Deciphering CD30 ligand biology and its role in humoral immunity

Abstract

Ligands and receptors in the tumour necrosis factor (TNF) and tumour necrosis factor receptor (TNFR) superfamilies have been the subject of extensive investigation over the past 10–15 years. For certain TNFR family members, such as Fas and CD40, some of the consequences of receptor ligation were predicted before the identification and cloning of their corresponding ligands through in vitro functional studies using agonistic receptor-specific antibodies. For other members of the TNFR family, including CD30, cross-linking the receptor with specific antibodies failed to yield many clues about the functional significance of the relevant ligand–receptor interactions. In many instances, the subsequent availability of TNF family ligands in the form of recombinant protein facilitated the determination of biological consequences of interactions with their relevant receptor in both in vitro and in vivo settings. In the case of CD30 ligand (CD30L; CD153), definition of its biological role remained frustratingly elusive. Early functional studies using CD30L⁺ cells or agonistic CD30-specific antibodies logically focused attention on cell types that had been shown to express CD30, namely certain lymphoid malignancies and subsets of activated T cells. However, it was not immediately clear how the reported activities from these in vitro studies relate to the biological activity of CD30L in the more complex whole animal setting. Recently, results from in vivo models involving CD30 or CD30L gene disruption, CD30L overexpression, or pharmacological blockade of CD30/CD30L interactions have begun to provide clues about the role played by CD30L in immunological processes. In this review we consider the reported biology of CD30L and focus on results from several recent studies that point to an important role for CD30/CD30L interactions in humoral immune responses.

Introduction

CD30 ligand (CD30L) and CD30 are interacting cell-surface glycoproteins that are members of the tumour necrosis factor (TNF) and tumour necrosis factor receptor (TNFR) superfamilies, respectively.^1,2 Molecular interactions among members of these families have important immune regulatory functions and have been reviewed extensively.^3–5 Although CD30L and CD30 were identified and cloned over a decade ago, it has been challenging to define the consequences of CD30/CD30L interactions at both the molecular and cellular levels. In addition to its well-documented association with several lymphoid malignancies, CD30 has been described as a marker of memory T cells but can also be expressed by activated B cells. Early human in vitro studies suggested that T-cell CD30 is specifically associated with T helper 2 (Th2)-type T cells and is elevated in ‘Th2-type’ diseases.^6–9 However, subsequent studies demonstrated that activation-induced CD30 can be expressed on Th1 and Th0, as well as Th2, clones.^10,11 Furthermore, a number of studies have confirmed elevated CD30 expression in diseases with ‘Th1’,^12–14 as well as ‘Th2’, characteristics.^15,16

The observation that a viral form of CD30 is one of a number of ‘viroreceptors’ encoded by several orthopoxviruses provides additional evidence for an immune regulatory role for CD30/CD30L interactions.¹⁷ The viroreceptors are related to cellular receptors and are thought to act as immunomodulators by competing for ligands that promote antiviral immunity or inflammatory responses. The role of viral CD30 as a determinant of host range, virulence or pathogenicity has not yet been reported.¹⁸

Various in vitro studies implicate CD30L in the regulation of humoral immune responses but the conclusions regarding the mechanism(s) and outcomes of such regulation are not consistent among the studies.^19–23 Recent studies employing CD30 gene disruption^24,25 and CD30L transgenic [C.R.W., manuscript in preparation] animal models provide evidence that points to a regulatory role for the CD30/CD30L pathway in secondary antibody responses. Several recent reviews focus on CD30/CD30L signalling²⁶ and on the potential role of this pathway in the biology of T cells and CD30⁺ tumours.^27–30 In this review we focus primarily on the role of CD30/CD30L interactions in the regulation of humoral immune responses.

