B cell depletion in diffuse progressive systemic sclerosis: safety, skin score modification and IL-6 modulation in an up to thirty-six months follow-up open-label trial

Although the pathogenesis of systemic sclerosis (SSc) remains unknown, the B cell abnormalities characterized by autoantibody production [1], hyper-γ-globulinemia and polyclonal B cell hyperactivity [2] are thought to play an important role in the disease. It has been previously described that SSc patients have distinct abnormalities of blood homeostasis and B cell compartments, characterized by expanded naïve cells and activated, but diminished, memory B cells [3]. Furthermore, the expression of CD19, a critical signal transduction molecule of B cells that regulates autoantibody production, is significantly increased in memory and naïve B cells in SSc patients [3, 4]. Analysis of DNA microarrays of cutaneous biopsies from diffuse SSc (dSSc) patients demonstrated a higher expression of clusters of genes of CD20-positive cells [5].

In the tight-skin mice, a genetic model of human SSc, the CD19 signaling pathway appeared to be constitutively activated [6, 7] and the loss of CD19 expression significantly up-regulated surface IgM expression, completely abrogated hyper-γ-globulinemia and autoantibody production, and also inhibited IL-6 production [7]. Additionally, in this animal model, the down-regulation of B cell function led to a decrease in skin fibrosis during the disease onset [8]. Likewise, in a bleomycin-induced SSc mouse model, another animal model that shares many characteristics with human SSc, CD19 deficiency inhibited the development of skin and lung fibrosis, hyper-γ-globulinemia, and autoantibody production [9]. Thus, B cells could have a relevant impact on the development of fibrotic changes as reported in the mouse scleroderma models [6‐9] and also in CCl₄-induced liver injury, in an antibody- and T cell-independent manner [10].

In several studies focusing on the pathogenesis of SSc, the increased levels of IL-6 in the skin, serum, and bronchoalveolar lavage fluid of SSc patients suggest a role of this cytokine in promoting fibrosis by enhancing inflammation [11‐13]. Furthermore, immunohistochemistry data demonstrated an over-expression of IL-6 on endothelium and fibroblasts of involved skin of scleroderma patients compared with normal skin [14]. SSc dermal fibroblasts constitutively produce about a four-fold increase in IL-6 levels with respect to healthy controls fibroblasts [15] and secretion of IL-6 from lung fibroblast is induced by SSc lung-derived B cells [16]. Recently, it has been reported that B-cell activating factor (BAFF), an essential component of B cell homeostasis and a potent B-cell survival factor associated with autoimmune disease in humans, is increased in SSc patients compared with healthy controls [17]. In the tight-skin mice, BAFF antagonist augmented anti-fibrogenic cytokines and inhibited the development of skin fibrosis. Finally, after BAFF stimulation, B-cells had a significantly enhanced ability to produce IL-6 [18].

Two recent open-label studies reported the safety of anti-CD20 treatment in SSc patients; despite both studies describing a decrease in myofibroblast score on serial skin biopsies after treatment, only one reported an improvement in skin score [19, 20]. In these two studies, lung function remained stable during follow up, whereas a case report suggested a possible beneficial role of rituximab on lung involvement in scleroderma disease [21].

The primary aim of the current prospective study was to evaluate the changes in the skin score from baseline to at least 6 up to 36 months of follow up after anti-CD20 therapy. Secondary aims were to assess the potential efficacy of rituximab on lung function, to investigate the modification in IL-6 and BAFF serum levels as biological parameters of disease activity, and to correlate the clinical characteristics with the immune cell infiltrate detected by immunohistochemistry.

The results of our study suggest that B cell depletion in patients with early and progressive dSSc, leads to a clinically relevant decrease in skin involvement and to a stabilization of organ function. The only patient who showed a less clear-cut response either in terms of severity and activity indexes was the one with a long-standing disease.

