CD45 regulates apoptosis in peripheral T lymphocytes

Programmed cell death (apoptosis) is a key mechanism for regulating lymphocyte numbers. Murine lymph node lymphocytes cultured in vitro without added stimuli show significant levels of apoptosis over 24 h, detectable by staining with Annexin V. CD4 and CD8 T lymphocytes from transgenic (Tg) mice expressing single CD45RABC or CD45RO isoforms show increased apoptosis and the extent of apoptosis is inversely correlated with the level of CD45 expression. CD45 Tg cells exhibit phosphatidyl serine translocation and DNA oligonucleosome formation, and can be partially rescued from apoptosis by culture in caspase inhibitors or common gamma-chain-binding cytokines. We conclude that CD45 is an important regulator of spontaneous apoptosis in T lymphocytes and this mechanism may contribute to the disease associations reported for individuals expressing CD45 variant alleles.     

Protein kinase C isoform expression in CD45RA+ and CD45RO+ T lymphocytes.

Differences in levels of specific enzymes utilized in intracellular signalling could be a factor in the distinct signalling properties observed in memory and naive T cells. We have studied the expression of both classical and non-classical protein kinase of C (PKC) isoenzymes in CD45RA and CD45RO cells using a combination of Western blot and flow cytometric analysis. These data indicate that CD45RA cells express higher levels of PKC alpha, PKC beta and PKC delta than CD45RO cells. In addition, CD45RA+ cells show greater proliferative activity when stimulated with phorbol myristate acetate (PMA) and calcium ionophore than their CD45RO+ counterparts. Variations in the levels of these isoenzymes could be implicated in functional differences, such as proliferation and cytokine production, in these cell subsets.     

Association of CD2 and T200 (CD45) in mouse T lymphocytes

A monoclonal antibody (mAb 12-15) reactive with the mouse CD2 was found to co-precipitate a high-molecular-weight glycoprotein from mouse thymocyte, splenic lymphocyte, Con A blast, and T cell tumor detergent lysates which was identified as the leukocyte common T 200 glycoprotein (CD45). The reactivity was specific for CD2 since antibodies to CD3 did not co-precipitate the T200 glycoprotein. mAb 12-15 did not react with immunoaffinity-purified T200 glycoprotein, ruling out the possibility that the antibody detected a cross-reactive epitope. Biochemical data indicated that the association of CD2 with T200 was not generated during lysis of the cell and that the molecular complex was non-covalently linked since it could be destroyed by high salt washing or boiling in SDS. Distribution analysis in Triton X114-H2O revealed that, in contrast to free T200 molecules, the complexed T200 was enriched in the detergents phase. To investigate the CD2-T200 association in more detail at the cell surface, modulation of CD2 and T200 was studied. Modulation could be induced on Con A blasts by monoclonal antibodies followed by cross-linking with a FITC-conjugated second antibody. Within 24 h the expression of CD2 or T200 was reduced to approximately 10-20% of the initial value on the majority of cells. However, two-color fluorescence showed that modulation of CD2 did not lead to co-modulation of CD3 or T200. A possible physiological role of CD2-T200 complexes is discussed.     

Isoform-specific associations of CD45 with accessory molecules in human T lymphocytes.

Association of CD45 with surface molecules was investigated in human T lymphocytes by co-capping. CD45 appeared to be associated with the CD3/T cell receptor complex and with CD4 or CD8 molecules in memory, but not in naive T cells, as previously reported in the mouse. Associations of CD45 isoforms with accessory molecules were then identified with seven anti-CD45R monoclonal antibodies (mAb). An isoform-specific association pattern was observed: CD2 co-capped with CD45 molecules recognized by UCHL1 mAb (CD45R0). LFA-1 with molecules bound by 2H4 mAb (CD45RA), and both CD4 and CD8 with molecules reacting with MCA.347 mAb (whose isoform specificity was not known). Further information on the CD45 isoform(s) associated to CD4 and CD8 was sought by assessing the isoform specificity of MCA.347. Cross-competition experiments showed that it reacts with an epitope clearly different from those recognized by 2H4 and UCHL1, and only partially overlapping the PD7/26 epitope (CD45RB). Moreover, the competition between MCA.347 and PD7/26 was maximal in naive T cells and minimal both in memory T cells and in a subset expressing CD11b, a marker of granular lymphocytes. Immunoprecipitation experiments showed that MCA.347 binds to CD45 molecules with a molecular mass of 220, 205 and 190 kDa, the 190-kDa molecules not being recognized by 2H4, PD7/26 or UCHL1. These data indicate that MCA.347 recognizes amino acid sequences different from those coded by the exon A or B of the gene, and not expressed by CD45R0, suggesting that it binds to sequences coded by the exon C. In conclusion, this work shows that in human T cells different CD45 isoforms are associated to different surface molecules: LFA-1 is associated to CD45RA, CD2 to CD45R0 and CD4 and CD8 presumably to CD45RC. This peculiar behavior of CD45 suggests that it may play a crucial role in lymphocyte activation, probably by modulating the signals delivered to the cell by different receptor systems.     

