Topic Article
B-Cell Depletion Immunotherapy in Pemphigus: Effects on Cellular and Humoral Immune Responses
Hugo Mouquet, Philippe Musette, Marie-Lyse Gougeon, Serge Jacquot, Brigitte Lemercier, Annick Lim, Danièle Gilbert, Ingrid Dutot, Jean C Roujeau, Michel D’Incan, Christophe Bedane, François Tron and Pascal Joly
Journal of Investigative Dermatology (2008), 128, 2859-2869; doi: 10.1038/jid.2008.178

Why Rituximab Works

Julia Tzu ¹, Nancy Kim ¹ and Robert S. Kirsner ¹

Journal of Investigative Dermatology (2008) 128, 2744. doi:10.1038/jid.2008.359

Pemphigus encompasses a group of autoimmune blistering diseases in which circulating immunoreactants—most notably IgG (particularly IgG1 and IgG4)—bind to the extracellular portions of desmosomal keratinocyte adhesion molecules (Sekiguchi et al., 2001) known as desmogleins and disrupt the adhesion between adjacent keratinocytes, producing intraepidermal blisters. In the two idiopathic forms of pemphigus, pemphigus foliaceus (PF) and pemphigus vulgaris (PV), antibodies initially target desmogleins 1 (Dsg1) and 3 (Dsg3), respectively, leading to skin disease in patients with PF and skin and mucosal disease in patients with PV.

Recently, rituximab, a chimeric monoclonal antibody that targets the CD20 molecule, has been reported to benefit patients with PV (Joly et al., 2007a). CD20 is found exclusively on the surfaces of normal and neoplastic B cells. Rituximab is thought to act via a B-cell-depleting mechanism that includes complement-dependent cytotoxicity, antibody-dependent cellular cytotoxicity, and the induction of apoptosis (Johnson and Glennie, 2003). Thus, rituximab has been used in various B-cell-mediated malignancies such as B-cell non-Hodgkin’s lymphoma (Cheson et al., 2002) and in autoimmune diseases such as rheumatoid arthritis, systemic lupus erythematosus, and dermatomyositis (Leandro et al., 2002; Looney et al., 2005; Levine, 2005). Rituximab’s benefit in patients with PV is thought to be due to the depletion of B cells that would otherwise give rise to pathogenic antibody–producing plasma cells.

Mouquet et al. (2008) studied 21 PV patients treated with rituximab and analyzed subsequent immunological changes, correlating them with clinical outcome. Total B-cell depletion was accompanied by significant decreases in IgM, but not IgG, levels. Over time, re-emergence of naive blood B lymphocytes occurred. Clinical responses were seen in patients who had decreased levels of anti-Dsg autoantibodies after treatment. Antimicrobial IgG levels did not change with rituximab treatment. The investigators also examined the VH-IgM and VH-IgG B-cell-receptor repertoire in two patients with PF and one with PV, both before and after rituximab treatment. In these patients, distortions of VH-IgM and VH-IgG were found before treatment, but were not found after B-cell reconstitution, following rituximab therapy. This suggests a mechanism by which rituximab may work and exhibit long-lasting effects: by initially depleting autoreactive B cells, leading to elimination of anti-Dsg autoantibodies. Eventually restoration occurs, but with naive (as opposed to primed) B lymphocytes with a diverse B-cell repertoire.

Through the following questions, we examine this paper in greater detail.

QUESTIONS
1. What is the evidence for a mechanism of pemphigus pathogenesis, and which anti-desmoglein antibodies are responsible?
2. Why was rituximab considered a potential treatment for pemphigus?
3. Describe the methodological techniques employed in this study.
4. What were the findings of this study?
5. What may be the clinical implications of this article?

