Allografts stimulate cross-reactive virus-specific memory CD8 T cells with private specificity.

Viral infections have been associated with the rejection of transplanted allografts in humans and mice, and the induction of tolerance to allogeneic tissues in mice is abrogated by an ongoing viral infection and inhibited in virus-immune mice. One proposed mechanism for this 'heterologous immunity' is the induction of alloreactive T cell responses that cross-react with virus-derived antigens. These cross-reactive CD8 T cells are generated during acute viral infection and survive into memory, but their ability to partake in the immune response to allografts in vivo is not known. We show here that cross-reactive, virus-specific memory CD8 T cells from mice infected with LCMV proliferated in response to allografts. CD8 T cells specific to several LCMV epitopes proliferated in response to alloantigens, with the magnitude and hierarchy of epitope-specific responses varying with the private specificities of the host memory T cell repertoire, as shown by adoptive transfer studies. Last, we show that purified LCMV-specific CD8 T cells rejected skin allografts in SCID mice. These findings therefore implicate a potential role for heterologous immunity in virus-induced allograft rejection.

Human peripheral blood leucocyte non-obese diabetic-severe combined immunodeficiency interleukin-2 receptor gamma chain gene mouse model of xenogeneic graft-versus-host-like disease and the role of host major histocompatibility complex.

Immunodeficient non-obese diabetic (NOD)-severe combined immune-deficient (scid) mice bearing a targeted mutation in the gene encoding the interleukin (IL)-2 receptor gamma chain gene (IL2rgamma(null)) engraft readily with human peripheral blood mononuclear cells (PBMC). Here, we report a robust model of xenogeneic graft-versus-host-like disease (GVHD) based on intravenous injection of human PBMC into 2 Gy conditioned NOD-scid IL2rgamma(null) mice. These mice develop xenogeneic GVHD consistently (100%) following injection of as few as 5 x 10(6) PBMC, regardless of the PBMC donor used. As in human disease, the development of xenogeneic GVHD is highly dependent on expression of host major histocompatibility complex class I and class II molecules and is associated with severely depressed haematopoiesis. Interrupting the tumour necrosis factor-alpha signalling cascade with etanercept, a therapeutic drug in clinical trials for the treatment of human GVHD, delays the onset and progression of disease. This model now provides the opportunity to investigate in vivo mechanisms of xenogeneic GVHD as well as to assess the efficacy of therapeutic agents rapidly.

Non-obese diabetic-recombination activating gene-1 (NOD-Rag1 null) interleukin (IL)-2 receptor common gamma chain (IL2r gamma null) null mice: a radioresistant model for human lymphohaematopoietic engraftment.

Immunodeficient hosts engrafted with human lymphohaematopoietic cells hold great promise as a preclinical bridge for understanding human haematopoiesis and immunity. We now describe a new immunodeficient radioresistant non-obese diabetic mice (NOD) stock based on targeted mutations in the recombination activating gene-1 (Rag1(null)) and interleukin (IL)-2 receptor common gamma chain (IL2rgamma(null)), and compare its ability to support lymphohaematopoietic cell engraftment with that achieved in radiosensitive NOD.CB17-Prkdc(scid) (NOD-Prkdc(scid)) IL2rgamma(null) mice. We observed that immunodeficient NOD-Rag1(null) IL2rgamma(null) mice tolerated much higher levels of irradiation conditioning than did NOD-Prkdc(scid) IL2rgamma(null) mice. High levels of human cord blood stem cell engraftment were observed in both stocks of irradiation-conditioned adult mice, leading to multi-lineage haematopoietic cell populations and a complete repertoire of human immune cells, including human T cells. Human peripheral blood mononuclear cells also engrafted at high levels in unconditioned adult mice of each stock. These data document that Rag1(null) and scid stocks of immunodeficient NOD mice harbouring the IL2rgamma(null) mutation support similar levels of human lymphohaematopoietic cell engraftment. NOD-Rag1(null) IL2rgamma(null) mice will be an important new model for human lymphohaematopoietic cell engraftment studies that require radioresistant hosts.

A new immunodeficient hyperglycaemic mouse model based on the Ins2Akita mutation for analyses of human islet and beta stem and progenitor cell function.

