Are lupus animal models useful for understanding and developing new therapies for human SLE?

Abstract

Systemic lupus erythematosus is a systemic autoimmune disease driven by a complex combination of genetic, environmental, and other immunoregulatory factors. The development of targeted therapies is complicated by heterogeneous clinical manifestations, varying organ involvement, and toxicity. Despite advances in understanding the mechanisms contributing to SLE, only one biologic drug, belimumab, is FDA-approved. The identification and development of potential therapies have largely been driven by studies in lupus animal models. Therefore, direct comparison of both the therapeutic and immunological findings in human and murine SLE studies is critical and can reveal important insights into indeed how useful and relevant are murine studies in SLE drug development. Studies involving belimumab, mycophenolate mofetil, abatacept, rituximab, and anti-interferon strategies generally demonstrated analogous findings in the attenuation of SLE manifestations and modulation of select immune cell populations in human and murine SLE. While further basic and translational studies are needed for identifying SLE patient subsets likely to respond to particular therapeutic modalities and in dissecting complex mechanisms, we believe that despite some inherent weaknesses SLE mouse models will continue to be integral in developing targeted SLE therapies.

1. Introduction

1.1. SLE

Systemic lupus erythematosus (SLE) is a chronic autoimmune disease that primarily affects women, particularly those of childbearing age.¹ In the US, approximately 20 to 150 people out of 100,000 will be diagnosed with SLE, with increased frequency in Asian, African American, African Caribbean, and Hispanic American populations.² Patients may present with a wide array of clinical manifestations ranging from mild arthralgia and rash to more severe features including renal disease, hematologic abnormalities, and neuropsychiatric involvement.

A complex combination of genetic, environmental, hormonal, and other immunoregulatory factors contribute to the loss of tolerance and formation of anti-nuclear autoantibodies observed in SLE. Collectively, evidence implicates an imbalance between the production and clearance of apoptotic material causing activation of toll-like receptors or other nucleic acid recognition receptors that trigger an inflammatory response. This inflammatory response leads to the subsequent upregulation of Type I and II interferons, activation of autoreactive T and B cells, chemokine and cytokine release, and formation of pathogenic autoantibodies, leading in concert to tissue injury and damage in target organs.³

1.2. Challenges in the Treatment of SLE

Current SLE treatments are limited given their incomplete effectiveness, non-curative nature, and toxic side effects. However, developing targeted therapies for SLE patients is challenging given the multiple associated immune abnormalities, the variety of organs that can be involved, and the many potential clinical manifestations. Furthermore, SLE patients are often diagnosed at a relatively advanced stage from an immunological perspective, at which point many immune cell subsets including dendritic cells, myeloid cells, and T and B cells have already been aberrantly activated and memory lymphocytes have been formed. While there is a heritable component observed in SLE, usually limited genetic information is available during the initiation and subclinical stages of the disease prior to the formal diagnosis. Despite these well-recognized challenges, understanding the underlying contributing mechanisms should allow for the development of more precise treatments for affected patients with improved efficacy and less toxicity.

1.3. SLE Animal Models

To study the pathogenesis of SLE, researchers have traditionally utilized mouse models to examine mechanisms contributing to disease. Table 1 summarizes the key clinical features of some commonly used lupus animal models in comparison to human disease. These are also highlighted in several excellent review articles.^4–6 Briefly, the classical spontaneous models include the NZB/W F1 and derived congenic strains (e.g. B6.SLE1.SLE), MRL/lpr, and BXSB/Yaa strains, which are genetically predisposed to develop prototypical SLE-like symptoms including high serum anti-nuclear antibody (ANA) levels and immune complex-mediated nephritis. The pristane-induced and the nephrotoxic nephritis models are both useful induced models to study strong type I interferon (IFN) responses and lupus nephritis, respectively. Table 2 calls attention to several key differences between murine lupus models in widespread use and emphasizes some of the characteristics which have contributed to preferential use of one model over the other, depending on the particular study and the scientific question being asked. Additional details, which are not the focus of this paper, can be found elsewhere.^4–8

Table 1:

Comparison of Human and Mouse SLE Manifestations^*

	Human	Mouse Models
		NZB/WF1	MRL/lpr	BXSB	Pristane-Induced
Year of first description/discovery	1833143	1966144	1976145	1978146	1994147
Female (F)/Male (M) ratio	F>>M	F>>M	F>M	M>>F	F>M
Age of onset4–6,148–150	15–44 years	25 weeks	8 weeks	9 weeks	4 weeks
Mean Survival4–6,148–150		45 weeks	17 weeks	30 weeks	25 weeks
IFN signature5,151	Strong	Weak	Weak	Unclear	Strong
Clinical Manifestations5
ANA/ anti-dsDNA	+	+	+	+	+
Low complement	+	+	+	+	+
Nephritis	+	+	+	+	+
Neurological involvement	+	+	+	−	−
Skin manifestations	+	−	+	−	−
Arthritis	+ <	−	+	−	−
Relative popularity**		5	26	4	3

^*Adapted in part from References 5 and 148.

