SLE nephritis – learning from murine models

Introduction

Nephritis is a challenging problem affecting between 30–60% of SLE patients^{1, 2}. Outcomes after broad spectrum immunosuppression remain unsatisfactory with complete remission rates of 50% at best³, relapse rates of up to 30% over a two year period⁴ and unacceptable toxicities. Despite the implementation of maintenance regimens^5–7 the incidence of end stage renal disease (ESRD) in SLE patients is increasing^{8, 9} especially in non-Caucasians. Barriers to performing informative clinical trials of new drugs for SLE nephritis remain formidable. Disease heterogeneity, the confounding effects of other immunosuppressive medications and concurrent medical problems contribute to these difficulties.

The last decade has witnessed an exciting increase in our understanding of the immunopathogenesis of autoimmunity. Despite many new discoveries, the path to a new approved therapy for SLE is strewn with a growing list of failures, even of drugs with efficacy in other diseases. Because some disease mechanisms are shared between humans and mice, the study of diverse mouse models has enabled acquisition of pre-clinical data that support clinical trials of novel therapies (Table 1)¹⁰. The major disadvantages of relying on murine models are differences in physiology between mouse and man and the extensive genetic heterogeneity in an outbred human population that complicates predictions of treatment efficacy. Nevertheless, mechanistic studies of therapeutic interventions in murine SLE may help explain the difficulties in identifying new drugs for human SLE.

Pathogenesis of SLE nephritis

SLE nephritis is initiated by the glomerular deposition of immune complexes (IC) that trigger a cascade of inflammatory events including activation of Fc receptors¹¹ and complement¹² recruitment of inflammatory cells and eventual fibrosis. IC also activate resident renal cells through Toll-like receptors (TLRs) to produce inflammatory mediators¹³. Secondary lymphoid tissue has been reported in the renal parenchyma in some patients^{14, 15}. Interstitial nephritis occurs in one murine model in the complete absence of circulating immunoglobulins¹⁶; similarly, pauci-immune nephritis has been observed in humans¹⁷. Renal microvascular damage and thromboses also occur, especially in patients with anti-phospholipid antibodies¹⁸.

Several renal cell types including endothelial cells, podocytes, interstitial cells and renal dendritic cells are the focus of new studies in nephritis (Figure 1)¹⁹. Renal migration of inflammatory cells requires endothelial cell activation²⁰. Hypoxia, due to loss of glomerular and peritubular capillaries contributes to cell stress and cell death and induces molecules that further activate innate immune receptors²¹. There is currently an increased interest in molecules that protect the endothelium from hypoxia, such as oxygen sensors that regulate endothelial leak and microthrombus formation²² and receptors that regulate endothelial adhesiveness. Patients with active lupus have increased levels of angiopoietin-2, an antagonist of Tie2, a receptor that maintains endothelial integrity and prevents leukocyte recruitment to the kidney; this may be a more specific marker of renal involvement than soluble VCAM-1²³. Signals delivered via the vagus nerve to nicotinic acetylcholine receptors (nAChRs) may exert a renoprotective effect by decreasing inflammatory cytokines and preserving endothelial integrity²⁴. All these pathways are potential therapeutic targets for SLE nephritis. In addition, circulating endothelial cells²⁵ and endothelial protein C receptor^{26, 27} may be biomarkers of endothelial dysfunction and damage.

The role of podocytes in renal inflammation is increasingly recognized²⁸ and their loss in chronic disease causes glomerulosclerosis. Podocytes form a critical barrier to proteinuria and protect the endothelium by producing both glomerular basement membrane and VEGF²⁹. Activation of angiotensin receptors on podocytes induces the release of inflammatory mediators and cell death³⁰; blockade of angiotensin receptors protects podocytes and may be beneficial for SLE nephritis³¹.