Co-expression of the CD45RA and CD45RO antigens on T lymphocytes in chronic arthritis

The site of T lymphocyte activation in chronic arthritis is unknown. Peripheral blood (PB) lymphocytes from chronic arthritis patients are in a ‘naïve’ or non-activated state, as defined by expression of the CD45RA antigen and lack of HLA class II expression. In contrast, most synovial fluid (SF) T lymphocytes express a ‘memory’ or activated phenotype, as defined by the CD45RO antigen and high HLA class II expression. Following stimulation, naive cells lose CD45RA and gain CD45RO expression to become memory cells with a transitional stage of dual CD45RA, CD45RO antigen expression. To localize where this change in phenotype occurs we used dual colour immunofluorescence labelling to compare the percentage of dual CD45RA, CD45ROpositive T lymphocytes in PB and SF from chronic arthritic patients and from normal PB, assuming this population would be increased at the primary site of T lymphocyte activation. Expression of the intermediate and late activation marker. HLA-DR, was also analysed using dual colour immunofluorescence labelling. The percentage of dual positive T lymphocytes was similar between arthritic PB, SF. and normal PB, as was the density of both CD45RA and CD45RO antigens. Thus, CD45 isoform expression did not indicate where T lymphocytes were activated. However, we identified a previously unreported population of CD45RA⁺ CD45RO⁺ HLA-DR^- T lymphocytes in arthritic and normal PB. In SF, this population was absent, but a substantial number of dual CD45RA, CD45RO-positive HLA-DR⁺ T lymphocytes were identified. This population would not be predicted by the current model of T lymphocyte activation. Division of T lymphocytes into functional groups on the basis of CD45 isoform expression is likely to be more complicated than previously thought. Based on our findings we propose an alternative model of T lymphocyte differentiation.     

Human CD4+ T cells expressing CD45RA acquire the lymphokine gene expression of CD45RO+ T-helper cells after activation in vitro.

CD4+ T cells were separated into subpopulations according to their expression of different isoforms of the CD45R molecule, i.e. CD45RA and CD45RO. The separated cells were activated with staphylococcal enterotoxin A (SEA) in the presence of formalin fixed Raji cells. Each set of cells was activated twice with a 6-day interval, and the lymphokine gene expression during the first 3 days after initiation of each stimulation was followed by use of polymerase chain reaction (PCR) technology. The lymphokine messenger RNA (mRNA) profiles were found to differ between the subsets, since after the first stimulation the CD45RA+ cells produced mRNA encoding interleukin-2 (IL-2) and IL-1 alpha, whereas the CD45RO+ cells transcribed genes for IL-1 alpha, IL-2, IL-4, IL-5 and interferon-gamma (IFN-gamma). After 6 days of SEA stimulation both populations were mainly CD45RO reactive, and when restimulated displayed the lymphokine mRNA profile restricted to this subset. These results indicate that the CD45RA subset is a precursor of the CD45RO and further strengthen the hypothesis that the former cell population represents naive whereas the latter subset represents memory T cells within the CD4 subset.     

CD4+ CD45RA+ T cells modulate allergen-induced interleukin 2 responsiveness in human lymphocytes.