^**Relative popularity depicts the number of Pubmed research articles (rounded, × 100) containing the respective strain/model names

Table 2:

Key features of SLE Animal Models^*

Category	Mouse Model Family	Associated strains/models	Genetic Loci9–11,152	Implicated Pathways	Organ Manifestations Usually Studied	Advantages	Disadvantages
Spontaneous	NZB/WF1	NZM2328, NZM2410, SLE1-3	Spg3, Nba2, SLE1-3	Genetic susceptibility	nephritis, vasculitis	Multigenic disease; Can be accelerated by interferon	Slow development of disease; Requires cross of two strains to generate lupus prone mice
	MRL/lpr	MRL/+, B6. lpr	Lmb3, Fas	Fas-mediated apoptosis	Nephritis; neuropsychiatric manifestations, skin, vasculitis, arthritis, salivary gland infiltration modeling Sjogren's	Recapitulates many of the clinical manifestations observed in human SLE; Early disease onset; Severe disease	Fas mutations do not drive human SLE; Lympadenopathy, splenomegaly, and double negative T cells are absent or much less common in human lupus
	BXSB	TLR7-tg	Bxs1-6	TLR7-induced production of IFN	nephritis	Evaluation of TLR7-driven mechanisms	Male predominance does not reflect human SLE
Induced	Pristane		---		nephritis, arthritis, pulmonary vasculitis	Inducible in non-autoimmune strains; Model of environmentally-driven disease	Strains are differentially susceptible; Significant early (non-lupus related) mortality; Slow onset; Clinical manifestations are mild
	Resiquimod/ Imiquimod		---	TLR7-mediated inflammation	nephritis, splenomegaly	Induces systemic autoimmunity in both male and female	Limited organ manifestations
	Nephrotoxic serum nephritis (NTN)	Anti-GBM	---		nephritis	Inducible in non-autoimmune strains; Rapid onset; Predictable timing	Strains are differentially susceptible; Limited to nephritis;
Other	Chronic graft versus host disease		---	graft-vs-host reactions; T cell- mediated	nephritis	Inducible in B6; Useful for evaluation of T cell driven mechanisms	Slow onset; Mild disease, limited to nephritis

^*Adapted in part from References 5 and 148.

Both spontaneous and induced models are considered to be useful in studying the etiology and mechanisms involved in SLE. The involvement of T cells, B cells, antigen presenting cells, and macrophages in these mouse models has facilitated the subsequent identification of drug targets theoretically translatable to human disease. Furthermore, genetically manipulating mouse strains has provided strong and convincing evidence supporting genetic contributions to mechanisms of disease in SLE. Genetic studies have identified susceptibility loci in spontaneous SLE mouse models, validated susceptibility genes found by GWAS approaches in human lupus patients in the pathogenesis of the disease (i.e. PTPN22, STAT4), and have functionally characterized the contributions of individual proteins and/or pathways in the development of clinical manifestations.^9–11 A fuller description of the genetic contributions to mouse and human lupus is outside the scope of the current review and can be found elsewhere.^10,11 It is important to note that while many of the genes driving autoimmunity in mouse models don’t have exact parallels in human disease (e.g. Fas mutations in humans are primarily associated with a non-lupus phenotype), nevertheless mice expressing such susceptibility loci can be valuable tools by accelerating the loss of tolerance and making many studies more feasible.⁵ However, these and other important differences between the mouse and human immune systems raise potentially significant concerns regarding the usefulness of SLE mouse models in the development of effective mechanism-driven therapeutics.¹²

1.4. Purpose of Review

This review seeks to determine how helpful SLE animal models have actually been in the development and success of novel drugs in SLE patients. At first glance, it is discouraging as only one drug has been FDA approved in the past 60 years, despite advances in technology and in our understanding of disease. Nevertheless, many drugs are used off-label for lupus and several more are currently under active investigation. Therefore, whether or not the effects of such medications in lupus animal models is applicable and relevant to human disease is a key question with important implications for drug development and clinical research. In this review, the therapeutic and immunological effects of treatments including belimumab, abatacept, mycophenolate mofetil, rituximab, and anti-IFNs will be compared between human and mouse studies, with particular focus on lupus nephritis. By carefully evaluating the successes and failures of these therapies, we can assess the utility of SLE mouse models in the development of effective therapies for human lupus patients.

2. Belimumab

2.1. Introduction

Belimumab, commercially known as Benlysta, was the first targeted biologic approved for the treatment of lupus.¹³ Belimumab is a human IgG1λ. recombinant monoclonal antibody targeting BAFF (B cell-activating factor, also referred to as B Lymphocyte Stimulator (BLyS)). BAFF promotes B cell survival and differentiation and can bind to three potential receptors, BAFF Receptor 3, TACI, and B-cell maturation antigen.¹³ BAFF is upregulated in a number of autoimmune diseases, including rheumatoid arthritis,¹⁴ Sjogren’s,¹⁵ and SLE.^16–19

2.2. BAFF Inhibition in Human SLE

2.2.1. Therapeutic effects

Two prominent multi-centric Phase III clinical trials, BLISS-52 and BLISS-76, evaluated the efficacy of intravenous belimumab in SLE patients over the course of 52 and 76 weeks respectively.^20,21 In both trials, belimumab treatment was associated with improvements in disease activity, reduced incidence and severity of disease flares, and steroid-sparing effects. ^20–22 Further post hoc univariate and multivariate analyses of pooled data from these two trials demonstrated that belimumab was effective, based on SRI (SLE Responder Index) responses at weeks 52 and 76. These analyses also identified baseline factors associated with increased benefit including more severe disease, low complement levels, anti-dsDNA positivity, and baseline corticosteroid use.²³ Mucocutaneous, musculoskeletal, and immunological domains showed a significant reduction in disease activity in the belimumab-treated groups.²⁴

