Immune Class Regulation and Its Medical Significance Part II of a Report of a Workshop on Foundational Concepts of Immune Regulation.

We discussed different proposals for how the nature of the Th1/Th2 phenotype of an immune response is determined, and favoured one, the Threshold Hypothesis, as plausible and so useful as the basis for further discussions. The activation of a target CD4 T cell can be facilitated by helper CD4 T cells when the CD4 T cells interact via an antigen-presenting cell. The Threshold Hypothesis states that tentative and robust antigen-mediated CD4 T cell cooperation results in the target CD4 T cell, respectively giving rise, upon activation, to Th1 and Th2 cells. We primarily discussed four topics. We briefly discussed in the background section certain limitations of the Th1/Th2 paradigm in understanding immune class regulation, and the remarkable anti-inflammatory properties of human IgG4 antibody. Secondly, we assessed the role of class II MHC molecules in determining the number of mature CD4 T cells and so affecting the Th1/Th2 phenotype of immune responses. We also discussed the controversial role of CD8 T cells in affecting the Th1/Th2 phenotype of responses to MHC and other antigens, and the potential role of their relative scarcity in neonates in biasing responses towards an antibody, Th2 mode. Lastly, we examined the regulation of the Th1/Th2 phenotype of both primary and ongoing immune responses in the context of the intriguing proposal that antigen initially generates different classes/subclasses of immunity and then selects, by a feedback mechanism, the most effective class. We found this interesting idea difficult to reconcile with various observations.

Immunological Tolerance. Part I of a Report of a Workshop on Foundational Concepts of Immune Regulation.

This report, the first of two, arose from a one-week workshop directed at discussing concepts of immune regulation, and focuses on immunological tolerance. We first outline the major ideas we thought sufficiently plausible to provide a context for discussing more controversial issues around tolerance. We then report on our discussion of different experiments that appear paradoxical in terms of the different, contemporary models of CD4 T cell inactivation/activation, and how such observations might be resolved in terms of insights provided by these contemporary models. These discussions bear on the plausibility of the Pathogen-Associated Molecular Pattern (PAMP), Danger and Two Step, Two Signal Models for the activation of naïve CD4 T cells. Some of the observations considered appear paradoxical in terms of the PAMP and Danger Models, but not with the Two Step, Two Signal Model. For example, genetically immunodeficient mice have been given foreign, sterile ectopic grafts, and the immune system allowed to develop once these grafts were well-healed in, and so in the absence of PAMPs or danger. The grafts were rejected, unexpected on the PAMP or Danger Models. We also discussed considerations and observations bearing on the widely held idea that antigen must crosslink the membrane Ig receptors of a B cell to initiate the generation of signal 1, or the alternative possibility that monovalent binding by antigen can do so. We favored the latter possibility, and discussed a particular model, "the Elbow Model," for how this might be achieved.

Genetic Polymorphisms in the Toll-like Receptor 10, Interleukin (IL)17A and IL17F Genes Differently Affect the Risk for Tuberculosis in Croatian Population.

Proinflammatory conditions leading to activation of macrophages via interferon-γ bear an important role in host defence against intracellular bacteria such as Mycobacterium tuberculosis (Mt). Interleukin-17 plays a similar role, as it appears to be also an activator of macrophages. Recently, the TLR-10 was identified as an anti-inflammatory factor that exerts its action via association with the TLR-2 chain at the cell surface of macrophages, the latter being an Mt-binding protein. We have previously found that gene polymorphisms that either inactivate the TLR2 gene product or have a dominant-negative role are associated with tuberculosis (TB) in Croatian population. We have now extended our survey and found that single nucleotide polymorphism (SNP) in TLR10 (rs11096957) is associated with risk for TB. Homozygotes carrying the A allele are associated with predisposition to disease as analysed by the dominant model of inheritance. In contrast, SNPs in the proinflammatory IL17A and IL17F genes (rs2275913 and rs763780, respectively), found previously to correlate with the disease occurrence in Chinese population, were not significantly associated with tuberculosis in the Croatian population.

Polymorphisms in the interleukin-1 gene locus and chronic periodontitis in patients with atherosclerotic and aortic aneurysmal vascular diseases.

