a multifactorial disorder with an unknown aetiology. The aim of this study is to employ a murine model of IBD to identify pathways and genes, which may play a key role in the pathogenesis of IBD and could be important for discovery of new disease markers in human disease. Here, we have investigated severe combined immunodeficient (SCID) mice, which upon adoptive transfer with concanavalin A-activated CD4(+) T cells develop inflammation of the colon with predominance in rectum. Mice with increasing level of inflammation was studied. RNA from rectum of transplanted and non-transplanted SCID mice was investigated by a genome-wide gene expression analysis using the Affymetrix mouse expression array 430A (MOE430A) including 22,626 probe sets. A significant change in gene expression (P = 0.00001) is observed in 152 of the genes between the non-transplanted control mice and colitis mice, and among these genes there is an overrepresentation of genes involved in inflammatory processes. Some of the most significant genes showing higher expression encode S100A proteins and chemokines involved in trafficking of leucocytes in inflammatory areas. Classification by gene clustering based on the genes with the significantly altered gene expression corresponds to two different levels of inflammation as established by the histological scoring of the inflamed rectum. These data demonstrate that this SCID T-cell transfer model is a useful animal model for human IBD and can be used for suggesting candidate genes involved in the pathogenesis and for identifying new molecular markers of chronic inflammation in human IBD.

CD25 shedding by human natural occurring CD4+CD25+ regulatory T cells does not inhibit the action of IL-2.

Regulatory T (Treg) cells are important for the maintenance of peripheral tolerance and inhibition of pathogenic T-cell responses. Therefore, they are important for the limitation of chronic inflammation but can also be deleterious by e.g. limiting antitumour immune responses. Natural occurring Tregs are known to inhibit CD4+ T cell in a contact-dependent manner, but at the same time, various suppressive factors are secreted. We, here, demonstrate that human naturally occurring CD4+CD25+ Tregs are able to shed large amounts of soluble CD25 upon activation. Secretion of sCD25 could add to the inhibitory effect of Tregs as such secretion in other settings has been proposed to act as a sink for local IL-2. However, we here demonstrate that supernatant from human Tregs containing high concentration of sCD25 does not inhibit proliferation of CD4+CD25(-) T cells or inhibit the action of IL-2 in an in vitro bioassay.

Addition of TAT protein transduction domain and GrpE to human p53 provides soluble fusion proteins that can be transduced into dendritic cells and elicit p53-specific T-cell responses in HLA-A*0201 transgenic mice.

The protein p53 has been shown to be an efficient tumour antigen in both murine and human cancer vaccine studies and cancer vaccines targeting p53 based on major histocompatibility complex (MHC) class I binding p53-derived peptides that induce cytotoxic T lymphocytes (CTLs) without p53-specific CD4(+) T-cell help have been tested by several research groups including ours. To obtain such CD4(+) T-cell help and cover a broader repertoire of MHC haplotypes we have previously attempted to produce recombinant human p53 for vaccination purposes. However, attempts to refold a hexahis-tagged p53 protein in our laboratory were unsuccessful. Here, we show that fusion of an 11-amino-acid region of the human immunodeficiency virus TAT protein transduction domain (PTD) to human p53 increases the solubility of the otherwise insoluble p53 protein and this rTAT-p53 protein can be transduced into human monocyte-derived dendritic cells (DCs). The induction of a p53-specific HLA-A*0201 immune response was tested in HLA-A*0201/K(b) transgenic mice after immunization with rTAT-p53-transduced bone-marrow-derived DCs. In these mice, p53-specific CD4(+) and CD8(+) T-cell proliferation was observed and immunization resulted in the induction of HLA-A*0201-restricted CTLs specific for two human p53-derived HLA-A*0201-binding peptides, p53(65-73) and p53(149-157). Addition of GrpE to generate rTAT-GrpE-p53 led to a further increase in protein solubility and to a small increase in DC maturation but did not increase the observed p53-specific T-cell responses. The use of rTAT-p53 in ongoing clinical protocols should be applicable and offers advantages to current strategies omitting the use of HLA-typed patients.

Tolerogenic dendritic cells pulsed with enterobacterial extract suppress development of colitis in the severe combined immunodeficiency transfer model.