Expression of CD30L and CD30

The expression of CD30L and CD30 is restricted to cells of the immune system and is tightly regulated. In the mouse, CD30L is detected on in vitro T-cell receptor (TCR)-activated mouse T cells but not on resting or in vitro-activated mouse B cells or NK cells (ref. ³¹ and M.K.K., unpublished observations). CD30L appears to be expressed constitutively on at least some types of dendritic cells (DC) or DC-like cells. For example, CD30L is expressed constitutively on a mouse DC line with a mature phenotype [CD11c⁺ CD40⁺ CD86^hi CCR7⁺ major histocompatibility complex (MHC) class II^hi] but was not detected on immature (CD11c⁺ CD40⁻ CD86^lo CCR7⁻ MHC class II^lo) DC populations (David Fitzpatrick, Amgen Inc, personal communication). CD30L is also expressed constitutively on a unique population of mouse splenic CD4⁺ CD3⁻ accessory cells that are found in close contact with primed T cells in B-cell follicles.³² It is proposed that these accessory cells, which also express OX40L, play important roles in organization of lymphoid tissue as well as in promoting the survival of T cells that provide help to B cells during affinity maturation and memory B-cell responses.²⁵

In humans, CD30L is expressed at high levels on activated T cells. There is conflicting evidence with regard to cell surface expression of CD30L on primary human B cells. An initial report of constitutive CD30L expression on peripheral blood B cells³³ has not been substantiated. However, it appears that distinct subsets of B cells obtained from human lymphoid tissue can express CD30L in response to ex vivo activation or can express CD30L constitutively.²² Studies from Klein et al.³⁴ and Feldhahn et al.³⁵ demonstrated that there are dramatic differences in the relative gene expression of CD30L among human B-cell subsets. Although the findings with naive B-cell populations were conflicting (relatively low versus high expression), both groups observed higher relative gene expression of CD30L within germinal centre B-cell populations compared to memory B-cell populations. The findings from both groups are compatible with the observation that CD30L can be expressed constitutively on B-cell lines with a germinal centre-associated phenotype, such as established Burkitt's lymphoma lines (R.J.A., unpublished observations) and the Burkitt's lymphoma-derived line CL-01.²¹ Thus the expression of CD30L on B cells is likely to be both stage and context specific.

We believe that the expression of CD30L on normal cells is more restricted than suggested by review of the literature. CD30L expression has been reported on a variety of human cell types, including medullary thymic epithelial cells, mast cells, eosinophils, neutrophils and peripheral blood B cells. Not all studies were well controlled to prevent non-specific binding of antibodies and few, if any, demonstrated reversal of staining by competition with soluble CD30 or CD30L constructs. Using flow cytometry with directly labelled anti-human CD30L monoclonal antibody (mAb), we have been unable to confirm cell surface expression of CD30L on freshly isolated blood B cells, neutrophils or eosinophils, or on primary mast cells derived from CD34⁺ blood progenitors or two human mast cell lines.

The cell surface expression of CD30 appears to be restricted primarily to subpopulations of T and B cells. In normal human tissues, CD30 expression is observed only on activated blasts in parafollicular areas of lymphoid tissues and in the thymic medulla, mainly around the Hassall's corpuscles.³⁶ On in vitro activated human and mouse T and B cells, CD30 is a relatively late activation-induced antigen, with maximal expression observed 48–72 hr poststimulation.^37,38 In mice, CD30 is expressed on splenic T cells after activation in vitro with immobilized CD3 mAb³⁸ and on freshly isolated CD4⁺ CD25⁺ or CD4⁻ CD8⁻ regulatory T cells.^39,40 The ability of mouse T cells to express CD30 in response to in vitro stimulation with CD3 mAb is enhanced by CD28 engagement and interleukin-4 (IL-4), but is down-regulated by interferon-γ^41,42. In peripheral blood of normal human subjects, CD30 is present on only a small fraction of cells. Among CD4⁺ T cells, CD30 expression is restricted to those that coexpress CD45RO.³⁷ CD30 expression is inducible in vitro on a fraction of CD45RO⁺ human T cells in response to a variety of T-cell activators.³⁷ CD30 expression has also been reported on a subset of freshly isolated CD8⁺ blood T cells⁴³ and on CD8⁺ T cells in response to repeated ex vivo stimulation.⁴⁴ CD30 is not expressed on resting B cells but can be induced on both mouse and human B cells by CD40L (ref. ²¹; R.J.A., unpublished observations). CD40L induction of CD30 on B cells is further enhanced by IL-4. CD30 is also expressed on the human mast cell lines HMC-1 and LAD2, but is absent from SCF-dependent mast cells derived from CD34⁺ peripheral blood progenitors (R.J.A., unpublished observations).