ANSWERS
1. Pemphigus encompasses a group of autoimmune blistering diseases in which circulating IgG (particularly IgG 1 and IgG 4) binds to the extracellular/N-terminal portions of the desmosomal keratinocyte adhesion molecules known as desmogleins (Sekiguchi et al., 2001) and disrupts the adhesion between adjacent keratinocytes, producing intraepidermal blisters (Hashimoto et al., 1995). The three main categories of pemphigus include pemphigus foliaceus (PF), pemphigus vulgaris (PV), and paraneoplastic pemphigus. For our current purposes, we focus mainly on PF and PV. In PF, antibodies target desmoglein 1 (Dsg1), leading solely to cutaneous involvement due to the more superficial expression of this antigen within the epidermis and lack of this antigen in nonkeratinized mucosal skin. In PV, antibodies target Dsg3 and sometimes Dsg1 as well, leading to mucous and/or mucocutaneous involvement due to, respectively, deep and superficial expression of these antigens within the skin (Amagai et al., 1999; Mahoney et al., 1999).

The antibodies observed in pemphigus patients are now known to play a pathogenic role. Evidence in support of the pathogenicity of these antibodies comes from several studies. Roscoe et al. (1985) demonstrated that when IgG was purified from the sera of patients with fogo selvagem, an endemic form of PF, and injected into the peritoneum of neonatal BALB/c mice, blisters clinically, histologically, and immunologically similar to the human pemphigus counterpart developed. Injection with sera from patients with higher antibody titers produced more rapid blistering. Blisters subsequently resolved with discontinuation of the antibody injections.

Rock et al. (1990) further demonstrated that, even when the injected IgG antibodies from fogo selvagem patients were limited to monovalent Fab fragments, BALB/c neonatal mice continued to exhibit blistering. This suggested that complement activation and antibody crosslinking of surface desmogleins may not be necessary for dissolution of keratinocyte adhesion.

Schiltz and Michel (1976) reported that when IgG derived from patients with PV and PF was added to skin maintained in an organ-culture system, even in the absence of complement or inflammatory cells, blistering indistinguishable from biopsies obtained from pemphigus patients developed. Again, this confirmed the primary role of anti-desmoglein antibodies in the pathogenesis of intraepidermal blistering, with no necessity for other serum factors.

Neonates of mothers who have PV and PF exhibit a transient expression of the respective disease due to passive transfer of maternal IgG 4 antibodies. As these maternal antibodies eventually catabolize, the course of the disease also resolves in the infant (Hirsch et al., 2003; Parlowsky et al., 2003).

2. Understanding the mechanism by which rituximab effects its action requires some understanding of basic immunology. Human B lymphocytes are derived from pluripotent B cells and express different antigenic cell-surface markers at different stages of development. The CD20 molecule, also known as human B-lymphocyte-restricted differentiation antigen Bp35, first exhibits surface expression at the early pre-B-cell stage. Expression of this molecule persists through the mature B-lymphocyte stage and ceases once differentiation into plasma cells is complete. Rituximab is a chimeric monoclonal antibody of the IgG idiotype that targets the CD20 molecule found exclusively on the surface of normal B cells as well as neoplastic B cells (Reff et al., 1994). Its B-cell-depleting mechanism of action includes complement-dependent cytotoxicity, antibody-dependent cellular cytotoxicity, and induction of apoptosis (Johnson and Glennie, 2003). Because it depletes pathogenic and antigen-presenting B cells, rituximab has been used in various B-cell-mediated malignancies such as B-cell non-Hodgkin’s lymphoma (Cheson, 2002) and in autoimmune diseases such as rheumatoid arthritis, systemic lupus erythematosus, and dermatomyositis (Leandro et al., 2002; Looney et al., 2005; Levine, 2005). Pemphigus is an autoimmune condition caused by pathogenic anti-desmoglein IgG antibodies produced predominantly from plasma cells that differentiate from B cells. By depleting the B cells that give rise to pathogenic antibody-producing plasma cells, it is thought that rituximab may also be an effective treatment for pemphigus.

3. Twenty-one patients with PV and PF from 13 centers in France were recruited and enrolled on the basis of specific inclusion and exclusion criteria. The patients were treated with one cycle of four weekly infusions of rituximab 375 mg/m2 bovine serum albumin for 1 month, then followed every month for 1 year and every other month in the second year.