AIMS/HYPOTHESIS: To develop and validate a new immunodeficient mouse strain that spontaneously develops a non-autoimmune hyperglycaemia to serve as a diabetic host for human islets and human beta stem and progenitor cells without the need for induction of hyperglycaemia by toxic chemicals with their associated side effects. METHODS: We generated and characterised a new strain of immunodeficient spontaneously hyperglycaemic mice, the NOD-Rag1null Prf1null Ins2Akita strain and compared this strain with the NOD-scid Il2rgammanull (also known as Il2rg) immunodeficient strain rendered hyperglycaemic by administration of a single dose of streptozotocin. Hyperglycaemic mice were transplanted with human islets ranging from 1,000 to 4,000 islet equivalents (IEQ) and were monitored for normalisation of blood glucose levels. RESULTS: NOD-Rag1null Prf1null Ins2Akita mice developed spontaneous hyperglycaemia, similar to Ins2Akita-harbouring strains of immunocompetent mice. Histological examination of islets in the host pancreas validated the spontaneous loss of beta cell mass in the absence of mononuclear cell infiltration. Human islets transplanted into spontaneously diabetic NOD-Rag1null Prf1null Ins2Akita and chemically diabetic NOD-scid Il2rgammanull mice resulted in a return to euglycaemia that occurred with transplantation of similar beta cell masses. CONCLUSIONS/INTERPRETATION: The NOD-Rag1null Prf1null Ins2Akita mouse is the first immunodeficient, spontaneously hyperglycaemic mouse strain described that is based on the Ins2Akita mutation. This strain is suitable as hosts for human islet and human beta stem and progenitor cell transplantation in the absence of the need for pharmacological induction of diabetes. This strain of mice also has low levels of innate immunity and can be engrafted with a human immune system for the study of human islet allograft rejection.

Different mechanisms control peripheral and central tolerance in hematopoietic chimeric mice.

Regulatory T cells (Treg) are important in peripheral tolerance, but their role in establishing and maintaining hematopoietic mixed chimerism and generating central tolerance is unclear. We now show that costimulation blockade using a donor-specific transfusion and anti-CD154 antibody applied to mice given bone marrow and simultaneously transplanted with skin allografts leads to hematopoietic chimerism and permanent skin allograft survival. Chimeric mice bearing intact skin allografts fail to generate effector/memory T cells against allogeneic targets as shown by the absence of IFNgamma-producing CD44(high)CD8+ T cells and in vivo cytotoxicity. Depletion of Tregs by injection of anti-CD4 or anti-CD25 antibody prior to costimulation blockade prevents chimerism, shortens skin allograft survival and leads to generation of effector/memory cytotoxic T cells. Depletion of Tregs by injection of anti-CD4 or anti-CD25 antibody two months after transplantation leads to loss of skin allografts even though mice remain chimeric and exhibit little in vivo cytotoxicity. In contrast, chimerism is lost, but skin allografts survive following naïve T-cell injection. We conclude that hematopoietic chimerism and peripheral tolerance may be maintained by different mechanisms in mixed hematopoietic chimeras.

Costimulation and autoimmune diabetes in BB rats.

Costimulatory signals regulate T-cell activation. To investigate the role of costimulation in autoimmunity and transplantation, we studied the BB rat model of type 1 diabetes. Diabetes-prone BB (BBDP) rats spontaneously develop disease when 55-120 days of age. We observed that two anti-CD28 monoclonal antibodies (mAb) with different functional activities completely prevented diabetes in BBDP rats. Anti-CD154 mAb delayed diabetes, whereas treatment with CTLA4-Ig or anti-CD80 mAb accelerated disease. Anti-CD86 or anti-CD134L mAbs had no effect. Diabetes resistant BB (BBDR) rats are disease-free, but >95% of them develop diabetes after treatment with polyinosinic-polycytidylic acid and an mAb that depletes Treg cells. In the induced BBDR model, anti-CD154 mAb delayed onset of diabetes, whereas CTLA4-Ig, anti-CD134L or either of the anti-CD28 mAbs had little or no effect. In contrast, blockade of the CD134-CD134L pathway was highly effective for preventing autoimmune recurrence against syngeneic islet grafts in diabetic BBDR hosts. Blockade of the CD40-CD154 pathway was also effective, but less so. These data suggest that the effectiveness of costimulation blockade in the treatment of type 1 diabetes is dependent on both the costimulatory pathway targeted and the mechanism of induction, stage, intensity and duration of the pathogenic process.