CD30⁺ cells are the hallmark of anaplastic large cell lymphoma, Hodgkin's disease and a number of other lymphoid malignancies.^27,36 CD30⁺ cells are also present at inflammatory sites in several human diseases, including atopic dermatitis,^45,46 rheumatoid arthritis,¹³ chronic graft versus host disease (GVHD)⁹ and systemic sclerosis.^9,47

Soluble CD30

A soluble form of CD30 (sCD30) can be detected in the serum of most normal individuals and is elevated in serum and/or body fluids of patients with CD30⁺ haematopoietic malignancies, certain viral infections, and a variety of inflammatory conditions (reviewed in refs ^16,48). In Hodgkin's disease and anaplastic large cell lymphoma, serum sCD30 concentrations can be used as a marker of tumour burden and appear to have prognostic value. ^49–54 Similarly, pretransplant serum sCD30 concentrations are predictive of renal graft survival and post-transplant sCD30 is being evaluated as an early predictor of impending graft rejection.⁵⁵ Elevated concentrations of circulating sCD30 have been reported to correlate with disease activity in patients with systemic lupus erythematosus,⁵⁶ Wegener's granulomatosis,⁵⁷ rheumatoid arthritis,⁵⁸ Grave's disease,⁵⁹ Hashimoto's thyroiditis⁵⁹ and human immunodeficiency virus-1 infection.^60,61 In contrast, although elevated sCD30 concentrations are also observed in adult and juvenile atopic dermatitis,^45,62–66 atopic asthma⁶⁷ and ulcerative colitis,⁶⁸ there is no apparent association with disease severity in these settings.

Soluble CD30 is shed from the surface of CD30⁺ lymphoma cells in vitro in response to specific ligation of membrane-bound CD30 or by other stimuli such as PMA.^69–71 Human CD30 is cleaved from the cell surface via the action of the cell surface metalloproteinase TNF-α-converting enzyme (TACE).⁷⁰ The mechanisms that lead to such release have not been the subject of intense exploration. It is possible that in vivo, sCD30 is released from the cell surface as a consequence of appropriate activation signals and/or in response to interaction with CD30L⁺ cells. Evidence for CD30L-dependent release of human sCD30 has been demonstrated in vitro.⁷¹

The exact cellular source of sCD30 in settings other than malignancies is not known, although in vitro experiments suggest that CD30 can be shed from the surface of activated lymphocytes. In previous studies investigators have assumed that activated T cells (or specifically Th2 cells) are the probable source of sCD30 in the context of an immune response, but it is also likely that activated B cells are a source of sCD30 in vivo.

Potential roles for CD30/CD30L interactions in immune responses

Implications from in vitro studies

Early in vitro studies led to the classification of CD30 as a T-cell ‘costimulatory receptor’ based on observations that immobilized CD30-specific antibodies or CD30L-transfected cells enhance the proliferation of human T cells in response to suboptimal stimulation via the TCR.² The physiological relevance of these early findings is not clear, because anti-human or anti-mouse CD30L mAb do not appear to block antigen-presenting cell-dependent T-cell proliferation and/or function in a variety of in vitro systems. As CD30 is expressed on T cells rather late after in vitro activation, it is possible that CD30/CD30L interactions occurring relatively late after antigen encounter promote T-cell survival and/or establishment of strong memory responses.³

Additional reports using recombinant CD30L or CD30-specific antibodies described a variety of (often quite conflicting) consequences of cross-linking CD30 on various CD30⁺ cell lines. For example, triggering cells via CD30 has been shown to induce activation and proliferation, or growth arrest and apoptosis. An elegant and comprehensive review of these in vitro findings and the potential CD30-mediated signalling pathways to account for the observations is available elsewhere.²⁶ To complicate matters further, it was reported that cross-linking of CD30L transduces a signal to the ligand-bearing cell.⁷²‘Reverse signalling’, wherein a ligand also functions as a receptor, has also been reported for other TNF family members in in vitro settings; however, the molecular mechanisms to account for reverse signalling have not been elucidated.⁷³

Despite the fact that CD30 is an activation marker expressed on both T and B cells, the potential roles of CD30/CD30L interactions in the direct control of B-cell proliferation and differentiation have not been studied extensively. Results from in vitro studies implicate CD30L in the regulation of humoral immune responses, but the conclusions regarding the mechanism(s) and outcomes of such regulation are not consistent among the studies. Most of the in vitro studies used relatively complex assays, which make it difficult for others to replicate. Furthermore, interpretation of the results is complicated by the fact that T cells and certain subsets of B cells can both express CD30L and CD30 in response to appropriate stimulation. The findings from these studies are summarized below and are outlined in Fig. 1.