Immunomodulatory dendritic cells (DCs) that induce antigen-specific T-cell tolerance upon in vivo adoptive transfer are promising candidates for immunotherapy of autoimmune diseases. The feasibility of such a strategy has recently proved its efficacy in animal models of allotransplantation and experimental allergic encephalitis, but the effect in inflammatory bowel disease has not yet been demonstrated. In severe combined immunodeficient (SCID) mice, adoptively transferred CD4(+) CD25(-) T cells repopulate the lymphoid tissues and lead to development of chronic colitis characterized by CD4(+) T-cell proliferation against enterobacterial extract in vitro. In this model, we adoptively transferred in-vitro-generated bone-marrow-derived DCs exposed to interleukin-10 (IL-10) and an enterobacterial extract. We show that these cells are CD11c positive with intermediate expression of CD40, CD80 and CD86 and have a diminished secretion of IL-6, IL-12 p40/70, tumour necrosis factor-alpha and keratinocyte-derived chemokine (KC) compared to DCs treated with enterobacterial extract alone. In vivo, these cells prevented weight loss in SCID mice adoptively transferred with CD4(+) CD25(-) T cells, resulted in a lower histopathology colitis score and tended to result in higher serum levels of IL-1alpha, IL-10, IL-12, IL-13, IL-17, KC and monokine induced by interferon-gamma (MIG). These data underscore the potential of using immunomodulatory DCs to control inflammatory bowel disease and demonstrate its potential use in future human therapeutic settings.

The potential for induction of autoimmune disease by a randomly-mutated self-antigen.

The pathology of most autoimmune diseases is well described. However, the exact event that triggers the onset of the inflammatory cascade leading to disease is less certain and most autoimmune diseases are complex idiopathic diseases with no single gene known to be causative. In many cases, a relation to an infectious disease is described, and it is thought that microbes can play a direct role in induction of autoimmunity, for instance by molecular mimicry or bystander activation of autoreactive T cells. In contrast, less attention has been given to the possibility that modified self-antigens can be immunogenic and lead to autoimmunity against wildtype self-antigens. In theory, modified self-antigens can arise by random errors and mutations during protein synthesis and would be recognized as foreign antigens by naïve B and T lymphocytes. Here, it is postulated that the initial auto-antigen is not a germline self-antigen, but rather a mutated self-antigen. This mutated self-antigen might interfere with peripheral tolerance if presented to the immune system during an infection. The infection lead to bystander activation of naïve T and B cells with specificity for mutated self-antigen and this can lead to epitopespreading in which T and B cells with specificity for wildtype self-antigens are activated as a result of general inflammation.

Characterization of monocyte-derived dendritic cells maturated with IFN-alpha.

Dendritic cells (DC) are promising candidates for cancer immunotherapy. These cells can be generated from peripheral blood monocytes cultured with granulocyte macrophage-colony stimulating factor (GM-CSF) and interleukin-4 (IL-4). In order to obtain full functional capacity, maturation is required, but the most potent reagents such as LPS or polyriboinosinic polyribocytidylic acid (Poly I:C) are not approved for clinical use. We tested the ability of type I interferon (IFN) to induce such maturation. We found that 24-h IFN-alpha co-culture of day 7 monocyte-derived DC generated with GM-CSF and IL-4 induces increased numbers of DC positive for CD54 and CD40 together with the co-stimulatory molecule CD80 but not the activation marker CD83. Also, IFN-alpha maturation leads to an increase in IP-10 and MCP-1 chemokine secretion, but only a minor increase in IL-12p40 secretion. In line with this, maturation with IFN-alpha has only a small effect on induction of autologous T-cell stimulatory capacity of the DC. However, an increase in DC allogeneic T-cell stimulatory capacity was observed. These data suggest that IFN-alpha has a potential as a maturation agent used in DC-based cancer vaccine trials, but not as a single reagent.

Treatment of transplanted CT26 tumour with dendritic cell vaccine in combination with blockade of vascular endothelial growth factor receptor 2 and CTLA-4.