2.2.2. 2^nd Generation Clinical Trials

A number of second generation BAFF modulating therapies have been explored as additional B cell-modulating therapies. However, compared to the success in the belimumab trials, second generation therapies including blisibimod, tabalumab, and atacicept yielded mixed results.^25–28 Recently, a Phase II atacicept trial concluded that while primary endpoints were not met, there was a trend for increased response rates with atacicept, particularly in patients with high disease activity at baseline.²⁸

2.2.3. Immunological Activity

Exploratory studies in a Phase II belimumab trial found that SLE patients with continuous belimumab treatment had significant median percentage reductions in the CD19+, CD20+, naïve (CD20+ CD27-), activated (CD20+ CD69+), and plasmacytoid (CD20+ CD138+) B cell subsets. However, there were no changes in plasma cells. Memory B cells were found be significantly increased at Day 28 but returned to baseline by week 52. Decreased serum IgG, IgA, IgM, and IgE, as well as anti-dsDNA titers, were also observed in the belimumab-treated groups.²⁹ Additional serological findings and cytometry by time-of-flight (CyTOF, also known as mass cytometry) confirmed that the initial B cell number reduction resulted from the preferential decrease in naïve and transitional B cells.³⁰ However, with prolonged belimumab treatment, non-class-switched memory B cells and plasma cells decreased after 18 months while BAFF-independent B cells, including class-switched CD27+ IgD- memory B cells, decreased after >7 years of treatment.^31,32 CyTOF immunophenotyping additionally revealed that CD11c+ CD21- B cell clusters, which resemble age-associated B cells, significantly decreased with belimumab treatment.

Despite depletion of approximately 90% of naïve B cells with belimumab, no differences were observed in V, D, or J family gene usage in unmutated IgM heavy chain sequences with treatment. While belimumab had minimal effects on the naïve B cell repertoire, expression of VH4–34, an immunoglobulin heavy chain gene commonly overrepresented in SLE patients, was lower in plasmablasts of patients treated chronically with belimumab.³² Multiple studies, however, found no significant changes in the number or proportion of T cells or monocytes in belimumab-treated SLE patients.

2.3. The Role of BAFF in Murine SLE

2.3.1. Therapeutic Effects

Several mouse studies have demonstrated the therapeutic potential of targeting BAFF in SLE. Similar to human SLE disease, circulating BAFF levels were elevated in both NZBWF1 and MRL/lpr strains with disease onset and progression.³³ Furthermore, constitutive overexpression of BAFF either in non-autoimmune mice^33–35 or in autoimmune-prone mice³⁶ led to SLE-like manifestations, including hypergammaglobulinemia, increased anti-dsDNA antibodies and circulating immune complexes, splenomegaly, and accelerated development of renal pathology.

The first in vivo studies for belimumab were not performed in mice due to its lower affinity for murine BAFF.³⁷ However, alternative methods to inhibit BAFF pathways were utilized by blocking either the specific BAFF receptor (BAFF-R) or the nonspecific receptor, TACI. In NZBWF1 mice, treatment with TACI-Ig resulted in reduced proteinuria, increased survival, and a significant decrease in peripheral B cells. However, no differences were observed in anti-dsDNA titers between TACI-Ig and PBS groups.³³ In NZM2410, selective BAFF-R blockade and nonspecific TACI blockade both delayed disease onset and induced remission after proteinuria development.³⁸ In BXSB mice, BAFF-R-Ig treatment led to increased survival, decreased renal disease, and reduced autoantibody production.³⁹

Recent studies in mice offer additional insight into the requirement of BAFF in SLE manifestations. BAFF was not be required in the development of SLE-like disease as long as B cell survival was independently preserved.⁴⁰ This BAFF-independent SLE mouse model provides at least a partial explanation for the heterogeneity of responses with BAFF antagonists, and implies that resistant B cells may continue to promote SLE despite BAFF blockade.

2.3.2. Immunological Activity

In regard to the belimumab’s mechanism of action in murine SLE, BAFF blockade in the NZM2410 strain resulted in reduced splenomegaly with significant depletion in T2, marginal zone, follicular B cells, and plasma cells.³⁸ While TACI blockade had more profound plasma cell depletion, particularly IgG-secreting bone marrow cells, no significant differences were seen in serum IgG levels.³⁸ In both this study and in the BXSB mice, activation and expansion of T cells was not affected by BAFF-R blockade.^38,39

3. Mycophenolate Mofetil

3.1. Introduction

Mycophenolate Mofetil, also referred to as MMF or Cellcept, is an immunosuppressant prodrug of mycophenolic acid (MPA). MPA is an inosine monophosphate dehydrogenase (IMPDH) inhibitor, and therefore exerts a cytostatic effect on T and B lymphocytes which have increased dependence on IMPDH for de novo guanosine nucleotide synthesis. Initially used to prevent acute allograft rejection, MMF has subsequently been used to treat multiple rheumatic diseases, including inducing and maintaining SLE remission.⁴¹

3.2. MMF in human SLE

3.2.1. Therapeutic Effects

Both ACR and EULAR recommendations position MMF as the first line drug of choice for the treatment of proliferative LN.⁴² In systematic reviews and meta-analyses, MMF was found to be associated with higher response rates and fewer adverse events of leukopenia, alopecia, and ovarian failure in induction therapy compared to cyclophosphamide (CYC). For maintenance therapy, comparisons between MMF- and CYC-treated patients revealed that while those receiving MMF had a decreased rate of relapse and leukopenia compared to azathioprine (AZA), there was no difference in the rate of end-stage kidney disease or mortality between the two groups.⁴³