Peripheral blood lymphocytes from nonallergic individuals acquired responsiveness to interleukin 2 (IL2) after stimulation with ovalbumin (OVA) or Dermatophagoides farinae (Df) antigens when they were pretreated with the CD45RA antibody, which has been shown to define the suppressor inducer subset of CD4+ cells and also to block its suppressor activity. The effect provided by the CD45RA antibody was lost if the lymphocytes had initially been activated with the OVA of Df antigens. The magnitude of the responses was comparable to the allergen-induced responses observed in OVA- or Df-sensitized lymphocytes from allergic patients. The pre-existing IL2 responsiveness in the patients was not increased by the CD45RA antibody pretreatment. However, the CD45RA antibody pretreatment gave rise to Df-induced IL2 responsiveness in the lymphocytes of the patients sensitized with OVA but not with Df; conversely, OVA-induced IL2 responsiveness was enhanced in Df- but not in OVA-sensitized lymphocytes. The CD45RA antibody apparently acts on CD4+ T cells, but not on CD8+ T cells, to induce the IL2 response. A further dissection of normal CD4+ T cells indicated that CD4+45RA- T cells preferentially respond to IL2 after stimulation with OVA or Df antigens. Since normal CD4+45RA+ T cells did not show antigen-induced IL2 responsiveness even after pretreatment with the CD45RA antibody, it is unlikely that the CD45RA antibody stimulates CD4+45RA+ T cells to become responsive to IL2 after antigenic challenge. Alternatively, CD4+45RA+ T cells may modulate the activity of CD4+45RA- T cells, which are potentially responsive to IL2 by antigenic stimulation and thus provide tolerance in nonallergic lymphocytes. Collectively, a defective suppressor activity of CD4+45RA+ T cells may exist in patients with hen-egg allergy and/or bronchial asthma, which may cause lymphocytes to be hyperreactive to OVA or Df antigens.     

CD50 and CD62L adhesion receptor expression on naive (CD45RA+) and memory (CD45RO+) T lymphocytes in the elderly.

A complex reshaping characterizes cellular immunity in the elderly. In particular, the hallmark of the "senescence" of the T cell compartment is a decrease in the proportion of CD45RA+ naive T lymphocytes concomitantly with an expansion of CD45RO+ memory T cells. However, in addition to age-dependent changes in their representation, phenotypical and functional anomalies also characterize naive and memory T cell populations in the elderly. Since cell adhesion molecules (CAMs) are multifunctional receptors which play important roles not only in cell-to-cell and cell-to-matrix interactions but also in signal transduction and cell activation, we analysed, by means of a three-colour flow cytometry method, the proportion, absolute number and density expression or mean fluorescence intensity (MFI) of CD50 (ICAM-3) and CD62L (L-selectin homing receptor) adhesion receptors on CD45RA+ and CD45RO+ peripheral blood CD3+ T cell subsets from 10 healthy elderly subjects and 10 young controls. Our aim was to investigate age-dependent changes in the expression pattern of these CAMs on naive and memory lymphocytes which might contribute to the remodelling of the immune system in the elderly. We considered the mean values +/- standard deviations of the percentage, absolute number and MFI of positive cells. The percentage of naive T cells expressing CD50 was not significantly modified in aged (94.8 +/- 5.0%) compared to young individuals (97.8 +/- 3.2%). On the contrary, the percentage of memory T cells exhibiting CD50 was lower in elderly than young donors (92.0 +/- 6.4 vs. 98.3 +/- 2.2%; p < 0.01). The percentage of naive T cells expressing CD62L was decreased in the elderly donors (53.3 +/- 18.8 vs. 80.8 +/- 11.0%; p < 0.001), whereas the proportion of CD62L+ memory T lymphocytes was substantially comparable between the two age groups (63.5 +/- 15.7 vs. 54.7 +/- 12.3%). The absolute number per mm(3) of CD50+ naive T cells from aged individuals was decreased (251.9 +/- 141.9 vs. 621.8 +/- 238.0/mm(3); p < 0.001), whereas memory peripheral blood T lymphocytes expressing CD50 were substantially unchanged (863.8 +/- 260.9 vs. 802.7 +/- 139.6/mm(3)). On the contrary, the absolute numbers per mm(3) of naive and memory peripheral blood T lymphocytes exhibiting CD62L were respectively decreased (190.8 +/- 133.4/mm(3)) and increased (515.1 +/- 146.8/mm(3)) in elderly donors compared to young controls (601.3 +/- 129.1 and 351.8 +/- 195.0/mm(3); p < 0.001 and p < 0.05, respectively). Finally, CD50 MFI values of naive as well as memory T cell subpopulations from aged subjects were increased compared to young donors (14.0 +/- 2.0 vs. 9.8 +/- 1.2 and 14.0 +/- 2.0 vs. 11.6 +/- 1.3; p < 0.001 and p < 0.01, respectively). CD62L was also overexpressed in both naive (8.4 +/- 1.6 vs. 6.7 +/- 1.4; p < 0.05) and memory (10.3 +/- 2.5 vs. 5.4 +/- 1.1; p < 0.001) T subsets in the elderly. CD50 and CD62L upregulation could be interpreted as a compensatory mechanism for a decreased responsiveness and a greater requirement for activation signals rather than an age-related anomaly.     