Interstitial cell infiltration, tubular atrophy that follows vascular compromise, activation of smooth muscle myofibroblast cells and interstitial fibrosis are predictors of poorer prognosis and/or progression to ESRD in SLE nephritis^{32, 33}. Activation of intrinsic renal dendritic cells occurs in both murine and human SLE^{20, 34} and is associated with expression of inflammatory cytokines, proteolytic activity, increased caspase 1 activity^{20, 35} and upregulation of cell surface molecules including CD11b²⁰. Genetic differences involving both immune and non-immune genes may influence the long term outcome of renal inflammation although there are as yet no rigorous studies in SLE nephritis; these studies are challenging to perform even in the setting of renal transplant in which many more patients are available³⁹. Recent identification of polymorphisms of genes that enhance or protect from SLE nephritis such as ITGAM (the α chain of CD11b)³⁶, IL-18³⁷ and kallikreins³⁸ suggest new avenues for investigation. Further studies of the mechanisms involved in intrinsic renal cell activation, interstitial renal inflammation and recruitment of fibrosis pathways should lead to new strategies for renal protection from the consequences of antibody deposition.

New immunologic therapies for SLE

Although the trigger initiating SLE in a genetically susceptible host is not known, initiation of murine lupus can occur in a variety of ways. As disease progresses interactions between innate and acquired immune activation pathways create amplification loops and interacting networks that become more difficult to regulate (Figure 2).

T cell directed therapies

CTLA4Ig, an antagonist of B7-CD28 costimulation is beneficial in murine SLE, especially when given before class switched and somatically mutated autoantibodies arise^{63, 64}. The follicular T helper cells (T_FH) that drive B cell help in germinal centers are completely dependent on CD28 costimulation⁶⁵. However TLR mediated signals together with combinations of cytokines including BAFF, IL-21 and IL-17 can substitute for B7-CD28 costimulation and drive reactivation of memory B cells^{55, 66}. Thus administration of CTLA4Ig may be futile in active lupus or chronic disease with persistent memory T cells and long-lived B cells. This might account for the failure of monthly abatacept (CTLA4Ig), given with tapering doses of steroids, to alter autoantibody titers or rates of flare at one year in patients with non-renal flares. These studies confirm the challenges of using CTLA4Ig alone in established disease. A study of abatacept in lupus nephritis is in progress. Remission of active SLE nephritis can be achieved in mice using a combination of cyclophosphamide and CTLA4Ig. Activated T and B cells are depleted by this regimen, with reversal of the activation phenotype of intrinsic renal cells^{20, 64}. A clinical trial of this combination in humans is also currently in progress.

Follicular T helper cells (T_FH) might be targeted with IL-17, IL-21, CD40 and ICOS antagonists, some of which are in early development in human clinical trials. While IL-21 is required for disease onset in some murine models⁶⁷ it is dispensable for others, even when T_FH are involved⁶⁸ and its inhibition has only modest effects on late disease. Anti-CD40L is effective at preventing murine SLE but has less value for established nephritis⁶⁹. Thus, while T_FH are required for disease initiation either they are not sufficient for disease perpetuation or their inhibition in late disease requires blocking more than one molecule. It remains to be determined whether targeting of T_FH is a viable therapeutic option.

The role of regulatory T cells (T_reg) in SLE nephritis is still controversial. Initial studies in spontaneous murine models have shown expansion of regulatory T cell subsets in the periphery and in the kidney over time⁷⁰ and adoptive transfer of large numbers of CD4+/CD25+ T_reg in the NZB/W model had only a very modest effect on disease onset and survival⁷¹. This may be because T_reg can be inactivated or convert to TH₁₇ cells in an inflammatory environment⁷². In contrast, induction of T_reg by non-depleting anti-CD3 antibodies had very robust effects in the SNF₁ model; in this case suppression was mediated by inducible CD4+/CD25−/LAP+/Foxp3- T_reg ⁷³. Intriguing new data suggests that partial inhibition of the mTOR pathway may induce T_reg ⁷⁴ pointing to an additional mechanism for the efficacy of mTOR inhibitors in murine SLE models. Clearly much remains to be learned about the potential therapeutic role of different types of T_reg in SLE.