In regard to non-renal manifestations, MMF has been found to improve systemic disease activity, and flares were rarely observed.⁴⁴ Furthermore, MMF treatment was associated with clinical improvement and remission of mucocutaneous, cardiovascular, vasculitis, and musculoskeletal manifestations.^44,45 These non-renal effects are seen in patients with or without lupus nephritis.⁴⁶

3.2.2. Immunological Activity

MMF treatment has a significant effect on circulating B cell subsets, particularly CD27^highCD38^high antibody-secreting cells (ASCs). Of those ASCs, a marked decrease was observed in the HLA-DR^high population, typically the predominant ASC population in flaring lupus patients, compared to the HLA-DR^low, which saw only a moderate change. Correlated to this depletion, MMF treatment also affected serum IgG levels, as compared to the AZA-treated or non-immunosuppressive treatment groups which had elevated ASCs.⁴⁷ Additional studies demonstrated that MMF directly inhibits both the proliferation and differentiation of ASCs. Moreover, significantly elevated numbers and percentages of transitional and naïve B cells were noted in the MMF group compared to the AZA group.⁴⁷

With CyTOF immunophenotyping, a significant reduction in the total number of B cells in almost all B cell subsets was observed with MMF treatment, with the exception of IgD⁻ CD27⁻ double-negative memory B cells. Additionally, T cells, particularly T_h17 and T_reg, were found to be significantly decreased, perhaps in response to MMF modulating STAT3 pathways.⁴⁸ Expression of VEGF, PDGF-BB, CXCL12, and CXCL9 was significantly reduced in MMF-treated patients, inferred to be the result of B cell and STAT3 pathway modulation.⁴⁸

3.3. MMF in Murine SLE

3.3.1. Therapeutic Effects

Similar to humans, an increased dependence to IMPDH is observed in mouse lymphocytes.⁴⁹ In the MRL/lpr and NZBWF1 strains, MMF treatment significantly improved survival and decreased albuminuria and serum anti-dsDNA levels.^50–55 Glomerulonephritis in these two strains were ameliorated, with decreased immune complex deposition and glomerulosclerosis.^51,52,54,56 Furthermore, MMF decreased kidney and salivary infiltrates in MRL/lpr mice⁵⁷ and dermal infiltration in NZBWF1.⁵⁸ In the NZBWF1 strain, mice treated with MMF were protected from leukopenia or anemia.⁵⁴ Furthermore, studies have implicated MMF in attenuating premature atherosclerosis in murine SLE.^59,60

3.3.2. Immunological Activity

Further analyses have not yielded a unifying mechanism of action for MMF in murine SLE. In the MRL/lpr strain, conflicting trends were observed in the effect of MMF on nitric oxide and inducible nitric oxide synthetase production, urine nitrate excretion,^51,56 and number of splenocyte T cells.^52,57 Additional comparisons revealed significantly decreased percentages of circulating double-negative T cells, increased serum levels of IL-12, and increased IFNγ and IL-10 expression in the spleen.⁵⁷ In MMF-treated NZBWF1 mice, no differences were observed in the percentage of CD4, CD8, or IgM positive splenocytes. However, the expression of VLA-4 and ICAM-1 was significantly decreased in CD4+ T cells in treated mice. When immunized, the MMF-treated NZBWF1 mice could not mount a prominent antibody response but cytokine production was unchanged.⁵⁵ Additional studies revealed that the inhibited expression of urokinase receptor in podocytes⁶¹ and/or abrogated expression of protein kinase C and fibronectin deposition in the glomeruli and interstitium⁶² could be contributing to the benefit of MMF in lupus nephritis.

4. Abatacept

4.1. Introduction

Abatacept, commercially referred to as Orencia, is the soluble form of two CTLA-4 molecules fused to an immunoglobulin constant region.⁶³ Cytotoxic T lymphocyte Antigen-4 (CTLA-4) is an inhibitory costimulatory molecule that competes to bind CD80 (B7–1) and CD86 (B7–2), thereby playing a critical role in tolerance mechanisms. In addition, CTLA-4 has been shown to play a role in T_reg function. Abatacept is FDA-approved for rheumatoid arthritis, juvenile idiopathic arthritis and psoriatic arthritis.⁶⁴ In SLE patients, soluble plasma CTLA-4 concentrations are significantly elevated compared to healthy controls and are positively correlated to SELENA-SLEDAI disease activity scores.⁶⁵

4.2. Abatacept in Human SLE

4.2.1. Therapeutic Effects

The efficacy of abatacept in human SLE was evaluated in three randomized, double-blind trials: two evaluating lupus nephritis^66,67 and the third evaluating non-life threatening manifestations.⁶⁸ The primary outcome for the 12-month lupus nephritis study was the time to confirmed complete response. Despite being well tolerated, the abatacept arm did not meet its primary outcome, although improvements from baseline in anti-dsDNA, C3, and C4 levels were observed in the abatacept group.⁶⁶ In the second lupus nephritis trial (the ACCESS trial) similar findings were observed, with the abatacept group not achieving the desired impact on the proportion of subjects who achieve complete renal response at week 24.⁶⁷ However, in a reanalysis, the abatacept-treated patients had a >20% response rate compared to the 6% rate in the placebo group, using the rituximab LUNAR trials’ definition for response.⁶⁹ While this alone does not prove abatacept’s efficacy, recent evidence provides support for potentially revisiting abatacept controlled trials as a small cohort of refractory lupus patients saw improvement in SLEDAI scores and articular involvement.⁷⁰