Skin allograft maintenance in a new synchimeric model system of tolerance.

Treatment of mice with a single donor-specific transfusion plus a brief course of anti-CD154 mAb uniformly induces donor-specific transplantation tolerance characterized by the deletion of alloreactive CD8+ T cells. Survival of islet allografts in treated mice is permanent, but skin grafts eventually fail unless recipients are thymectomized. To analyze the mechanisms underlying tolerance induction, maintenance, and failure in euthymic mice we created a new analytical system based on allo-TCR-transgenic hemopoietic chimeric graft recipients. Chimeras were CBA (H-2(k)) mice engrafted with small numbers of syngeneic TCR-transgenic KB5 bone marrow cells. These mice subsequently circulated a self-renewing trace population of anti-H-2(b)-alloreactive CD8+ T cells maturing in a normal microenvironment. With this system, we studied the maintenance of H-2(b) allografts in tolerized mice. We documented that alloreactive CD8+ T cells deleted during tolerance induction slowly returned toward pretreatment levels. Skin allograft rejection in this system occurred in the context of 1) increasing numbers of alloreactive CD8+ cells; 2) a decline in anti-CD154 mAb concentration to levels too low to inhibit costimulatory functions; and 3) activation of the alloreactive CD8+ T cells during graft rejection following deliberate depletion of regulatory CD4+ T cells. Rejection of healed-in allografts in tolerized mice appears to be a dynamic process dependent on the level of residual costimulation blockade, CD4+ regulatory cells, and activated alloreactive CD8+ thymic emigrants that have repopulated the periphery after tolerization.

Levels of Art2+ cells but not soluble Art2 protein correlate with expression of autoimmune diabetes in the BB rat.

ART2a and ART2b are isoenzymes expressed on the surface of mature T cells and intraepithelial lymphocytes (IELs) in the rat. They exhibit both adenosine diphosphoribosyltransferase and nicotine adenine dinucleotide (NAD) glycohydrolase activities, and both can generate a transmembrane signal that modulates T cell activation. The presence or absence of ART2+ T cells modulates the expression of autoimmune diabetes in the BB rat. ART2 also circulates in a soluble form whose function is unknown. We tested the hypothesis that circulating ART2 protein regulates the expression of autoimmunity. We compared the kinetics, regulation, and source of soluble ART2 in normal rats and in rats with autoimmune diabetes. Basal levels of soluble ART2 varied greatly among strains of rats and were lowest in the diabetes-prone BB (BBDP/Wor) rat. In diabetes-resistant BB (BBDR/Wor) rats, administration of anti-ART2a antibody, which is known to induce diabetes, resulted in transient clearing of soluble ART2a that was followed rapidly by a rebound increase. Repeated treatment of BBDR/Wor rats with anti-ART2a antibody resulted in sustained supraphysiologic levels of soluble ART2a. Although the number of peripheral ART2a+ T cells is known to correlate with the expression of diabetes in BBDR/Wor rats, the level of soluble ART2a protein did not. The source of the soluble ART2 protein in the rat appeared to be the gut. The results suggest that ART2+ T cells and soluble ART2 protein may subserve different immunomodulatory functions.

Requirement of the JIP1 scaffold protein for stress-induced JNK activation.

The c-Jun N-terminal kinase (JNK) signal transduction pathway is activated in response to the exposure of cells to environmental stress. Components of the JNK signaling pathway interact with the JIP1 scaffold protein. JIP1 is located in the neurites of primary hippocampal neurons. However, in response to stress, JIP1 accumulates in the soma together with activated JNK and phosphorylated c-Jun. Disruption of the Jip1 gene in mice by homologous recombination prevented JNK activation caused by exposure to excitotoxic stress and anoxic stress in vivo and in vitro. These data show that the JIP1 scaffold protein is a critical component of a MAP-kinase signal transduction pathway.

An atypical population of NK cells that spontaneously secrete IFN-gamma and IL-4 is present in the intraepithelial lymphoid compartment of the rat.