Chronic periodontitis (CP) and atherosclerotic and aortic aneurysmal vascular diseases (VD) are inflammatory conditions that share a number of predisposing factors. They have a complex genetic heritability and may share genetic risk factors, but a well-defined relationship is still not determined. In addition, distinct genetic patterns of predisposition have been associated with these diseases. Here, we investigated the association of polymorphisms in the IL-1 gene locus with CP in a case-case study analysing VD patients with or without CP. Seventy-four patients with VD of whom 36 had CP were genotyped for single nucleotide polymorphisms in the IL1A -889 (rs1800587), IL1B +3954 (rs1143634) and IL1B at -511 (rs16944) genes and for VNTR polymorphisms in the IL1RN gene. A significantly higher frequency (17%) for allele 1 (four repeats) of the IL1RN VNTR gene was found among the VD patients with CP compared to those without CP. In addition, the frequency of the IL1RN VNTR genotypes 1/1 (4/4 repeats) and 2/2 (2/2 repeats) were significantly higher and lower, respectively, in VD patients with CP. These findings suggest an association of genetic polymorphisms in the IL1-gene locus with risk for CP in patients with VD. The carriage of the risk genotypes, the development and the subsequent influence of CP on systemic health may constitute an additional burden in the pathogenesis of VD. This emphasizes the importance of effective periodontal treatment in patients with VD.

On recognizing 'shades-of-gray' (self-nonself discrimination) or 'colour' (Integrity model) by the immune system.

The aim is to discuss Cohn's T-cell receptor (TCR) Tritope model of recognition, propose a novel suggestion for prior-to-positive selection of thymocytes contributing to inherent major histocompatibility complex (MHC) reactivity of a T-cell repertoire and clarify the Integrity model about the function of the immune system. If we compare the perception of light with the recognition of nonself, we could imagine that the opacity might be a measure of docking interaction between specific receptors for antigen on T or B cells (TCR/peptide-MHC or BCR/antigen). From this viewpoint, the self-nonself discrimination (S-NS) metaphor would be perception of black (self) versus white (nonself). However, whereas detection of shades-of-gray suffices to describe S-NS discrimination principle, colour vision of the antigenic world portrays best the Integrity model. In concert with recognition of opacity, the Integrity model proposes detection of at least three colours (signals): red (harmful), blue (useful) and yellow (the rest, including homoeostatic ones). As a result, recognition of nonself is transferred into communication within self while deciding on type of the immune response. Hence, the S-NS discrimination model seems to be an oversimplification, because it fails to see colours and consequently lacks the need for suppressor/regulatory function. Similarly, the Danger model stops short of detecting being useful signals that confer immune asylum to helpful micro-organisms like commensals. I suggest that the immune system's repertoire for recognition, in general, has evolved by a novel drive called 'natural integrity' alongside natural selection, thus facilitating communication between cells of the immune system.

Associations of the interleukin-1 gene locus polymorphisms with risk to hip and knee osteoarthritis: gender and subpopulation differences.

Genetic predisposition to the complex hereditary disease like osteoarthritis (OA) of the large joints (hip and knee) includes the interleukin-1 gene (IL-1) cluster on chromosome 2. Using a case-control study with 500 OA patients (240 knee and 260 hip OA patients, all with joint replacement), we analysed frequencies of IL-1 gene cluster polymorphisms in Croatian Caucasian population. The control samples came from 531 healthy individuals including blood donors. We genotyped two single nucleotide polymorphisms in the IL-1 gene locus at IL-1A (-889, C>T, rs1800587) and IL-1B (+3594, C>T, rs1143634) and compared their frequencies between patients and controls. We predicted haplotypes by combining current data with our previous results on gene polymorphisms (IL-1B, rs16944 and the IL-1 receptor antagonist gene [IL-1RN] variable number tandem repeat [VNTR]) for the same population. Haplotype analyses revealed gender disparities and showed that women carriers of the 1-2-1-1 haplotype [IL-1A(rs1800587) - IL-1B(rs1143634) - IL-1B(rs16944) - IL-1RN(VNTR)] had sixfold lower risk to develop knee OA. However, carriers of the 1-1-1-2 haplotype of both sexes had over twofold higher predisposition to hip OA. Our results differ from some earlier studies in Caucasian subpopulations, which may be due to the fact that this is the first study to separate genders in assessing the IL-1-locus genetic risk of OA. The results suggest that inflammatory mediators like IL-1 might be implicated in the pathogenesis of primary OA in large joints and that as yet unidentified gender-specific factors exist in a Croatian Caucasian population.