4.2.2. 2^nd Generation Clinical Trials

While no formal studies have been performed with belatacept, a second generation CTLA-4 biologic inhibitor, belatacept was initiated for renal indications in six SLE patients. In this cohort, five patients observed creatinine level stabilization after 6 months of therapy, and three patients saw improvements in SLEDAI-2KG, anti-dsDNA, and C3 levels. This retrospective study provides at least preliminary support for exploring the use of belatacept as an alternative therapy to calcineurin inhibitors and their associated toxicities.⁷¹

4.2.3. Immunological Activity

Limited information is available in regard to abatacept’s mechanistic effects in human SLE. However, baseline samples collected in the abatacept trial for non-life threatening SLE manifestations revealed four immunophenotypic clusters of SLE patients.⁷² Notably, in SLE patients characterized by high levels of plasma cells, activated dendritic cells, neutrophils, and natural killer cells, abatacept treatment was associated with improvement in BILAG scores, time to flare, and C3 and C4 levels. However, in patients characterized by B and T cell abnormalities at baseline, abatacept treatment was not associated with disease improvement.

4.3. Abatacept in Murine SLE

4.3.1. Therapeutic Effects

Initial SLE mouse model studies demonstrated the potential efficacy of CTLA-4 inhibition in mitigating SLE disease.⁷³ CTLA-4Ig suppressed the production of autoantibodies, attenuated lupus nephritis, and prolonged life when given preventatively and as treatment in the NZBWF1 mice.⁷⁴ Interestingly, the benefits of CTLA-4Ig in NZBWF1 mice extended beyond treatment as anti-dsDNA levels were suppressed 3 months after therapy cessation. Furthermore, combination therapy of CTLA-4Ig with CYC in NZBWF1 mice was effective in achieving nephritis remission.⁷⁵ Similar trends in disease attenuation were seen in the MRL/lpr mice, as well as improvements in end organ disease in the kidney and salivary glands.⁷⁶ In the BXSB model, CTLA-4Ig treated mice showed a similar suppression of glomerulonephritis and autoantibody production (both during and after treatment cessation).⁷⁷

4.3.2. Immunological Activity

In non-SLE mice, the role of CTLA-4 in regulating immune responses has been well- documented. Indeed, homozygous CTLA-4 deficiency is fatal around 4 weeks of age due to multi-organ lymphocyte infiltration and tissue damage (e.g., severe myocarditis and pancreatitis).^78,79 In mouse transplantation studies, CTLA-4Ig not only alters T cell activation and proliferation, but also subsequently affects B cell activation as antibody production to T cell-dependent antigens is impaired.^80,81 In B6.MRL/lpr mice, CTLA-4Ig treatment significantly decreased ANA, anti-dsDNA, and IL-17A levels, particularly with the co-administration of IL-10 expressing dendritic cells.⁸² Furthermore, a similar significant trend was observed the proportion of T_h17 to T_reg cells in treated mice.

In the CTLA-4Ig-treated BXSB male mice, the predominant CD4+ T cell population was found to be a naïve T cell phenotype (CD44^low CD45RB^high CD62L^high), compared to the control mice that had predominantly activated/memory T cell characteristics (CD44^high CD45RB^low CD62L^low).⁷⁷ In NZBWF1 mice, the injection of a CTLA-4Ig-expressing adenoviral vector led to a decreased expansion of both IgM and IgG autoreactive B cells, inhibited immunoglobulin class switching, altered pattern of somatic hypermutation, and decreased activated CD69+ CD4 T cell numbers.⁸³ Bone marrow IgG-secreting B cells were unaffected by CTLA-4Ig, likely not requiring T cell-mediated help. In the pristane-induced lupus model, treatment with a B7–1 short hairpin RNA lentivirus or a neutralizing anti-B7–1 antibody led to a significant reduction of serum ANA, IFNγ, and IL-4 levels as well as attenuating the expression of CD11b, CD11c, Gr1, CD21, CD86, MHC II in splenic B cells.⁸⁴

5. Rituximab

5.1. Introduction

Rituximab, commercially known as Rituxan, is a murine-human chimera anti-CD20 monoclonal antibody designed to deplete cells expressing CD20+, ⁸⁵ which is an identifying B cell surface marker. CD20 is widely expressed in most B cell subsets except for pro-B cells and terminally differentiated plasmablasts and plasma cells.⁸⁶ The extensive role of B cells in SLE has been widely documented, both the expansion of abnormal B cells and the role of B cells in the pathogenesis. ^87,88

Despite being the first B cell differentiation antigen discovered, the function and/or regulation of CD20 has not been fully elucidated. CD20 is believed to be involved in the regulation of B cell activation and proliferation^89,90, as well as to constitute part of the multimeric cell surface complex that regulates Ca²⁺ transport across the plasma membrane.^91,92 The use of rituximab has been FDA approved in rheumatoid arthritis⁹³, granulomatosis with polyangiitis, and microscopic polyangiitis.⁹⁴