Figure 1.

Potential outcomes of CD30L/CD30 interactions based on published results from in vitro studies.

Using a mouse system, Shanebeck et al. demonstrated that stimulation of mouse splenic B cells by CV-1 cells transfected with CD30L in the presence of cytokines (IL-2, IL-4 and/or IL-5) induced the B cells to proliferate, differentiate and produce antibodies [immunoglobulin G1 (IgG1), IgA, IgG3 and IgE].¹⁹ In addition, activated T-cell clones from CD40L-deficient mice also promoted immunoglobulin production, which was blocked partially in the presence of a CD30.Fc fusion protein. A general model based on this study predicts that CD30L⁺ cells interact with CD30⁺ B cells to promote B-cell proliferation, differentiation and/or antibody production.

In contrast, results from human in vitro studies suggest that CD30/CD30L interactions inhibit class switching and immunoglobulin production. Models based on the human studies can be subdivided into two general categories: one in which CD30⁺ B cells receive an inhibitory signal from CD30L⁺ cells^20,21 and another in which CD30L⁺ B cells receive an inhibitory signal from CD30⁺ cells (via ‘reverse signalling’).^22,23 To account for these seemingly conflicting findings, it has been proposed that non-antigen-selected B cells receive inhibitory signals from CD30L⁺ cells, whereas antigen-experienced B cells receive inhibitory signals from CD30⁺ cells.²²

In studies in which reverse signalling has been proposed,^22,23 activated T cells from normal individuals or from patients with chronic lymphocytic leukaemia inhibited immunoglobulin class switching in activated tonsil B cells or in the B-cell line CL-01. This inhibitory effect could be blocked by the addition of a CD30-specific antibody. The authors concluded that the observed effects were the result of interactions between CD30⁺ T cells and CD30L⁺ B cells; however, it is important to note that interpretation of these experiments can be complicated by the fact that activated T cells also express CD30L and activated B cells can express CD30.

Implications from in vivo studies

As noted above, in vitro studies have identified a plethora of effects attributed to CD30/CD30L interactions, yet such studies have not shed much light on the role of such interactions in vivo. Studies in gene-deficient or transgenic mice or in normal mice treated with antibodies specific for CD30 or CD30L are also being used to the explore the role of CD30/CD30L interactions in immune responses.

Studies in CD30L^−/− and CD30^−/− mice

The initial report on CD30^−/− mice described a partial defect in negative selection of thymocytes.⁷⁴ However, others propose that there is little if any role for CD30 in negative selection.^75,76 With regard to other defects, Amakawa et al.⁷⁴ reported that CD30^−/− mice have no obvious defects in peripheral T- or B-cell homeostasis, and they mount effective primary and secondary cytotoxic T-lymphocyte responses and neutralizing antibody responses (IgM and IgG) to vesicular stomatitis virus. Basal serum concentrations of IgM, IgG1, IgG2a, IgG2b and IgG3 and in vitro T-cell proliferative responses to CD3 mAb or mitogens were comparable to controls. The authors concluded that CD30 is not required for production or maintenance of CD8⁺ memory T cells or the maturation and class switching of normal B cells.

Flórido et al.⁷⁷ reported that CD30 contributes to immune responses to Mycobacterium avium. Throughout infection, CD30^−/− mice had reduced CD4⁺ T-cell expansion and decreased interferon-γ production by splenocytes stimulated in vitro with myobacterial proteins. Granulomas formed normally in the CD30^−/− mice, but at late times after infection (> 3 months) the granulomas were reduced in size relative to those of the controls. The authors note that the immune deficiencies associated with the lack of CD30 were small in magnitude and appear to play a marginal role compared to other elements of the immune system in protective immunity to M. avium.

An early study using CD30^−/− mice as donors in an adoptive transfer model of diabetes led to the concept that CD30 was an important ‘biological brake’ for controlling CD8⁺ T-cell-mediated responses in the periphery.^78,79 The investigators later reported that their original observations were caused by a genetic disparity between the host and recipient mice rather than the absence of CD30 on the donor cells.⁸⁰ However, the original findings are regularly cited in manuscripts and reviews as evidence of a central role for CD30 in maintaining peripheral tolerance by controlling the expansion of autoreactive CD8⁺ T cells.