Single nucleotide polymorphism in the interleukin 12B gene is associated with risk for breast cancer development.

We analysed the association of a single nucleotide polymorphism (SNP) in the gene encoding the IL-12 subunit p40 (IL12B, rs3212227, A>C) with breast cancer. The SNPs allelic and genotypic frequencies were compared between patients (n = 191) and healthy (n = 194) women in a case-control study from Croatia. The major allele (A) was associated with susceptibility to breast cancer (P = 0.003; OR = 1.67; 95% CI: 1.17-2.38). Likewise, the minor allele (C) was significantly correlated with protection (P = 0.003; OR = 0.60; 95% CI: 0.42-0.86). At the genotype level, AA homozygosity was significantly associated with predisposition to disease (P = 0.013; OR = 1.68, 95% CI: 1.09-2.59), whereas the minor allele homozygosity (CC) was correlated with protection to disease (P = 0.020, OR = 0.28, 95% CI: 0.09-0.91). The heterozygous genotype showed no significant correlation with disease. The product of the IL12B gene (IL-12 p40) can either form a homodimeric cytokine or be part of two pro-inflammatory (IL-12 and IL-23) cytokines. It is presently unclear whether the major allele is associated with higher or lower protein levels of IL-12 p40 and IL-12 p70, which are critical in inflammation and adaptive immune responses. However, as the A allele is high producer of IL12B (p40) mRNA, these results might imply that higher levels of IL-12 p40 (either as homodimers or joined with one or both of the other two subunits) predispose to breast cancer.

Association studies of gene polymorphisms in toll-like receptors 2 and 4 in Croatian patients with acute myocardial infarction.

The aim of the study was to assess the frequency of SNP896A/G in the Toll-like receptor (TLR) 4 gene and SNP1350T/C in the TLR2 gene in patients with acute myocardial infarction (AMI) and to analyse the association of these SNPs with risk factors for atherosclerosis and clinical aspects of AMI in a sample of the Croatian population. We included 240 participants in the study: 120 AMI patients and 120 sex- and age-matched healthy blood donor controls. The SNP1350T/C variant in the TLR2 gene showed a lower frequency in the AMI patient group than in the control group (P = 0.033). The frequency of SNP896A/G variants in the TLR4 gene between the patients and the controls did not differ (P = 0.286). Significantly, fewer people had SNP1350T/C in the TLR2 gene (P = 0.003) among the participants with arterial hypertension than those without it. The frequency of SNP896A/G in TLR4 was the same in hypertensive patients compared with normotensive subjects (P = 0.088). SNP1350T/C in TLR2 was less frequent in the AMI patients and in those with hypertension. Thus, SNP1350T/C in TLR2 might play a protective role against AMI and arterial hypertension. The frequency of SNP896A/G in the TLR4 gene was not associated with AMI and arterial hypertension. Other risk factors for atherosclerosis and clinical aspects of myocardial infarction were not associated with the genotype distribution of the examined genes.

Toll-like receptor 2 (P631H) mutant impairs membrane internalization and is a dominant negative allele.

We have sequenced 416 Toll-like receptor-2 (TLR2) alleles in 208 subjects in a tuberculosis case-control study in Croatian Caucasian population. We found ten single nucleotide polymorphisms (SNP) among which three were novel (S97S, T138I and L266F). The genotype containing TLR2-P631H SNP was significantly overrepresented in patients with tuberculosis when compared to contact controls, suggesting a small yet increased risk to disease. The causative agent of tuberculosis is Mycobacterium tuberculosis, which can bind to TLR2 with its lipoprotein coat. The TLR2-P631H mutant has a dominant negative effect on the wild type TLR2 signalling in transfected HEK293 kidney cells using the NF-kappaB-driven luciferase as a reporter gene with ligands like M. avium extracts, Pam3CysSK4 or FSL-1 that bind TLR2/TLR1 or TLR2/TLR6 heterodimers, respectively. Studies on internalization from the Regular Madine Darby Canine Kidney cell surface into the early endosomal compartments showed a lower rate of the mutant compared to the wild type. Our data, in combination with a report by others show that the TLR2-P631H allele could be associated with protection to meningococcal meningitis, suggest that by dominantly inhibiting the response of cells important in the immune response this mutant might confer either protection or susceptibility to meningitis or tuberculosis, respectively.