5.2. CD20 Inhibition in Human SLE

5.2.1. Therapeutic Effects

Quite a few trials have been conducted to determine the efficacy of rituximab in SLE, albeit with mixed success.^86,95–105 EXPLORER and LUNAR were randomized, double-blind clinical trials that evaluated the efficacy and safety of rituximab in SLE patients with moderately-to-severely active extrarenal disease¹⁰⁶ or proliferative lupus nephritis,¹⁰⁷ respectively. In both these trials, there were no significant differences between the rituximab-treated and the placebo groups in their respective primary or secondary outcomes, despite robust depletion of CD19+ cells by 2 weeks after initial infusions. Nevertheless, in the LUNAR trial statistical improvements in anti-dsDNA titers and complement levels were seen in the rituximab-treated group.¹⁰⁷ Similarly, a post hoc analysis revealed a reduction in anti-cardiolipin antibodies and an increase in serum complement and BAFF levels in the rituximab-treated group.¹⁰⁸

5.2.2. 2^nd Generation Clinical Trial Findings

A second generation of anti-CD20 therapies have been designed to be more effective, better tolerated, and less immunogenic.¹⁰⁹ These antibodies include ocrelizumab, obinutuzumab, and ofatumumab, which is fully human. Trials using the fully humanized ocrelizumab were halted either due to a lack of response or an increased risk of severe infections.¹⁰⁹ There is a phase II trial for obinutuzumab currently underway that shows promise for lupus nephritis. Early findings are that 34.9% of obinutuzumab-receiving patients meet the trial’s primary outcome and 91% of obinutuzumab-receiving patients have no detectable B cells by flow cytometry at 52 weeks.^110–112 A single-center retrospective case series described the potential for the use of ofatumumab as a potential alternative agent, with initial evidence of safety and efficacy in B cell depletion.¹¹³

5.2.3. Immunological Activity

Analyzing immune cells population may elucidate rituximab’s effects in human SLE, particularly those B cells resistant to depletion. The resistant B cells were memory, double negative (IgD- CD27-), and CD5+ phenotype, suggested to be CD19+ plasmablasts with little to no CD20 expression.¹¹⁴ Furthermore, rituximab treatment was associated with decreased expression of the costimulatory molecules CD40 and CD80 on B cells. This down-regulation could impact T cells and their activation. However, significant increases in activated CD4+, CD8+, and T regulatory cells were observed with rituximab treatment in two separate studies.^99,114

The role of CD20 in Murine SLE

5.3.1. Therapeutic Effects

In the MRL/lpr and its congenic control, MRL/+, B cell deficiency was achieved by inhibiting heavy chain formation and subsequently B cell maturation. Both B cell-deficient strains were protected from development of glomerulonephritis and interstitial nephritis, and had lower serum IgG and anti-dsDNA levels.¹¹⁵ Additionally, anti-CD20 approaches ameliorated clinical disease in MRL/lpr,¹¹⁶ NZBWF1,^116,117 and pristane-accelerated NZBWF1 strains.¹¹⁸ However, in NZBWF1 mice, anti-CD20 failed to decrease anti-dsDNA or total IgG levels.¹¹⁷

5.3.2. Immunological Activity

For B cell deficient MRL/lpr and MRL/+ strains, additional immunological changes included reduced activated and memory T cell populations, while the percentage of naïve T cells increased in these strains.¹¹⁵ In MRL/lpr mice, a depleting anti-CD20 antibody similarly substantially reduced B cell subsets, which in turn diminished T cell activation.¹¹⁶ However, resistant B cells were found in secondary lymphoid tissue, which was also a feature in MRL/+ and NZBWF1 mice treated with anti-CD20 antibody. Unless treatment was maintained for a long period of time, splenic B cells in MRL/lpr mice were recalcitrant to depletion.¹¹⁶ In NZBWF1 mice, these findings were corroborated but also demonstrated that mice with more severe nephritis had increased resistance to B cell depletion. The residual B cells in the spleen were found to be predominantly marginal zone (CD21^high, CD23^low) and T2 B cells (CD21^high, CD23^high).¹¹⁷ In combination, these findings infer that B cell intrinsic factors in autoimmune mice contribute to the survival of select resistant B cell subsets.

6. Anti-IFN

6.1. Introduction

Anifrolumab is a fully human monoclonal antibody that binds to the Interferon-α receptor 1 (IFNAR) and prevents signaling from all type I interferons.¹¹⁹ As a hallmark of SLE, 60–80% of adult SLE patients have a Type I interferon (IFN) signature, defined as a collection of interferon-stimulated genes upregulated in PBMCs.^120–122 Plasmacytoid dendritic cells and their production of Type I IFN are considered to be critical players in the pathogenesis of SLE.¹²³ Not only have genetic and epigenetic studies identified multiple IFN loci associated with SLE susceptibility, but in some studies the IFN signature correlated with disease activity.^120–122 Furthermore, patients with hepatitis C or melanoma who received IFNα therapeutically have experienced SLE-like manifestations including high ANA titers and arthritis.¹²⁴

6.2. Anti-IFN in Human SLE

6.2.1. Therapeutic Effects

Both the MUSE and TULIP Anifrolumab clinical trials illustrate the potential of anti-IFN biologics in SLE. As a phase IIb, randomized, double-blind study, the MUSE trial met its primary outcome, the percentage of patients achieving a SRI response by 24 weeks of treatment with sustained reduction of corticosteroid use.¹²⁵ At 52 weeks of treatment, anifrolumab-treated patients had significant improvements in SRI response and in organ-specific disease such as cutaneous lupus and arthritis manifestations.¹²⁶ Furthermore, patients with a high IFN signature at screening experienced greater efficacy with anifrolumab.