CD30^−/− cells have also been used as a source of donor cells in lethal, T-cell-mediated models of GVHD.⁸¹ There was no apparent difference in the outcome of GVHD when CD30^−/− or C57BL/6 control cells were used as donors to induce CD8⁺ T-cell-dependent GVHD in sublethally irradiated, MHC class I-only disparate bm1 recipients, or to induce CD4⁺ plus CD8⁺ T-cell-dependent GVHD in lethally irradiated, fully allogeneic B10.BR recipients. In contrast, the absence of CD30 on donor cells resulted in a reduced mortality in CD4⁺ T-cell-dependent GVHD induced in sublethally irradiated, MHC class II-only disparate bm12 recipients. A survival advantage was also observed in lethally irradiated, CD30L^−/− recipients injected with fully allogeneic, BALB/c CD4⁺ T cells.⁸¹ In both settings, the survival advantage of the recipients was overcome by increasing the number of CD4⁺ donor T cells.

A recent study,²⁴ as well as our own studies (see below), revealed that CD30^−/− mice have defects in their secondary humoral immune responses. Gaspal et al.²⁴ demonstrated that CD30^−/− mice fail to sustain follicular germinal centre responses and secondary antibody responses are reduced dramatically compared to controls. The defect in secondary humoral responses in these mice has been attributed to a deficiency in CD4⁺ T-cell memory that stems from the inability of CD30^−/− T cells to receive adequate survival signals from OX40L⁺ CD30L⁺ accessory cells found in B-cell follicles. We propose that the deficiency in the humoral response in these mice results, at least in part, from the inability of activated B cells from CD30^−/− mice to respond to signals from CD30L⁺ cells (activated T cells and/or the accessory cells referred to above).

The generation of CD30L^−/− mice has been reported recently^81,82 but a comprehensive analysis of the immune phenotype associated with a deficiency in CD30L has not been described. The CD30L^−/− mice have been used as recipients in adoptive transfer models to determine whether CD30/CD30L interactions are required for CD4⁺ T-cell-mediated, lethal GVHD⁸¹ (see above) or for expansion and maintenance of CD8⁺ T cells in vivo.⁸² In the latter study, a reduction in antigen-induced expansion of TCR transgenic CD8⁺ T cells was observed in CD30L^−/− mice relative to wild-type mice. More recently, Nishimura et al. reported that CD30L^−/− mice backcrossed onto a BALB/c background have normal generation of effector CD8⁺ T cells in response to infection with Listeria monocytogenes, but have a defect relative to BALB/c mice in their ability to maintain long-lived, Listeria-specific memory CD8⁺ T cells.⁸³ The effects of CD30L deficiency on humoral immune responses have not been described.

Although comparative studies using gene-deficient mice are useful for identifying potential contributions of CD30/CD30L interactions to T- or B-cell-mediated responses, additional evidence is needed to clearly establish the role of these interactions in the immune response. Unless littermate control mice are available for every system under study, attributing differences in outcomes of adoptive transfer or disease models, such as those mentioned above, to a single genetic difference between gene-deficient and control mice can be difficult. The possibility remains that additional genetic disparities may exist between the gene-deficient mice and the control mice and/or that variations in environmental factors influence the immune response of a particular strain of mouse (such could occur when gene-deficient mice bred within an investigator's facility are compared to control mice ordered from a vendor).

Pharmacological blockade of CD30/CD30L interactions in vivo

The in vivo relevance of CD30/CD30L interactions can also be explored in mice using antibodies that block interactions between CD30- and CD30L-bearing cells. Antibodies to TNFR family members are potentially agonistic in vivo, thus an antibody specific for CD30L rather than CD30 is perhaps the better reagent for such studies. However, it is important to keep in mind that an antibody directed against CD30L could modulate immune responses in vivo by blocking interactions between CD30⁺ and CD30L⁺ cells, by depleting CD30L⁺ cells, or perhaps even via agonistic activity on CD30L⁺ cells (‘reverse signalling’).