Beginning of the end of (understanding) the immune response.

The algorithm about the workings of the immune system should incorporate in its basic form the ability to explain several basic features: the ability to reject pathogens, to respond differently to invasions by class switching, to retain memory of rejecting past infections, the regulatory T-cell formation and the acquisition of a life-long antigen tolerance in adults. The Danger and the two self-non-self theories - Pathogenicity (Janeway) and the Associated Antigen Recognition (Langman and Cohn) - cannot explain satisfactorily one or more of these issues. By failing to find appropriate solutions to help their corresponding basic algorithms to be operative, these theories negate, instead of finding explanations (no matter how complicated) for interpretations of certain results, one of them being the existence of a regulatory cell as a physiological component of the immune system. On the other hand, the integrity model can incorporate regulatory T cells in a basic algorithm of the workings of the immune system, explaining it as a downregulation of the immune response provided advantageous commensalism prevails over disadvantages of parasitism.

Interferon-gamma gene (T874A and G2109A) polymorphisms are associated with microscopy-positive tuberculosis.

Genetic susceptibility to tuberculosis includes several unknown yet different loci each contributing to a small extent. Intronic polymorphisms within the interferon-gamma (IFN-gamma) gene IFNG T+874A and IFNG G+2109A correlate with the IFN-gamma production in vitro, and the frequency of potential high IFN-gamma producers was previously reported by others to be lower in patients than in controls from Sicily. The aim of this study was to determine whether there is an association between polymorphisms in the IFN-gamma gene and predisposition to tuberculosis. We analysed two IFNG SNPs (T+874A and G+2109A) in patients (n = 253) hospitalized in Rijeka (Croatia) and controls (n = 519) from the same area. One-fifth of the controls were healthy contacts of the diseased, and the rest were blood donors. IFNG alleles, their predicted haplotypes or genotypes were not associated with disease susceptibility. Thus, we could not reproduce results from Sicilian case-control study. However, T/T+874 (possible high IFN-gamma producer) and +874A/A (putative low producer) genotypes were associated with microscopically positive-negative forms of disease. Haplotypes (T+874A and G+2109A) based on a prediction by software phase and subsequent genotype analysis corroborated these findings. Patients had significantly higher frequency of genotypes without T at +874 (AA/AA; AA/AG and AG/AG) in microscopy- or bacterial culture-positive groups compared with their negative counterparts. These data suggest an association with disease severity rather than susceptibility to tuberculosis in Croatian Caucasian population.

Interferon-gamma receptor-1 gene promoter polymorphisms (G-611A; T-56C) and susceptibility to tuberculosis.

We analysed frequencies of two single-nucleotide polymorphisms (SNP) in the interferon-gamma (IFN-gamma) receptor-1 (IFNGR1) gene promoter (G-611A, T-56C) in tuberculosis patients (n = 244) and compared them with controls (n = 521). These frequencies were not significantly different, whether analysed independently or as haplotypes. Because these SNP affect transcription, the results suggest that the expression of the IFNGR1 gene does not confer susceptibility to disease in patients from Croatia. Further analysis revealed a significant association between the protective (CA)(n) polymorphism (22 repeats, 192 FA(1)), located in the fifth intron of the IFNGR1 gene (+16682), and GT promoter haplotype (-611; -56) that showed the strongest expression capacity. In addition to this cis relationship, the (CA)(22) allele was correlated in trans with an IFN-gamma SNP (IFNG G + 2109A), which might affect the transcription of the IFNG gene. These results suggest that a particular combination of IFNG and IFNGR1 SNP might offer a better protection against tuberculosis in this population.

Response to Cohn: The immune system rejects the harmful, protects the useful and neglects the rest of microorganisms.

The immune system is seen as a guardian of tissue integrity. It would analyse the extent and quality of damage and respond adequately. If no ill effects were found, the system would ignore disturbance, but if beneficial effects were found, it could protect certain microorganisms (establishing commensalism), perhaps via regulatory cells. The Integrity hypothesis proposes three basic groups of intercellular signals for cells of all tissues and assumes that they govern communication between dendritic cells, T cells and B cells. Signal-1 would be the main information source resulting with generation of intracellular mediators that are bound to travel into the nucleus to achieve reaction. Signal-2 represents the generation of additional signal transducers representing a modifier at the level of cytosol. And, signal-3 would be a modifier at nuclear level, perhaps guarding accessibility to chromosome or genetic locus.