In our study, the safety of anti-CD20 treatment in SSc patients was also confirmed in up to 36 months of follow up. The observed skin score improvement is more than expected as the spontaneous improvement in patients with similar disease, and comparable with the study by Smith and colleagues [19]. Recently, two studies assessed the safety of anti-CD20 treatment in scleroderma patients. In the first open-label trial, eight SSc patients experienced a skin score improvement up to 43% after 24 weeks from the beginning of anti-CD20 treatment [19], while in the second, a cohort of 15 SSc patients, with a follow up of 12 months, showed no improvement in the skin score [20]. Only the first group used the corticosteroids premedication.

In these two studies, all SSc patients, as in our study, had an early diffuse disease and patients were similar for age, disease duration and clinical characteristics [19, 20]. It is interesting to note that despite little changes reported in the skin score after rituximab treatment in the largest cohort, a decrease in myofibroblast score was observed in several patients [20]. As the myofibroblast score correlates with the skin thickness score [20, 30], these data suggest that a decrease in myofibroblast score could be a preclinical indicator of improvement of scleroderma skin fibrosis. Furthermore, Lafyatis and colleagues reported the presence of B cells in all but one skin specimen at baseline and a complete or nearly complete depletion of dermal B cells six months after administration of rituximab [20]. This suggests a biological effect on the skin after drug administration that could with new courses of the drug lead to a clinical skin improvement. In fact, we treated patients with a progressive cutaneous disease after conventional cyclophosphamide therapy. Moreover, we decided to re-treat two of our patients, because they presented a slower improvement of the skin score in the first six months of follow up and an earlier repopulation of B cells, similar to the data reported by the Lafyatis and colleagues, in which the majority of patients presented a precocious recovery of B cells between 6 and 12 months [20].

Interestingly, none of the SSc patients in the current study treated with anti-CD20 showed a progression of major end-organ involvement in a population with early diffuse disease that had a relatively high risk of organ complication. Parameters of internal organ involvement remained stable, but a further follow up in a more consistent group of patients is needed before drawing any conclusions.

The clear fall in IL-6 levels observed in our study is in agreement with findings obtained in a mouse model after B cell depletion [8]. This fall could be related, at least for the first stages, to the high dose of methylprednisolone used for the premedication in ours and the cohort of patients in the study by Smith and colleagues [19], but considering the follow times of evaluation (3 to 6 to 12 months) it has to be related to the rituximab treatment. This may suggest that IL-6 might contribute to the active phase of the disease. The decrease in IL-6 at the systemic levels could be the biological premise of the improvement in skin fibrosis. In fact, it has been previously reported that chronic IL-6 administration induces an increased synthesis of collagen in dermal fibroblasts [31] and in the liver [32]. Furthermore, IL-6 has been demonstrated to enhance resistance of lung fibroblasts to apoptosis, contributing to the fibrotic effect [33].

Immunohistochemistry clearly demonstrated the presence of T cells either in uninvolved or in involved skin, but B cells were seen only in some patients, as previously reported [5, 19]. These data suggest that the most relevant contribution of B cells comes from the systemic pool. In fact, it appears clear that the response of skin fibrosis to B cell depletion does not rely on the presence of B cells in the skin, because most of our treated patients had no B cells, but very likely depends upon the general contribution to the autoimmune derangement given by the B cell compartments in lymphoid organs. B cells, with their multiple mechanisms as antibody-producing cells, antigen-presenting cells and profibrotic and proinflammatory cytokines producing cells (IL-6, IL-4, transforming growth factor-β), seem to be of great impact in the development of fibrosis. Thus, their modulation could inhibit skin fibrosis, as reported in the scleroderma mouse model [8], but the data on BAFF levels need to be interpreted since, as observed in patients with Sjogren's syndrome [34] or rheumatoid arthritis [35], the levels went up after B cell depletion.

Our data suggest that anti-CD20 treatment is well tolerated and that dSSc patients experience an improvement of the skin score and of clinical symptoms. The clear fall in IL-6 levels may contribute to the skin fibrosis improvement, while the presence of B cells in the skin seems to be irrelevant with respect to the outcome after B cell depletion. Although we cannot draw any conclusion due to the limited number of cases, the response in the early disease patients was striking suggesting that a trial is warranted to confirm these preliminary data.