Therapies directed at the innate immune system

Many SLE patients manifest a Type I interferon (IFN) signature in the peripheral blood, induced by exposure of PBMCs to IFNα and/or immune complexes⁷⁵. Type I IFNs can induce either anti-DNA or anti-Sm/RNP antibodies and accelerate disease in murine models by activating antigen presenting cells, enhancing expression of TLRs and BAFF receptors, enhancing germinal center formation and inducing release of BAFF and IL-6^{75, 76}. IFNα inhibitors can inhibit IFNα accelerated disease⁷⁷ but it is still unknown if IFNs drive disease after the onset of inflammation. Clinical trials of anti-IFNα antibodies in SLE are underway. In mice, Type I IFNs confer resistance to both CTLA4Ig and BAFF inhibitors (unpublished data). This has two important implications. First high serum levels of BAFF or of T cell derived cytokines do not necessarily indicate sensitivity to inhibition of those cytokines. Second it may be necessary to randomize patients based on their interferon signature in order to analyze the effects of other immune modulators.

While SLE has classically been considered a T cell mediated disease, initiation of SLE in BAFF transgenic mice depends only on signaling through the TLR adaptor molecule MyD88⁷⁸. Intracellular TLRs can be targeted using short synthetic oligodeoxynucleotides (ODN). ODN are more effective for disease prevention than for inducing remission in mice and inhibitors of TLR7 are as effective as inhibitors of both TLR7 and TLR9⁷⁹. The precise role for specific TLR inhibitors will need to be addressed in human clinical trials. Selective TLR9 inhibition may not be an appropriate strategy because it may increase IFN production by immature dendritic cells and thereby induce B cells to secrete pathogenic anti-RNA antibodies⁸⁰. Moreover, humans express a functional TLR8 in monocytes and myeloid DCs⁸¹ and it is unclear whether this is blocked by currently available ODN. Nevertheless, the disease modifying effects of hydroxychloroquine, an inhibitor of autophagosome formation (in which TLRs encounter nucleic acid antigens)⁸², suggest that such reagents may be beneficial.

Conclusions

Despite advances in our understanding of lupus pathogenesis we have not yet identified an effective new therapy for SLE nephritis. A number of clear themes are emerging from the murine studies that bear upon this difficulty. First, mouse models are heterogeneous in their responses to therapies. In particular, mice with excessive activation of the innate immune system are resistant to T cell directed therapies. Second, depletion of circulating autoantibodies does not always translate into an immediate clinical effect. Conversely, if antibody effector functions are blunted it is not essential to remove autoantibodies in order to improve nephritis outcomes. Third, therapies that deplete naïve B cells do not have robust efficacy for active disease and could lead to expansion of autoreactive B cells in the face of generalized immunosuppression. Fourth, while many therapies prevent disease onset, most have significantly compromised efficacy when used at later stages. For established disease, combinations of therapies are more effective than single therapy alone but are considerably more immunosuppressive. Because the inflammatory milieu of chronic SLE profoundly alters T and B cell function and sets up dysregulated amplification loops, restoration of normal tolerance is difficult to achieve once disease has begun. Some changes may be persistent since they are associated with the generation of populations of long-lived effector cells and with alterations in DNA methylation⁸³. Finally, genetic polymorphisms that alter the susceptibility of the kidneys to inflammation and fibrosis will affect prognosis. The data from the murine models suggest that we need to reconsider our current paradigms for the management of SLE. At present we reserve immunologic interventions for patients with active disease and withdraw medications in inactive disease. The most logical strategy is to prevent activation of autoreactive cells before class switching and somatic mutation of immunoglobulin genes, formation of long-lived populations and establishment of multiple dysregulated inflammatory pathways have occurred. Approaches for the treatment of newly diagnosed SLE after remission of the first flare could include relatively safe forms of long-term immune modulation (Table 1) as well as some of the reagents discussed above. Concurrent medical problems such as hypertension and obesity need to be aggressively treated at early stages. In addition, genetic and phenotypic profiling of patients may allow identification of subgroups more likely to respond to particular therapies. Once inflammation is present, intensive immunotherapy can deplete activated B and T cells but strategies need to be developed to avoid reconstitution of an autoreactive repertoire and/or homeostatic proliferation of long-lived populations. Therapies directed at effector pathways need to be developed and used early to protect target organs from damage. Both short and long term outcomes need to be examined in clinical trials. The recent success of belimumab, an agent directed at both autoreactive B cells and target organs, exemplifies the principle that measuring the results of moderate immunomodulation for the maintenance of long term remission requires large and prolonged clinical trials of reagents that target more than one immune pathway.