The TULIP program was comprised of two phase III clinical trials, titled TULIP 1 and 2. Based on the belimumab trials and success in the MUSE trial, the primary outcome for both were originally identified to be a SRI response. However, mixed success was observed in the TULIP 1 trial as the primary outcome was not met and yet the secondary end points suggested treatment efficacy.¹²⁷ Subsequently, the TULIP 2 trial protocol was amended, prior to data analysis, designating BICLA response as its primary endpoint. Notably, it was recently published that the latter primary outcome as well as some secondary end points including improved skin disease and glucocorticoid use were met in the TULIP 2 trial.¹²⁸ Overall, improvements in the BICLA response, skin disease, and flare reduction were observed in all three trials, with a SRI response observed only in the TULIP-2 and MUSE trials.^126–128

6.2.2. Additional Anti-IFN therapies

Additional anti-IFN therapies have been explored in SLE, including rontalizumab and sifalimumab, human anti-IFN-α monoclonal antibodies, and AMG811, a human anti-IFNγ antibody. Similar to the anifrolumab trials, sifalimumab met its primary outcome in a phase IIb randomized, double-blind trial, reducing organ-disease manifestations and other disease activity composites.¹²⁹ In contrast, rontalizumab trials have been discontinued after the rontalizumab ROSE phase II study failed to meet either primary or secondary outcomes even in SLE patients with high IFN signature metric scores.¹³⁰ Initial data in Phase Ib AMG811 SLE trials demonstrated efficacy and reduced IFNγ-associated biomarkers but yielded no evidence for clinical improvement.¹³¹

6.2.3. Immunological Activity

Anifrolumab treatment rapidly and sustainably reversed SLE-associated neutropenia, lymphopenia, monocytopenia, and thrombocytopenia.¹³² Specifically, significant increases in class-switched memory B cells and in a number of T cell subsets including CD4, CD8, and CXCR5+/− memory cells were observed in anifrolumab-treated patients. Furthermore, T and B cell-targeting chemokines were significantly decreased including CXCL13, BAFF, CCL19, and endothelial cell markers (i.e. VCAM-1). However, for patients with low interferon signature scores, no significant differences were observed in lymphocyte or neutrophil numbers or subsets. The ability of anifrolumab to suppress plasmacytoid dendritic cell-driven Type 1 interferon production was evaluated in healthy controls.¹³³

6.3. Anti-IFN in Murine SLE

6.3.1. Therapeutic Effects

Numerous studies have been performed in SLE mouse models to elucidate the role of interferons in the pathogenesis of murine SLE. IFNAR-deficient NZBWF1, B6/MRL-lpr, and NZM2328 mice showed reduced nephritis and less kidney IgG deposition, as well as reduced anti-dsDNA levels.^134–136 In BXSB male mice treated with anti-IFNAR antibodies, similar improvements in glomerulonephritis and kidney deposits were observed.¹³⁷ However, inconsistent results have been observed in the MRL/lpr strain. A transient therapeutic effect was noted with anti-IFNAR treatment as anti-RNA autoantibodies and proteinuria were significantly reduced at 12 weeks of age but not at 16 weeks.¹³⁷ In genetic studies however¹³⁸, the Type I IFN receptor deficiency worsened lymphoproliferation, autoantibody production, and end organ damage, while the Type II IFN receptor deficiency protected MRL/lpr mice from autoimmune manifestations. Some of the variability in the murine studies has been attributed to the fact that different from human SLE, IFN signatures vary in SLE animal models, ranging from low/weak (MRL/lpr and NZBWF1) to strong (pristane-induced).¹³⁹

6.3.2. Immunological Activity

In the IFNAR −/− NZBWF1 mice, significant decreases in T, B cells, and macrophages were identified, correlating with the observed decrease in splenomegaly. While generally the proportions of these populations were not affected, the frequency of activated CD19+ CD69+ B cells were reduced with IFNAR deficiency which correlated to decreased levels of autoantibodies. Furthermore, IFNAR deficient mice had decreased proliferative in vitro and in vivo responses of B and T cells.¹³⁴ Comparatively, in IFNAR −/− NZM2328 mice, both the percentage and number of activated CD40^hi plasmacytoid dendritic cells were reduced in the renal lymph node compared to the control NZM2328 mice at 2 months of age, suggesting an early IFNAR-dependent response.¹³⁵ At 5 months of age, splenic dendritic cells were reduced in IFNAR deficient NZM mice and had decreased CD40 expression.

In the BXSB mice treated with anti-IFNAR antibody, the decrease in splenic B cells correlated with the attenuated splenomegaly observed.¹³⁷ Furthermore, the percent of activated CD69+ CD4 and CD8 T cells, number of CD11b+ CD11c- monocytes, and MHC I expression in cDCs were reduced in these mice.

7. Discussion

Many challenges limit the development of targeted therapies for SLE, including the heterogeneous manifestations observed in SLE patients and the complex immunological mechanisms underlying this disease. The development of mechanism-driven therapies has been predominantly based on evidence in human SLE patients (i.e. the expansion of autoreactive B cells and/or Type I IFN signature) as well as immunological aberrations in SLE mouse models. However, a number of biologics and immunosuppressants have been tested in SLE clinical trials with little success.