The intestinal lymphoid compartment of the rat is large and diverse, but the phenotype and functions of its constituent cell populations are not fully characterized. Using new methodology for the isolation and purification of rat intestinal intraepithelial lymphocytes (IELs), we previously identified a population of alphabeta- and gammadelta-TCR- NKR-P1A+ NK cells. These cells were almost completely restricted to the CD4-CD8- IEL population, and unlike peripheral NK cells in the rat, they were CD2-. We now report that rat intraepithelial NK (IENK) and peripheral NK cells are similar in morphology, in their ability to lyse NK-sensitive targets, and in their ability to suppress a one-way mixed lymphocyte culture. In contrast, however, intraepithelial and splenic NK cells differ markedly in two respects. First, IENK cells express high levels of ADP-ribosyltransferase 2 (a marker of regulatory T cells in the rat) and CD25, whereas peripheral NK cells do not. Second, unlike splenic NK cells, a substantial fraction of IENK cells appear to spontaneously secrete IL-4 and/or IFN-gamma. We conclude that the rat IEL compartment harbors a large population of NKR-P1A+CD3- cells that function as NK cells but display an activated phenotype and unusual cytokine profile that clearly distinguish them from splenic NK cells. Their phenotypic and functional characteristics suggest that these distinctive IENK cells may participate in the regulation of mucosal immunity.

Nicotinamide adenine dinucleotide (NAD) and its metabolites inhibit T lymphocyte proliferation: role of cell surface NAD glycohydrolase and pyrophosphatase activities.

The presence of NAD-metabolizing enzymes (e.g., ADP-ribosyltransferase (ART)2) on the surface of immune cells suggests a potential immunomodulatory activity for ecto-NAD or its metabolites at sites of inflammation and cell lysis where extracellular levels of NAD may be high. In vitro, NAD inhibits mitogen-stimulated rat T cell proliferation. To investigate the mechanism of inhibition, the effects of NAD and its metabolites on T cell proliferation were studied using ART2a+ and ART2b+ rat T cells. NAD and ADP-ribose, but not nicotinamide, inhibited proliferation of mitogen-activated T cells independent of ART2 allele-specific expression. Inhibition by P2 purinergic receptor agonists was comparable to that induced by NAD and ADP-ribose; these compounds were more potent than P1 agonists. Analysis of the NAD-metabolizing activity of intact rat T cells demonstrated that ADP-ribose was the predominant metabolite, consistent with the presence of cell surface NAD glycohydrolase (NADase) activities. Treatment of T cells with phosphatidylinositol-specific phospholipase C removed much of the NADase activity, consistent with at least one NADase having a GPI anchor; ART2- T cell subsets contained NADase activity that was not releasable by phosphatidylinositol-specific phospholipase C treatment. Formation of AMP from NAD and ADP-ribose also occurred, a result of cell surface pyrophosphatase activity. Because AMP and its metabolite, adenosine, were less inhibitory to rat T cell proliferation than was NAD or ADP-ribose, pyrophosphatases may serve a regulatory role in modifying the inhibitory effect of ecto-NAD on T cell activation. These data suggest that T cells express multiple NAD and adenine nucleotide-metabolizing activities that together modulate immune function.

Translating data from animal models into methods for preventing human autoimmune diabetes mellitus: caveat emptor and primum non nocere.

Type 1 diabetes in humans is a serious autoimmune disorder of children that is still poorly understood, unpreventable, and irreversible. Study of its animal models, notably the NOD mouse and BB rat, has generated a wealth of information concerning genetics and immunopathogenesis, but that information has still not altered the way in which we treat children with diabetes. In this review we attempt to identify the most promising avenues of continuing research in these models and the most important issues that must be faced by the designers of human therapies based on the animal dataset.

BB rat thymocytes cultured in the presence of islets lose their ability to transfer autoimmune diabetes.