Anti-class II antibodies, but not cytotoxic T-lymphocyte antigen 4-immunoglobulin hybrid molecules, prevent rejection of major histocompatibility complex class II-negative myeloma in T-cell receptor-transgenic mice.

We have previously shown that tumour-specific CD4+ T cells protect against subcutaneous injections of major histocompatibility complex (MHC) class II-negative MOPC315 myeloma cells. Here, we have interfered with the immunologic events that lead to successful rejection of MOPC315 challenges in T-cell receptor (TCR)-transgenic mice. The CD4+ T cells have a transgene-encoded TCR specific for a MOPC315 V-region idiotypic (Id) peptide presented on the MHC class II molecule E(d). A side-by-side comparison indicated that DNA-recombination-deficient TCR-transgenic mice were better protected against MOPC315 tumour development than recombination-sufficient counterparts, suggesting that B cells or endogenous TCR chains might facilitate tumour progression in this model. Intraperitoneal injections of E(d)-specific antibodies over a period of initial 24 days, abrogated protection against tumours in both strains of mice. By contrast, injections of anticostimulatory molecules (cytotoxic T-lymphocyte antigen 4-immunoglobulin hybrid molecules) had no effect. The findings demonstrate that tumour rejection depends on the presence of MHC class II molecules, despite the fact that MOPC315 tumour cells themselves do not express them. The results are consistent with the idea that secreted myeloma protein is processed and presented by class II+ antigen-presenting cells to Id-specific naïve CD4+ T cells that become activated and kill the myeloma cells by a bystander mechanism. While Id presentation on class II molecules is absolutely required for tumour rejection, costimulatory CD80/CD86 molecules might be dispensible in this process.

Liver metastasis of cancer facilitated by chemokine receptor CCR6.

When injected subcutaneously, mouse plasmacytoma (MOPC315) grew rapidly in situ, and metastatic cells became detectable first in the lymph nodes (LNs) and bone marrow, and later in the liver and lungs. We studied MOPC315 cell migration by tracking metastatic cells labelled with green fluorescent protein (GFP). We measured the levels of their chemokine receptor mRNA (by semiquantitative and real-time quantitative reverse transcriptase-polymerase chain reaction (RT-PCR), because chemokines can regulate organ predilection of metastasis. Freshly sorted metastatic cells and tumour cell lines derived from the liver of BALB/c mice overexpressed functional CCR6 and CCR7 molecules compared with primary tumour. Preincubation with the CCR6 ligand (CCL20) induced liver-sorted tumour cells to preferentially colonize the liver, demonstrating an association between liver metastasis and CCR6 expression in the mouse. Because the liver is a common site for metastasis, second only to draining LNs, we wished to ascertain whether this finding could be generalized, i.e. whether other cancers can use the similar mechanism of metastasis to the liver, and whether it holds true for humans. We found that CCR6 is overexpressed in small liver metastases of colon, thyroid and ovarian carcinomas compared with normal liver. Because human liver constitutively expresses CCL20, it could attract and select CCR6+ cancer cells. We suggest that chemotaxis via CCR6 might be a common mechanism by which malignant cancers metastasize to the liver. As metastasis in patients with cancer poses the biggest peril for survival, inhibition of CCR6 signalling, either during or after medical or surgical treatment, might be useful in preventing liver metastasis.

Interferon-gamma receptor-1 gene polymorphism in tuberculosis patients from Croatia.

Recent studies have indicated that the interleukin-12/interferon-gamma (IFN-gamma) axis is important in mycobacterial infection susceptibility. Using an intronic (CA)n polymorphic microsatellite marker within the IFN-gamma receptor-1 (IFNGR1) gene, we have compared the allelic frequencies of this marker in hospitalized tuberculosis patients (n = 120) with that of controls (n = 87) from Rijeka, Croatia. We identified 13 (CA)n alleles in the tuberculosis patients, whereas only 10 were found in the controls. A significant difference between one allelic marker and the control group was observed (P = 0.02, 95% confidence interval 0.14-0.94), suggesting a possible protective association. In contrast, several other allelic markers showed a trend towards association with the disease. We also found a trend towards an increased frequency in homozygosity of one allelic marker in patients (11.7%) as compared with controls (4.6%). We conclude that there is no evidence for disease association of the IFNGR1 gene marker in Mendelian-type (single-allele) inheritance. However, our results also suggest that unidentified allelic variations in the IFNGR1 gene might elevate or decrease the risk in this ethnic population, as a part of the multigenic predisposition to tuberculosis.