Heparinized peripheral blood was collected from these patients after treatment with rituximab, and lymphocytes isolated from the blood were subjected to flow cytometry to detect quantities of CD19+ lymphocytes at various intervals, recorded as days (i.e., 0, 21, 65, 90, 120, 180…540, 720) after rituximab treatment. Flow cytometry was also used to detect percentages of lymphocytes expressing markers CD3, CD4, CD19, CD20, CD21, CD22, CD23, CD5, CD27, CD38, CD24, CD56, CD86, CD25, CD26, CD28, CD38, CD45RA, CD45RO, CD70, CD71, CD69, and CD134, using respective fluorochrome-labeled antibodies against each CD marker, at various time points up to 24 months after treatment. Total IgG versus IgM levels in patient sera at various time points (days 0, 180, 360, 540, and 720) were quantified using immunonephelemetry.

Immunoscope analysis, also known as spectratyping, a technique recently adapted by Lim et al. (2008) for use in B cells, was employed to examine the Vh-IgM and Vh-IgG B-cell-receptor (BCR) repertoire from two patients with PF and one patient with PV before and after rituximab treatment. In examining the BCR repertoire, the investigators analyzed the distribution of lengths of the CDR3s for the Vh chains of IgM and IgGs (specifically Vh3a, Vh3b, and Vh4) expressed by the B cells before and after rituximab treatment, at days 240 for PF patient P05, at day 500 for PV patient P20, and at day 600 for PF patient P21. It is unclear why the investigators chose different time points at which to analyze the BCR repertoire of each patient.

Anti-Dsg1 and Dsg3 titers were determined via ELISA from the sera of 20 patients at the same time points after rituximab treatment that were used in the flow cytometry analysis. In patients who had persistently anti-Dsg3 IgG despite partial and complete clinical remission, quantification of IgG subclasses was analyzed up to 1 year after rituximab treatment. Anti-pneumococcal capsular polysaccharide (PCP) and anti-tetanus toxoid (TT) antibody levels were also measured via ELISA in 10 of these patients at day 0 up to 1 year after rituximab treatment.

4. Rituximab induces depletion of peripheral B cells, as defined through CD19/20/22 marker positivity, observed as soon as 21 days and lasting for up to 2 years. Using immunonephelemetry, patient serum IgM levels were observed to drop steadily after rituximab therapy (as detected at days 180, 360, and 540), although IgG levels remained unchanged. These results are depicted graphically in Figure 1 of the article (Mouquet et al., 2009).

B-cell reconstitution occurred at 6 months after rituximab therapy. The reconstituted B cells are of CD19+27– (naive) phenotype, with a higher proportion than before rituximab treatment, and CD5+/CD38 ^high/CD24 ^high/CD19 cells (transitional) phenotype, also higher in proportion than before rituximab treatment. The percentage of activated B cells was the same before and after treatment. These results are depicted graphically in Figure 2 of the article.

Pre- versus post-rituximab BCR repertoires for VH-IgM and VH-IgG were analyzed in three (two PF, one PV) and two (one PF, one PV) patients, respectively. Interestingly, when the IgM and IgG CDR3 length distribution in these patients was analyzed, normalization—defined as restoration toward CDR3 length distributions found in control patients, with trends from monoclonal to polyclonal peaks—was found after rituximab therapy. The time point after rituximab therapy at which the data were obtained differed among all three patients; this lack of standardization was not explained by the authors. These results are depicted graphically in Figure 3 of the article.

Percentage of usage of the major VH families (VH3a, VH3b, and VH4) for IgM and IgG was also found, via immunoscope analysis, to differ before and after rituximab treatments in all three patients. No particular pattern could be drawn from the frequency of usage in any of the VH groups (for both immunoglobulin idiotypes) before and after rituximab treatment, nor can any inferences be drawn from the limited data set. Inclusion of this data probably was simply for interest. These results are depicted in Table 1 of the article.