The effects of a commercially available rat anti-mouse CD30L mAb (clone RM153;³¹ IgG2b isotype) have been reported in a number of settings. Anti-CD30L RM153 prevents or delays the spontaneous development of diabetes in non-obese diabetic (NOD) mice when treatment is initiated in young mice (2–4 weeks of age). In addition, the mAb inhibits the ability of NOD splenocytes or islet-specific cell lines to transfer diabetes into NOD-SCID mice.⁸⁴ In an allograft model designed to assess the regulatory function of CD4⁺ CD25⁺ T regulatory cells, anti-CD30L RM153 treatment blocked the ability of antigen-induced, adoptively transferred T regulatory cells to delay the skin allograft rejection mediated by the adoptive transfer of memory CD8⁺ T cells into splenectomized alymphoplastic (aly/aly) mice. In an infectious disease setting, C57BL/6 mice infected with M. avium and treated weekly with anti-CD30L RM153 for up to 3 months had similar bacterial burdens in tissues compared to control mice at approximately 4 weeks postinfection, but had increased bacterial burdens at 12 weeks postinfection.⁷⁷ Finally, treatment with anti-CD30L RM153 prolongs the survival of mice in a model of CD4⁺ T-cell-mediated lethal GVHD induced by injection of C57BL/6 cells into MHC class II-only disparate bm12 recipients.⁸¹

The authors of all these studies propose that the RM153 mAb acts in vivo to block interactions between CD30⁺ and CD30L⁺ cells, rather than by depleting CD30L⁺ cells. The ability of an antibody to deplete CD30L⁺ cells is a difficult question to address in normal mice because CD30L is expressed constitutively only on small subpopulations of cells (such as the accessory cells described by Lane et al.²⁵) or is expressed transiently on activated T cells. Our unpublished studies, however, indicate that a single injection of 500 μg of this particular mAb can effectively deplete CD30L⁺ cells in vivo within 3–4 days of administration. We tested the ability of anti-CD30L RM153 (rat IgG2b), as well as one of our own mAb, anti-CD30L M15 (rat IgG2a), to deplete cells in CD30L transgenic mice. In these mice, the expression of mouse CD30L expression is driven by the human CD2 enhancer and CD30L is essentially a ‘pan T-cell’ marker that is expressed at relatively high levels on the majority of T cells (C.R.W., manuscript in preparation). The anti-CD30L RM153 mAb effectively depleted T cells in the peripheral blood of the transgenic mice, whereas our anti-CD30L M15 mAb appeared to be non-depleting, or poorly depleting at best.

Thus, the commercially available anti-CD30L RM153 mAb is capable of depleting CD30L⁺ cells in vivo, and it is possible that the efficacy of this particular mAb in some in vivo settings results from depletion of CD30L⁺ cells rather than, or in addition to, blocking interactions between CD30⁺ and CD30L⁺ cells.

CD30/CD30L interactions promote secondary humoral immune responses in vivo

A great deal of emphasis has been devoted to determining the role of CD30 in regulating T-cell responses, perhaps because CD30 has been considered a ‘T-cell costimulatory receptor’ and/or ‘biological brake’ for T-cell responses. In addition, no obvious defect in humoral responses was noted in early studies in CD30^−/− mice. In contrast, aside from the in vitro studies summarized earlier in this review, relatively little emphasis has been placed on determining whether CD30/CD30L interactions influence B-cell responses.

The in vitro studies^19–23 discussed above and summarized in Fig. 1 suggest that CD30/CD30L interactions can have a direct positive¹⁹ or negative^20–23 effect on mouse and human B-cell proliferation and/or differentiation. To test the various models of CD30/CD30L-mediated regulation of B-cell responses, we examined humoral responses in CD30^−/− mice, in normal mice treated with an antibody to mouse CD30L, and in transgenic mice that overexpress CD30L on T cells.

Our own studies have shown that CD30^−/− mice have no alterations in basal serum concentrations of IgM or IgG3, slight reductions in serum IgG1, and dramatic reductions in basal serum concentrations of IgG2c (also known as IgG2a^b), IgA, and IgE relative to controls (M.K.K., manuscript in preparation). These aberrations become more apparent as the mice age, which might explain why the defect was not noted in the early studies with CD30^−/− mice. These mice mount normal trinitrophenyl (TNP)-specific IgM and IgG3 responses to a prototype T-independent type 2 antigen (TNP-Ficoll), but their TNP-specific IgE response to a prototype T-dependent antigen (TNP-KLH) is about one-third that of the controls.