Expression of CD45RB functionally distinguishes intestinal T lymphocytes in inflammatory bowel disease

The importance of CD45RB expression on T cells was already shown in mice where CD45RB(high) expression determines pathogenic potential. In this study, we analyzed the expression of CD45RA, CD45RB, and CD45RO on CD4(+) T lymphocytes in the intestinal mucosa and in the circulation of patients with inflammatory bowel disease (IBD). In addition, we studied the cytokine profile of these cells. In the circulation, virtually all CD4(+)CD45RB(high) T cells expressed the naive marker CD45RA, and circulating CD4(+)CD45RB(low) cells expressed the memory marker CD45RO in IBD patients and a control patient population. In contrast, the intestinal CD4(+) CD45RB(high) T cells are in normal controls for 90% CD45RO(+). However, in IBD, 27.7% [Crohn's disease (CD)] and 49% [ulcerative colitis (UC)] of the intestinal CD4(+) CD45RB(high) T cells are CD45RA(+). This special CD4CD45RA(+) T cell in IBD can be found in the lamina propria as well as in lymphoid follicles (confocal laser-scanning microscopy). The CD4(+)CD45RB(high) T lymphocytes produce significantly less interleukin (IL)-10 and IL-4 and produce more tumor necrosis factor alpha than CD45RB(low) T lymphocytes in control patients. CD4(+)CD45RB(low) T cells from IBD patients produced less IL-10 than CD4(+)CD45RB(low) T lymphocytes of controls, and interferon-gamma production by both T lymphocyte subsets was decreased in IBD. These data indicate that CD and UC are characterized by an influx of CD4(+)CD45RB(high) T lymphocytes. These CD4(+)CD45RB(high) T lymphocytes seem to be important in the pathogenesis of IBD, as they produce more proinflammatory cytokines and less anti-inflammatory cytokines compared with CD4(+)CD45RB(low) T lymphocytes.     

Human CD8+ intraepithelial T lymphocytes are mainly CD45RA-RB+ and show increased co-expression of CD45R0 in celiac disease

Expression of various CD45 isoforms (RA, RB and R0) on CD3+, CD4+ and CD8+ intraepithelial and lamina propria T cells was examined in situ by a three-color immunofluorescence technique in jejunal biopsy specimens from 32 patients with celiac disease and 18 controls. The median percentage of CD3+ intraepithelial lymphocytes (IEL) that expressed CD45R0 increased from 52% in controls to 69% in untreated celiac disease (p less than 0.01). Furthermore, the percentages of CD4+ and CD8+ IEL strongly positive for CD45R0 rose respectively from 94% and 24% in controls to 100% and 55% in untreated celiac disease. Conversely, CD45R0 was strongly expressed on most CD3+ lamina propria lymphocytes (LPL) both in control (81%) and diseased (77%-81%) mucosa. A variable fraction of the intraepithelial and lamina propria CD3+ T cells expressed mainly CD45RB (controls, 46% and 20%, respectively; celiac disease, 29% and 15%). Only 2% IEL and 4% LPL were positive for CD45RA. Expression of different CD45R isoforms thus identified three distinct CD8+ T cell subsets in human intestinal mucosa. In addition, our results suggested that antigen-primed CD8+CD45R0+ memory cells accumulate in the jejunal epithelium of patients with untreated celiac disease.     