Using ELISA quantification, anti-Dsg antibodies were found to be decreased in most patients. However, a subset of PV patients showed persistently high levels of anti-Dsg1 and anti-Dsg3 antibodies despite partial to complete remission. Most of these high levels of pathogenic antibodies were of the IgG 4 subtype, consistent with the literature on pathogenic IgG subtypes for pemphigus. Levels of antimicrobial antibodies as measured from the sera of 10 patients using the anti-PCP and anti-TT antibodies were found to be unchanged before and after rituximab treatment. These results are depicted graphically in Figure 4 of the article.

CD3+ and CD4+ T cells, T-cell subpopulations, natural killer cells, and T cell–produced cytokines (INF-γ, IL-2, and TNF-α) were unchanged before and after rituximab therapy for up to 1 year.

The premise of the study by Mouquet et al. is that rituximab is clinically efficacious for the treatment of pemphigus, and the article references a previously published article that discusses rituximab’s clinical efficacy (Joly et al., 2007a). The current study provides an immunologic basis for this clinical data. With minor exceptions, the immunologic data seem to be consistent with the clinical data. By its anti-CD20 activity, rituximab does effect B-cell depletion followed by reconstitution with a different set of transitional and naive B cells of a normalized and nonpathogenic repertoire. Hence, pathogenic precursors to anti-Dsg antibody-secreting plasma cells would be eliminated.

The authors do not explain why rituximab does not affect total serum IgG levels. They address this by invoking the existence of long-lived IgG-secreting plasma cells versus short-lived IgM-secreting plasma cells. If IgG is constantly produced by long-lived plasma cells and IgM is produced by mature B cells and short-lived plasma cells, once rituximab eliminates all B-cell precursors that give rise to these plasma cells, total IgM levels should decline, given that the short-lived plasma cells soon die off, whereas the long-lived plasma cells continue to secrete IgG (Radbruch et al., 2006; Vallerskog et al., 2007). Although no data in the article support this speculation, it helps explain the findings.

Even though rituximab does not affect total IgG levels, most patients who experience clinical remission from rituximab exhibit decreased (pathogenic) anti-Dsg IgG levels. One can attempt to explain this inconsistency by assuming that the pathogenic IgG subset is only a small percentage of total IgG levels and hence the technique used to quantify total sera IgG would not detect any differences. The authors did not note this in their discussion.

Furthermore, pathogenic anti-Dsg IgG4 antibodies persisted in a small percentage of patients who show clinical remission after being treated with rituximab. There is no clear explanation for this phenomenon, although Joly et al. (2007a,b) claim that these persistent antibodies no longer recognize the pathogenic epitope originally targeted and are therefore not the “pathogenic” anti-Dsg IgG4 we generally believe all anti-Dsg IgG4 antibodies to be.

5. This article argues for the addition of a new drug, rituximab, to the armamentarium of treatments for pemphigus. Rituximab can be used to produce clinical remission in patients with pemphigus, with no obvious effect on cell-mediated or humoral immunity. The effect of rituximab is selective for B cells, including the pathogenic subset involved in the disease, and the B-cell repertoire normalizes after treatment. However, one important aspect that this paper does not address is the effect of rituximab on the nonpathogenic B cells that are also depleted and that may have important roles in a patient’s overall humoral immunity and infection susceptibility, a concern pertinent to other immunodulating biologics therapies. In addition, although paraneoplastic pemphigus was not addressed in this article, rituximab may work for this disease as well, via the same immunologic mechanisms.

REFERENCES
Amagai M, Tsunoda K, Zillikens D, Nagai T, Nishikawa T (1999) The clinical phenotype of pemphigus is defined by the anti-desmoglein autoantibody profile. J Am Acad Dermatol 40(2 Pt 1):167–70

Cheson BD (2002) Hematologic malignancies: new developments and future treatments. Semin Oncol 29(4 Suppl 13):33–45

Hashimoto T, Amagai M, Garrod DR, Nishikawa T (1995) Immunofluorescence and immunoblot studies on the reactivity of pemphigus vulgaris and pemphigus foliaceus sera with desmoglein 3 and desmoglein 1. Epithelial Cell Biol 4:63–9