Thymocytes from adult BB rats can adoptively transfer autoimmune diabetes to athymic recipients. It is also known that the development of BB rat T-cells is recapitulated in adult thymus organ cultures (ATOCs). Based on these observations, we tested the hypothesis that cells capable of the adoptive transfer of diabetes would be present in long-term ATOCs but could be rendered nondiabetogenic by co-culture with appropriate antigens. We observed that cells recovered from adult diabetes-resistant BB (BBDR) rat thymi cultured for up to 14 days can adoptively transfer disease to athymic WAG-rnu/rnu rats treated with polyinosinic: polycytidylic acid and a monoclonal antibody to preclude development of ART2a+ regulatory T-cells. Co-culture of adult BBDR thymi in the presence of BBDR thyrocytes had no effect on the ability of recovered cells to induce diabetes in 70-80% of adoptive recipients. In contrast, co-culture in the presence of islets prevented transfer of diabetes, on average, in >90% of recipients. Fresh islets, frozen islets, and islets pretreated with streptozotocin to deplete insulin were equally effective in preventing diabetes, but none prevented insulitis in nondiabetic recipients. Co-culture in the presence of islets was not associated with detectable alterations in phenotype or in the secretion of gamma-interferon or interleukin-4, either in cultures or in cells recovered from adoptive recipients. We conclude that islet antigens involved in the initiation of autoimmune diabetes in BB rats may be absent or deficient in BB rat thymi. Exposure of ATOCs to exogenous islets may lead to deletion or anergy of diabetogenic T-cells or to the positive selection of regulatory T-cells.

Fetal thymi from diabetes-prone but not diabetes-resistant BB/Wor rats fail to generate mature ART2+ T-cells in organ culture.

Diabetes-prone (BBDP) BB rats develop spontaneous autoimmune diabetes mellitus. They are lymphopenic and severely deficient in ART2+ T-cells. Diabetes-resistant BB (BBDR) rats do not develop spontaneous diabetes and have normal numbers of ART2+ T-cells. T-cell lymphopenia in BBDP rats results from hematopoietic stem cell defects leading to abnormal intrathymic T-cell maturation. To study this process, we established rat fetal thymic organ cultures (FTOC). Like mouse FTOC, cultures of BBDR rat thymi yielded approximately 10(5) cells per lobe. The majority of cells were CD8+ART2+ T-cells. In contrast, BBDP rat FTOC yielded 60% fewer cells (approximately 0.3 x 10(5)/lobe), a smaller percentage of CD8+ and TcRalphabeta+ T-cells, and almost no detectable ART2+ T-cells. ART2 mRNA was detectable in BBDR but not BBDP FTOC. In contrast, expression of mRNAs encoding bcl-2 and a panel of cytokines was comparable in BBDP and BBDR FTOC. Addition of anti-ICAM-1 (CD54) antibody reduced T-cell number in BBDR rat FTOC by approximately 70%, but addition of IL-7 or IL-1beta had no effect. The data demonstrate that BBDP thymocytes fail to generate mature ART2+ T-cells in rat FTOC, a system that can now be used to study the mechanism of this process.

Rat xenograft survival in mice treated with donor-specific transfusion and anti-CD154 antibody is enhanced by elimination of host CD4+ cells.

BACKGROUND: Treatment with a donor-specific transfusion (DST) and a brief course of anti-mouse CD154 (anti-CD40-ligand) monoclonal antibody (mAb) prolongs the survival of both allografts and rat xenografts in mice. The mechanism by which allograft survival is prolonged is incompletely understood, but depends in part on the presence of CD4+ cells and the deletion of alloreactive CD8+ T cells. Less is known about the mechanism by which this protocol prolongs xenograft survival. METHODS: We measured rat islet and skin xenograft survival in euthymic and thymectomized mice treated with combinations of DST, anti-CD154 mAb, anti-CD4 mAb, and anti-CD8 mAb. Recipients included C57BL/6, C57BL/6-scid, C57BL/6-CD4null, and C57BL/6-CD8null mice. RESULTS: Pretreatment with a depleting anti-CD4 mAb markedly prolonged the survival of both skin and islet xenografts in mice given DST plus anti-CD154 mAb. Comparable prolongation of xenograft survival was obtained in C57BL/6-CD4null recipients treated with DST and anti-CD154 mAb. In contrast, anti-CD8 mAb did not prolong the survival of either islet or skin xenografts in mice treated with DST and anti-CD154 mAb. Thymectomy did not influence xenograft survival in any treatment group. Adoptive transfer of splenocytes from C57BL/6-CD4null recipients treated with DST and anti-CD154 mAb and bearing long-term skin xenografts revealed the presence of residual xenoreactive cells. CONCLUSIONS: These data suggest that treatment with DST and anti-CD154 mAb induces a state of "functional" transplantation tolerance. They also support the hypothesis that both the induction and maintenance of graft survival based on this protocol depend on different cellular mechanisms in allogeneic and xenogeneic model systems.

Islet cell transplantation tolerance.