We investigated the anti CT26 tumour effect of dendritic cell based vaccination with the MuLV gp70 envelope protein-derived peptides AH1 and p320-333. Vaccination lead to generation of AH1 specific cytotoxic lymphocytes (CTL) and some decrease in tumour growth of simultaneously inoculated CT26 cells. After combination with an antibody against VEGF receptor 2 (DC101), a significant increase in survival of the tumour cell recipients was observed. Also, monotherapy with an antibody against CTLA-4 (9H10), led to approximately 100% survival of tumour cell recipients. However, effective treatment of mice with already established tumours was only obtained after combination of vaccination, DC101 and 9H10 treatment in which setting 80% of the mice rejected their tumours.

Phenotypic and functional characterization of clinical grade dendritic cells generated from patients with advanced breast cancer for therapeutic vaccination.

Dendritic cells (DC) are promising candidates for cancer immunotherapy. However, it is not known whether in vitro-generated monocyte-derived DC from cancer patients are altered compared with DC from healthy donors. In a clinical phase I/II study, monocyte-derived DC were generated in vitro utilizing granulocyte macrophage colony-stimulating factor and rh-interleukin-4 (IL-4) and used for cancer immunotherapy. In this study, we tested the effect of various maturation cocktails and performed a comparative evaluation of the DC phenotype and functional characteristics. Polyriboinosinic polyribocytidylic acid (Poly I:C) + tumour necrosis factor-alpha (TNF-alpha) induced significant IL-12 p70 secretion, which was increased after addition of a decoy IL-10 receptor. The lymph node homing chemokine receptor CCR-7 expression was induced by TNF-alpha + IL-1beta + IL-6 + prostaglandin E2 but was not induced by Poly I:C + TNF-alpha. In general, DC from patients had an intermediate maturity phenotype with a significantly higher expression of CD40 and CD54 compared with healthy donors. In vitro analyses showed an unimpaired capacity of the patient-derived DC for antigen-specific (cytomegalovirus, tetanus and keyhole limpet haemocyanin) T-cell stimulation, whereas the allostimulatory capacity of patient-derived DC was significantly decreased. These data suggest that patient-derived DC are more differentiated but are less sensitive to maturation-inducing agents than DC obtained from healthy individuals.

Signal transduction by the major histocompatibility complex class I molecule.

Ligation of cell surface major histocompatibility class I (MHC-I) proteins by antibodies, or by their native counter receptor, the CD8 molecule, mediates transduction of signals into the cells. MHC-I-mediated signaling can lead to both increased and decreased activity of the MHC-I-expressing cell depending on the fine specificity of the anti-MHC-I antibodies, the context of CD8 ligation, the nature and cell cycle state of the MHC-I-expressing cell and the presence or absence of additional cellular or humoral stimulation. This paper reviews the biochemical, physiological and cellular events immediately after and at later intervals following MHC-I ligation. It is hypothesized that MHC-I expression, both ontogenically and in evolution, is driven by a cell-mediated selection pressure advantageous to the MHC-I-expressing cell. Accordingly, in addition to their role in T-cell selection and functioning, MHC-I molecules might be of importance for the maintenance of cellular homeostasis not only within the immune system, but also in the interplay between the immune system and other organ systems.

MHC class I cross-talk with CD2 and CD28 induces specific intracellular signalling and leads to growth retardation and apoptosis via a p56(lck)-dependent mechanism.

Ligation of the major histocompatibility complex class I molecules (MHC-I) on human T lymphoma cells (Jurkat) initiates p56(lck)-dependent intracellular signalling events (phosphotyrosine kinase activity; [Ca(2+)](i)) and leads to augmented growth inhibition and apoptosis. MHC-I ligation in concert with ligation of CD2 or CD28 augments, changes or modifies the pattern of activation. Ligation of MHC-I and CD2 alone resulted in growth inhibition, whereas CD28 ligation alone had no effect on cell proliferation. Ligation of MHC-I together with CD2 augmented growth inhibition and enhanced the level of apoptosis. In parallel experiments with the p56(lck)-negative Jurkat mutant cell, JCaM1.6, cross-linking neither influenced cell signalling nor cellular growth functions, indicating a cardinal role of the src kinases in signal transduction via MHC-I, CD2 and CD28 molecules. The results presented here provide evidence for the involvement of MHC-I molecules in the modulation of signal transduction via the CD2 and CD28 costimulatory molecules.