Comparisons between human and mouse SLE findings, particularly the immunological effects, provide insights into the successes and failures of the drugs evaluated in this review. These findings are summarized in Table 1. For the most part, similar therapeutic effects were observed between mouse SLE studies and corresponding clinical trials, including decreased autoantibody production and improved kidney disease. Furthermore, both depleted and resistant immune populations in the treated SLE patient populations were similarly reflected in the mouse studies. This is notable as mouse models could potentially identify enduring cell populations that can potentiate disease manifestations despite treatment. For example, immunological changes associated with belimumab and rituximab treatment revealed not only varied effects between B cell-mediated therapies, but also identified similar populations of B cells resistant to depletion in both mouse and human SLE. Therefore, while the varied clinical responses in treated SLE patients sometimes appear discordant with the more uniform beneficial effects reported in mouse studies, it further demonstrates the complexity of mechanisms driving SLE.

While no singular SLE mouse model perfectly recapitulates human SLE disease, the flexibility of genetic mouse studies can shed some light not only on the potential efficacy of novel drug targets, but also in discerning nuanced mechanisms after a clinical trial. For example, recent mouse studies have revealed the potential redundancy of BAFF in the pathogenesis of SLE-like manifestations, and therefore a possible explanation for a lack of response observed in select SLE patients.

An important facet in the development of a drug from mouse studies to clinical trials is deciding, in advance, on the primary outcomes of a trial, which as we have seen can greatly affect the overall success or failure of a novel drug. There are multiple examples of post hoc and meta-analyses of lupus trials in which previously unrecognized but significant differences were elucidated between the treatment and placebo groups. Recent clinical trial findings in the anifrolumab TULIP trials further highlight the importance of selecting the most appropriate primary outcome.

Recent technology advances are poised to further expand our understanding of both murine and human SLE. For the spontaneous SLE models, the use of the CRISPR/Cas9 system can increase the ease and feasibility of studies in the spontaneous SLE models with their unique genetic backgrounds. Additionally, immunophenotyping techniques, such as CyTOF, have revealed clusters of SLE patients who were more likely to respond to various treatment strategies. Moreover, single cell RNA sequencing methodologies can provide intricately detailed gene expression information from minute amounts of human biopsy tissue obtained for clinical indications.^140–142 Future clinical trials may benefit from these and other methodological advances and improvements in classification criteria in identifying subsets of SLE patients who will more likely benefit from more precise and targeted therapies.

In summary, the comparisons revealed that SLE mouse models are not only integral in demonstrating potential efficacy, but also in detailing possible mechanisms of action for both the success and failure of treatments after clinical trial completion. While no one single mouse model replicates the human SLE immune system in all its variability and complexity, we believe that lupus animal models are enormously valuable, are currently irreplaceable, and are likely to remain a cornerstone of drug development efforts for human SLE for years to come.

Table 3:

Summary of Drug Development Findings in Human and Mouse SLE

	Human SLE		Mouse SLE
	TheraDeutic	Immunolosical	Therapeutic	Immunoloeical
Belimumab	↓ : disease activity, organ system manifestations, incidence and severity of disease flares, steroid-sparing effects ↓ : anti-dsDNA titers	↓ : naïve, activated, plasmacytoid B cells - : plasma cells, class-switched memory B cells, T cells	↓ : disease activity, autoantibody production	↓ : T2, marginal zone, follicular B cells, plasma cells - : T cells
MMF	↓ : disease activity, organ system manifestations, incidence of flares	↓ : almost all B cell subsets, Th17 and Tres cells ↓ : VEGF, PDGF-BB, CXCL12, CXCL9, B cell and STAT3 pathways - : double-nesative memory B cells	↓ : disease activity, organ infiltrates and manifestations, Anti-dsDNA levels	Mixed results ↓ : podocyte urokine receptor expression, PKC/fibronectin deposition in glomeruli/interstititium
Abatacept	↓ : anti-dsDNA, complement levels - : efficacy	↓ : disease activity in patients with high plasma, dendritic, neutrophil, and natural killer cells	↓ : disease activity, organ manifestations, decreased autoantibody production	↓ : T and B cell activation/proliferation ↓ : serum IFNg, IL-4, CD11b, CD11c, Gr1, CD21, CD86, MHCII in splenic B cells
Rituximab	- : efficacy ↓ : anti-dsDNA, anti- cardiolipin ↓ : complement, BAFF levels	- : double negative and CD5+ memory B cells ↓ : CD40, CD80 expression on B cells ↓ : activated CD4, CD8, Treg cells	↓ : disease activity, autoantibody production	↓ : B cell subsets, T cell activation - : splenic MZ and T2 B cells
Anti-IFN	↓ : disease activity, organ system manifestations, incidence and severity of disease flares, steroid-sparing effects	↑ : class-switched memory B cells, CD4, CD8, CXCR5 +/- memory cells ↓ : CXCL13, BAFF, CCL19, endothelial cell markers	↓ : disease activity, organ manifestation, autoantibody production	↓ : activated T, activated B cells, macrophages, CD40hi pDC

Symbols: ↓ indicate a decrease or improvement; ↑ indicate an increase; - indicate no change

Highlights.

• Similar therapeutic and immunological effects are observed in human and murine SLE studies

• Studies with belimumab and rituximab treatment reveal resistant B cell populations, present in both human and murine SLE

• CyTOF immunophenotyping can identify potential SLE patient clusters more likely to respond to specific therapies

• Testing in lupus animal models remains an important avenue in SLE drug development