Our observations are consistent with those of Gaspal et al.²⁴ who demonstrated that CD30^−/− mice have an impaired capacity to sustain follicular germinal centre responses and show a reduced ability to mount an antibody response to recall antigen. These authors propose that the defect in humoral responses is secondary to a deficiency in CD4⁺ T-cell memory. In this model, T-cell memory is deficient because the T cells do not receive adequate survival signals from OX40L⁺ CD30L⁺ accessory cells found in B-cell follicles.^24,25 However, as noted above, activated B cells express CD30 in response to stimulation with CD40L and this expression can be enhanced by IL-4. In addition there is evidence that B cells can express CD30L under certain conditions or at certain stages of development. Thus B cells could be active participants in the CD30/CD30L cellular interactions that contribute to strong secondary humoral immune responses.

We have seen a very similar pattern of humoral responses in normal mice that are treated with the ‘non-depleting’ or ‘poorly depleting’ anti-CD30L M15 mAb described above. Such treatment inhibits, but does not abolish, the class-switched antibody response to T-dependent antigens but has no effect on primary antibody responses. Specifically, mice treated with anti-CD30L M15 mounted normal antigen-specific IgM and IgG3 responses to TNP-Ficoll and a relatively normal IgG1 response to a TNP-KLH, but their TNP-specific IgE response to the latter antigen was one-third that of the controls (M.K.K., manuscript in preparation).

In preliminary studies using anti-CD30L M15 in mouse models of disease, we examined the ability of this antibody to inhibit disease-associated increases in total serum IgE. Interestingly, the outcome of such treatment appears to be dependent upon the model under investigation, with observed effects ranging from no effect to dramatic inhibition of the IgE response. In at least one setting, the ability of CD30L to promote IgE responses appears to be independent of CD40L (M.K.K., manuscript in preparation). It is possible that CD30L works in concert with CD40L to promote secondary humoral responses in some settings, substitutes for CD40L in other settings, but has no role in others.

An opposite pattern of humoral immune responses was observed in CD30L transgenic mice (C.R.W., manuscript in preparation). We recently generated two CD30L transgenic lines on a C57BL/6 background, using a mouse CD30L transgene construct containing the human CD2 enhancer to direct the expression of CD30L to T cells. CD30L is expressed constitutively at relatively high levels on the majority of T cells in each line, and the two lines have a comparable phenotype.

CD30L transgenic mice show increased numbers and activity of splenic germinal centres and elevated basal serum concentrations of IgG2b, IgG2c, IgE, and IgA but not IgM, IgG3, or IgG1. The mice have an exaggerated response to TNP-KLH (IgE > IgG1) and an exaggerated IgE response in an ovalbumin-induced model of pulmonary inflammation. In general, the results from the CD30L transgenic mice are the opposite of what we have observed in CD30^−/− mice.

Figure 2 summarizes the predicted outcomes (based on the in vitro studies) as well as the observed outcomes from the in vivo studies described above. Taken together, our unpublished results and those of Gaspal et al.²⁴ support a role for CD30/CD30L interactions in promoting, rather than inhibiting, secondary humoral immune responses. Although none of these in vivo observations provide evidence that CD30L acts directly on CD30⁺ B cells to promote secondary humoral responses, taken together the results from these studies are consistent only with this scenario.¹⁹ In contrast, the overall observations do not support the proposals that ligation of CD30^20,21 on B cells or ‘reverse signalling’^22,23 via CD30L expressed on B cells has a negative effect on antibody production.

Figure 2.

Are hypotheses based on published in vitro findings predictive of CD30L activity in vivo?

Concluding remarks

Considerable progress has been made recently toward better understanding of the biological consequences of CD30L engagement with its receptor. Needless to say, immune responses occur as a result of complex interactions between many cell types in varied locations. While examination of events at the molecular and biochemical level is important for our understanding of individual ligand–receptor interactions, a full understanding of the biological significance of these events will only come from interrogation in an in vivo‘whole body’ context. Further characterization of CD30^−/−, CD30L^−/− and CD30L transgenic mice, together with the use of antagonistic CD30L-specific antibodies in vivo, will yield important information about the physiological and pathological aspects of the involvement of CD30L in normal and disregulated immune responses.