Deficiency of the expression of CD45RA isoform of CD45 common leukocyte antigen in CD4+ T lymphocytes in children with infantile cholestasis

The immunological background of the pathological changes that appear in infantile cholestasis (infections, inflammatory process in the liver) is largely unknown. With the use of double color flow cytometry, we assessed the distribution of functionally different lymphocyte subpopulations in the peripheral blood of 29 infants with extra and intra-hepatic cholestasis (12 and 17 patients, respectively), aged from 1 to 8.6 months. Control group consisted of 15 age-matched, healthy infants. We examined: (1) the expression of CD3, CD4, CD8, CD19 lymphocyte surface receptors; and (2) the distribution of lymphocyte subsets with distinctive surface Ag characteristics of 'naive' (CD45RA+) and 'memory' (CD45RO+) cells in both CD4+ and CD8+ cell populations. The surface markers expression was evaluated in terms of percentage of positive cells and receptor density. The following changes in the expression of lymphocyte surface markers are described: (1) a decrease in the percentage of total CD3+, CD4+ cells but normal percentage of CD8+ cells and elevated proportion of CD19+ B cells; (2) a reduction of the proportion of 'naive' CD4+ lymphocytes but normal percentage of 'naive' CD8+ as well as 'memory' CD4+ and CD8+ cell subsets; (3) a decrease in density of CD3, CD4+, CD8 receptors, and D45RA isoform in a subset of 'naive' CD4+ cells. We conclude that deficiency of 'naive' CD4+ T cell subset which possess important effector and immunoregulatory functions, and low expression of certain lymphocyte receptors known to be engaged in T cell activation, possibly reflect a defect of cell mediated immunity that may account for viral and bacterial infections, often observed in infants with cholestasis.     

Identical expression of CD45R isoforms by CD45RC+ 'revertant' memory and CD45RC+ naive CD4 T cells.

Naive and memory CD4 T cells are frequently defined by exon-specific monoclonal antibodies (mAb) which stain (or not) high- or low-molecular-weight (MW) isoforms of the leucocyte common antigen CD45. The link between isoform and the naive/memory designation is complicated by the fact that CD4 T cells with a 'memory' phenotype (CD45RA-, RB-, RC-, or CD45RO+) may revert ('revertants') and re-express the high mw isoform (CD45RA+, RB+, RC+). Isoform expression also changes during normal T-cell development. Furthermore, the picture may be incomplete since an exon-specific mAb will not detect all possible isoforms on a cell. We have used molecular techniques to determine whether revertant CD4 memory T cells were different from naive T cells with respect to CD45R isoform expression. Using the anti-CD45RC mAb OX22 to purify rat lymphocyte subsets, CD45R isoform expression was examined at the mRNA level in CD4 T cells at different stages of development and compared with that of B cells and unseparated lymphocytes. B cells contained abundant message for the highest MW 3-exon isoform ABC, the 2-exon isoforms AB and BC, and the null isoform O. Both immature CD45RC- (i.e. CD4+8- 'single positive' thymocytes, and peripheral Thy-1+ recent thymic emigrants) and mature CD45RC- 'antigen-experienced' CD4 T cells had message for single-exons B, possibly C and for the O exon. In contrast, CD45RC+ CD4 T cells contained mRNA coding for ABC (low level), AB, BC, B, C (low level) and O (low level). Importantly, there was no difference between CD45RC+ T cells that had not seen antigen ('truly native') and CD45RC+ antigen-experienced revertant memory T cells. This observation has implications for understanding long-term immunological memory.     

Selective increase of activation antigens HLA-DR and CD38 on CD4+ CD45RO+ T lymphocytes during HIV-1 infection.

Infection with HIV results in a progressive depletion of CD4+ T cells and leads to significant in vivo lymphocyte phenotype changes. In this regard, the expression of HLA-DR and CD38 on CD8+ T cells has been shown to increase dramatically with disease progression. We investigated the expression of both activation markers on CD4+ T cells in HIV-1-infected subjects at different clinical stages of infection and compared the in vivo activation of CD4+ T cells with parameters of viral activity and CD8+ T cell activation. Fresh peripheral venous blood was obtained from 54 HIV-infected subjects and from 28 uninfected healthy controls. Three-colour immunophenotyping of the CD4+ T cell subset showed that the proportion of CD4+ T cells expressing HLA-DR (10% in HIV-negative controls) or CD38 (62% in HIV-negative controls) was higher in asymptomatic (P < 0.05 for CD38) and symptomatic (P < 0.001 for HLA-DR and CD38) HIV-infected subjects than in controls, whereas the proportion of CD4+ T cells expressing CD45RO (54% in controls) remained relatively unchanged. Simultaneous expression of HLA-DR and CD38 on CD4+ T cells increased from 2.3% in controls to 11% (P < 0.001) in asymptomatic and 22% (P < 0.001) in symptomatic HIV-infected subjects. This relative increase of CD38 and HLA-DR expression occurred mainly on CD4+ T cells co-expressing CD45RO. Changes in expression of HLA-DR and CD38 on CD4+ T cells correlated with similar changes on CD8+ T lymphocytes, with the presence of HIV antigen in the circulation, and with the disease stage of HIV infection.     