Curative islet transplantation for type 1 diabetes currently requires lifelong systemic immunosuppression. Induction of islet transplantation tolerance would be far preferable. We have previously demonstrated that blockade of costimulation by the administration of a donor-specific transfusion in combination with anti-CD154 monoclonal antibody leads to permanent islet and prolonged skin allograft survival in mice. The protocol requires the presence of CD4+ T cells, interferon-gamma, and CTLA4, and involves the deletion of CD8+ alloreactive T cells. Translation of this strategy into clinical practice will, however, require attention to at least two issues. First, we have observed that the presence of viral infection during tolerance interferes with tolerance induction. Second, we have observed that our tolerance induction protocol is ineffective in autoimmune nonobese diabetic mice. We hypothesize that resistance to tolerance induction in nonobese diabetic mice is due to the presence of memory autoreactive cells. To overcome the deleterious effects of viral infection and of primed memory responses, it may be necessary to modify current tolerance induction strategies based on costimulatory blockade. These modifications may require patient isolation, the generation of hematopoietic chimerism, or treatments that target the specific T-cell populations, cytokines, and/or costimulatory factors responsible for resistance. Such modifications may make it possible to extend tolerance induction to the "real world" situation of individuals with type 1 diabetes who are likely to harbor both memory allo-and autoreactive immune cells.

Viral infection abrogates CD8(+) T-cell deletion induced by costimulation blockade.

BACKGROUND: Treatment with a single donor-specific transfusion (DST) plus a brief course of anti-CD154 monoclonal antibody (mAb) prolongs skin allograft survival in mice. It is known that prolongation of allograft survival by this method depends in part on deletion of alloreactive CD8(+) T cells at the time of tolerance induction. Recent data suggest that infection with lymphocytic choriomeningitis virus (LCMV) abrogates the ability of this protocol to prolong graft survival. METHODS: To study the mechanism by which viral infection abrogates allograft survival, we determined (1) the fate of tracer populations of alloreactive transgenic CD8(+) T cells and (2) the duration of skin allograft survival following treatment with DST and anti-CD154 mAb in the presence or absence of LCMV infection. RESULTS: We confirmed that treatment of uninfected mice with DST and anti-CD154 mAb leads to the deletion of alloreactive CD8(+) T cells and is associated with prolongation of skin allograft survival. In contrast, treatment with DST and anti-CD154 mAb in the presence of intercurrent LCMV infection was associated with the failure to delete alloreactive CD8(+) T cells and with the rapid rejection of skin allografts. The number of alloreactive CD8(+) cells actually increased significantly, and the cells acquired an activated phenotype. CONCLUSIONS: Interference with the deletion of alloreactive CD8(+) T cells mediated by DST and anti-CD154 mAb may in part be the mechanism by which viral infection abrogates transplantation tolerance induction.

Allogeneic hematopoietic chimerism in mice treated with sublethal myeloablation and anti-CD154 antibody: absence of graft-versus-host disease, induction of skin allograft tolerance, and prevention of recurrent autoimmunity in islet-allografted NOD/Lt mice.

We describe a tolerance-based stem cell transplantation protocol that combines sublethal radiation with transient blockade of the CD40-CD154 costimulatory pathway using an anti-CD154 antibody. With this protocol, we established hematopoietic chimerism in BALB/c mice transplanted with fully allogeneic C57BL/6 bone marrow. The percentage of donor-origin mononuclear cells in recipients was more than 99%. In addition, all chimeric mice treated with anti-CD154 antibody remained free of graft-versus-host disease (GVHD) and accepted donor-origin but not third-party skin allografts. It was similarly possible to create allogeneic hematopoietic chimerism in NOD/Lt mice with spontaneous autoimmune diabetes. Pancreatic islet allografts transplanted into chimeric NOD/Lt mice were resistant not only to allorejection but also to recurrence of autoimmunity. We conclude that it is possible to establish robust allogeneic hematopoietic chimerism in sublethally irradiated mice without subsequent GVHD by blocking the CD40-CD154 costimulatory pathway using as few as 2 injections of anti-CD154 antibody. We also conclude that chimerism created in this way generates donor-specific allograft tolerance and reverses the predisposition to recurrent autoimmune diabetes in NOD/Lt mice, enabling them to accept curative islet allografts. (Blood. 2000;95:2175-2182)