Hirsch R, Anderson J, Weinberg JM, Burnstein P, Echt A, Fermin J et al. (2003) Neonatal pemphigus foliaceus. J Am Acad Dermatol 49(2 Suppl Case Reports):S187–9

Johnson P, Glennie M (2003) The mechanisms of action of rituximab in the elimination of tumor cells. Semin Oncol 30(1 Suppl 2):3–8

Joly P, Mouquet H, Roujeau JC, D’Incan M, Gilbert D, Jacquot S et al. (2007a) A single cycle of rituximab for the treatment of severe pemphigus. N Engl J Med 356:545–52

Joly P, D’Incan M, Musette P (2007b) Rituximab for pemphigus vulgaris. N Engl J Med 356:521; author’s reply, 521–2

Leandro MJ, Edwards JC, Cambridge G (2002) Clinical outcome in 22 patients with rheumatoid arthritis treated with B lymphocyte depletion. Ann Rheum Dis 61:883–8

Levine TD (2005) Rituximab in the treatment of dermatomyositis: an open-label pilot study. Arthritis Rheum 52:601–7

Lim A, Lemercier B, Wertz X, Pottier SL, Huetz F, Kourilsky P (2008) Many human peripheral VH5-expressing IgM+ B cells display a unique heavy-chain rearrangement. Int Immunol 20:105–16

Looney RJ, Anolik J, Sanz I (2005) Treatment of SLE with anti-CD20 monoclonal antibody. Curr Dir Autoimmun 8:193–205

Mahoney MG, Wang Z, Rothenberger K, Koch PJ, Amagai M, Stanley JR (1999) Explanations for the clinical and microscopic localization of lesions in pemphigus foliaceus and vulgaris. J Clin Invest 103:461–8

Mouquet H, Musette P, Gougeon M-L, Jacquot S, Lemercier B, Lim A et al. (2008) B-cell depletion immunotherapy in pemphigus: effects on cellular and humoral immune responses. J Invest Dermatol 128:2859–69

Parlowsky T, Welzel J, Amagai M, Zillikens D, Wygold T (2003) Neonatal pemphigus vulgaris: IgG4 autoantibodies to desmoglein 3 induce skin blisters in newborns. J Am Acad Dermatol 48:623–5

Radbruch A, Muehlinghaus G, Luger EO, Inamine A, Smith KG, Dörner T et al. (2006) Competence and competition: the challenge of becoming a long-lived plasma cell. Nat Rev Immunol 6:741–50

Reff ME, Carner K, Chambers KS, Chinn PC, Leonard JE, Raab R et al. (1994) Depletion of B cells in vivo by a chimeric mouse human monoclonal antibody to CD20. Blood 83:435–45

Rock B, Labib RS, Diaz LA (1990) Monovalent Fab’ immunoglobulin fragments from endemic pemphigus foliaceus autoantibodies reproduce the human disease in neonatal Balb/c mice. J Clin Invest 85:296–9

Roscoe JT, Diaz L, Sampaio SA, Castro RM, Labib RS, Takahashi Y et al. (1985) Brazilian pemphigus foliaceus autoantibodies are pathogenic to BALB/c mice by passive transfer. J Invest Dermatol 85:538–41

Schiltz JR, Michel B. Production of epidermal acantholysis in normal human skin in vitro by the IgG fraction from pemphigus serum. J Invest Dermatol 67:254–60

Sekiguchi M, Futei Y, Fujii Y, Iwasaki T, Nishikawa T, Amagai M (2001). Dominant autoimmune epitopes recognized by pemphigus antibodies map to the N-terminal adhesive region of desmogleins. J Immunol 167:5439–48

Vallerskog T, Gunnarsson I, Widhe M, Risselada A, Klareskog L, van Vollenhoven R et al. (2007) Treatment with rituximab affects both the cellular and the humoral arm of the immune system in patients with SLE. Clin Immunol 122:62–74

¹ Department of Dermatology and Cutaneous Surgery, University of Miami Miller School of Medicine, Miami, Florida, USA