Evidence that the CD45 phosphatase regulates the activity of the phospholipase C in mouse T lymphocytes.

The importance of the tyrosine phosphatase CD45 in the regulation of lymphocyte activation was first demonstrated using antibodies against the extracellular domain of CD45 in functional assays. More recently it was reported that CD45-negative mutants were nonresponsive to stimulation through the T cell receptor-CD3 complex. We have studied the effect of CD45 cross-linking on the early signals induced by CD3 in mouse T cells. We show that CD45 cross-linking inhibits the increase in inositol phosphates and cytoplasmic Ca2+ induced by cross-linking of CD3. This indicates that CD45 is involved in the regulation of phospholipase C.     

Locomotor phenotypes of unstimulated CD45RAhigh and CD45ROhigh CD4+ and CD8+ lymphocytes in three-dimensional collagen lattices.

The spontaneous locomotion of immunomagnetically isolated resting CD4+ and CD8+ lymphocytes in three-dimensional collagen gels was recorded by time-lapse videomicroscopy. Two-dimensional projections of the paths of randomly selected individual cells were digitized, plotted and quantitatively analysed. Among five different donors 46 +/- 10% of CD4+ and CD8+ lymphocytes (n = 180) showed initial spontaneous locomotion (individual speed 5-25 microns/min; mean speed CD4+ 5.2 +/- 3.6 microns, CD8+ 3.23 +/- 2.72 microns). Active CD4+ cells were constantly migrating for more than 4 hr, whereas CD8+ lymphocytes significantly slowed down after 60-90 min in the collagen gel (P < 0.003). Quantitative analysis of the paths indicated at least three migratory phenotypes: (1) spontaneously locomoting cells exhibiting high speed and low frequency of stopping; (2) a major non-motile fraction without significant displacement; and (3) a subpopulation within CD4+ and CD8+ cells with intermediate activity of speed and stopping. Further subtype analysis of immunomagnetically isolated CD45RAhigh/ROlow or CD45RAlow/ROhigh lymphocytes showed that more than 90% of CD4+ CD45RAhigh/ROlow cells were actively locomoting. In contrast, only 20% of the CD4+ CD45ROhigh/RAlow phenotype showed spontaneous motility to a limited degree. The data indicate that resting CD4+ and CD8+ lymphocytes comprise further locomotory subpopulation related to the expression of different CD45 isoforms.     

Alloreactive CD4 T lymphocytes responsible for acute and chronic graft-versus-host disease are contained within the CD45RChigh but not the CD45RClow subset

Graft-versus-host disease (GvHD) is a major complication of allogeneic bone marrow transplantation and occurs when donor T cells react with histo-incompatible recipient's antigens. In the present study, we analyzed the contribution of CD4 T cell subsets, defined according to their CD45RC expression level, in the development of acute and chronic GvHD. For this purpose, we used the model of GvHD induced in rats when parental lymphocytes are transferred to irradiated (LEWxBN) F1 hybrid recipients. We showed that parental CD45RC(high) (naive cells) CD4 T cells induced both acute and chronic GvHD while CD45RC(low) (memory cells) subset did not. In vitro, only CD45RC(high) CD4 T cells proliferated and produced cytokines in response to alloantigen stimulation. LEW and BN CD45RC(high) CD4 T cells produced different cytokine profiles in response to in vitro allostimulation, which could explain their ability to induce different forms of GvHD. Finally, we showed that memory CD45RC(low) CD4 T cells, known to contain regulatory T cells, were unable to prevent GvHD induction. Together these data show that memory CD45RC(low) CD4 T cells do not contain functional alloreactive T cells and suggest that selective transfusion of donor memory cells could greatly improve post-transplant immune reconstitution without risk of GvHD induction.