Phagocytic dendritic cells from myelomas activate tumor-specific T cells at a single cell level.

Antigen-presenting cells (APCs) from subcutaneous mouse MOPC315 plasmacytoma phagocytosed immunoglobulin G-coated magnetic beads, enabling efficient isolation within 2 hours by magnetic separation (APC-MB). Cell morphology was heterogeneous, with some of the cells having dendrites. The surface phenotype of purified tumor APCs-MB was CD11b(+), CD11c(+), CD40(+), CD80(+), CD86(+), and MHC class II(+). Tumor APCs-MB expressed messenger RNA for fractalkine and ABCD-1 chemokines, and for CC-type chemokine receptors CCR5 and CCR7, indicating the presence of mature dendritic cells (DCs). Visualized at a single cell level within 4 hours after disruption of the tumor, APCs-MB induced rapid Ca(++) mobilization in MHC class II-restricted tumor idiotype (Id)-specific cloned CD4(+) T cells. In long-term assays, tumor APCs-MB induced proliferation of naive T cells from Id-specific T-cell receptor transgenic mice. The results suggest that tumor APCs-MB represent a heterogeneous cell population that includes myeloid-derived DCs of various stages of maturation. A considerable fraction (> or = 15%) of DCs is spontaneously primed with tumor-specific antigen.

Dendritic cells purified from myeloma are primed with tumor-specific antigen (idiotype) and activate CD4+ T cells.

Multiple myelomas produce tumor-specific antigen (TSA) in the form of idiotype (Id) on monoclonal Ig. CD4(+) T cells can recognize Id-peptide on MHC class II molecules and protect against challenges with MOPC315 cells, which are, as common for myelomas, class II-negative. The present study explains these previous results by demonstrating that Id can be transferred from myeloma cells to antigen-presenting cells (APC), which present processed Id-peptide on their class II molecules to Id-specific T cell receptor-transgenic (TCR-TG) CD4(+) T cells. Id-primed tumor APC were heterogeneous, the majority being dendritic cells with class II(+), CD11b(+) CD11c(+) CD40(+) CD80(+) CD86(+) markers. The APC were localized beneath CD31(+) endothelial cells of tumor microvessels, and their frequency declined with tumor progression. The APC could stimulate Id-specific naive TCR-TG, short-term polarized TCR-TG, and cloned CD4(+) T cells to proliferate and produce cytokines in vitro. Furthermore, small MOPC315 tumors established in Id-specific TCR-TG mice contained clusters of activated (CD69(+)CD25(+)) and proliferating (BrdUrd(+)) Id-specific transgenic CD4(+) blasts. The activated Id-specific T cells were located adjacent to Id-primed dendritic cells in the tumor. Thus, a TSA can be transferred in vivo from myeloma, and possibly other types of cancer cells to APC for MHC class II presentation to CD4(+) T cells.

Deletion of idiotype (Id)-specific T cells in multiple myeloma.

Mycloma cells secrete monoclonal immunoglobulin (Ig), called myeloma protein. The variable (V) regions of myeloma proteins are unique to each plasma cell tumor, and therefore contain highly tumor-specific antigenic determinants called idiotopes (Id). T cells with specificity for Id are thought to be of importance in eradication of multiple myeloma. In ongoing clinical trials, myeloma patients are vaccinated against the Id of their own myeloma protein, with the aim of inducing Id-specific T cells. However, this strategy will only succeed if Id-specific T cells are present in patients, and are able to respond. In an experimental animal model, we have shown that [d-specific T cells become progressively deleted as the myeloma protein serum concentration exceeds 50 microg/ml. This indicates that the ability of multiple myeloma patients to respond to Id-vaccination might be seriously handicapped. We suggest that Id-vaccination should be reserved for eradication of minimal residual disease, e.g. after high-dose chemotherapy and stem-cell transplantation.