Patients and treatment

Nine patients with progressive cutaneous SSc involvement, who showed a worsening of skin score higher than 10% after the conventional cyclophosphamide therapy [22] (up to 6 g), were treated with rituximab, two infusions of 1000 mg, two weeks apart, together with 100 mg methylprednisolone at each infusion, after three months of wash-out. All patients fulfilled the American College of Rheumatology classification criteria for scleroderma [23] and gave their informed consent to enter the study, which was approved by our Ethics Institutional Committee. All patients accepted that their biographical and clinical information could be eventually published.

Inclusion criteria were: age older than 18 years, a worsening in skin score higher than 10% after the conventional cyclophosphamide therapy, and a diffuse disease with trunk involvement. Exclusion criteria were: rest dyspnoea or signs and symptoms of heart failure, serious and uncontrolled coexisting diseases, infection, immunodeficiency or a history of tuberculosis contact, or cancer. None of the patients was taking corticosteroids daily. Three patients were re-treated with rituximab 1 g × 2 (days 1 to 15): the first patient because after 18 months she presented with a reactivation of her arthritis, while the other two patients were re-treated after 12 months because they presented a precocious and quicker B cell-recovery at months 3 and 7 (CD19 >4.5%).

The extent of skin involvement was evaluated by the Rodnan skin score, performed by two observers and their results averaged [25]. Every three months, activity index [26] and severity index were assessed [27] and Global Health Status (GH) and Health Assessment Questionnaire (HAQ) were administered to patients to evaluate the influence of the disease on daily functions. At the same time intervals, blood samples were collected to determine IL-6 and BAFF levels and to count CD19-positive cells by flow cytometry.

Internal organ involvement

All nine patients underwent pulmonary function tests to define forced vital capacity (FVC) and diffusing capacity for carbon monoxide (DLCO) before treatment and every six months. High-resolution computed tomography (HRCT) was performed before treatment and every 12 months. Renal involvement was defined as a scleroderma crisis or the presence of proteinuria or elevation in creatinine serum level. Creatinine levels and urine analysis were performed every three months. Cardiac involvement was defined as the presence of conduction disturbance, left ventricular ejection fraction (LVEF) less than 50%, pulmonary artery systolic pressure (PASP) more than 35 mmHg or presence of myocarditis; electrocardiography (ECG) and echocardiography were performed at the beginning of the treatment and every six months. Gastrointestinal involvement was defined as the presence of gastro-esophageal reflux symptoms or the evidence of gastrointestinal motility disturbance by barium swallow performed before treatment.

Biological marker detection

Serum levels of IL-6 and BAFF (R&D Systems, Minneapolis, MN, USA) were measured using an ELISA, as described by the manufacturer. Erythrocyte sedimentation rate, total immunoglobulin (Ig) G, IgM and IgA were part of the routine clinical care of each patient. ANA were determined by indirect immunofluorescence using Hep-2 cells as substrates and autoantibodies specificities were further assessed by ELISA (Shield, Dundee, UK). Peripheral blood CD19-positive cell count was obtained by flow cytometry every three months.

Skin biopsies and immunohistochemical analysis

Skin biopsies were performed in seven patients, who gave their informed consent, before treatment and in the five patients that achieved 12 months of follow up from the beginning of anti-CD20 therapy. Four healthy controls gave their informed consent to undergo forearm skin biopsy. In dSSc patients, cutaneous specimens were taken from the distal forearm for the clinically involved skin and from the buttock for clinically uninvolved skin. The biopsies were fixed into 10% formalin for two hours followed by paraffin inclusion for histological and immunohistochemical analysis.

Immunohistochemistry was carried out on 5 μm thick sections on polylysine-coated slides. After routine deparaffinization and rehydration, antigen retrieval was performed. Slide-mounted sections were heated in a microwave oven at 700 watt twice for four minutes in 10 mmol/L sodium citrate buffer (pH 6.0). Tissue sections were allowed to cool at room temperature (RT). Quenching of endogenous peroxidase activity was performed with Tris-buffered saline (TBS; pH 7.6) containing 2% hydrogen peroxide for 10 minutes at RT. Blocking was performed with 20% normal goat serum in TBS for 60 minutes at RT.

The sections were incubated with anti-CD3 and anti-CD20 mouse monoclonal antibodies (mAbs, Clone PS1 and L26 respectively; IgG_2a; Ylem, Rome, Italy) both 1:100 diluted in blocking solution (20% normal goat serum in TBS) for 60 minutes at RT. Then, the Super Picture Polymer detection kit (Zymed Laboratories, South San Francisco, CA, USA) was used for 30 minutes at RT. The chromogenic reaction was developed with 3,3'-diaminobenzidine tetrahydrochloride solution (Zymed Laboratories, South San Francisco, CA, USA). The nuclei were lightly counterstained with Mayer's hematoxylin. Negative controls without primary antibodies were performed for all reactions. As all mAbs were of IgG_2a isotype, mouse mAb IgG_2a served as an isotype-specific control. Human tonsil specimens were used as positive controls for both antibodies. All controls were run under the same conditions and the same IgG concentrations were used for the respective primary antibodies. Positive cells were counted by two independent observers in six randomly selected fields (total area: 7.38 mm²) for each section at × 400 magnification. Differences between observers about staining evaluation were resolved by consensus. The total number of positive cells was calculated.

Statistical analysis

All analyses were carried out using SPSS 15.0 (Chicago, IL, USA). Categorical variables were expressed as numbers, and quantitative variables as mean ± SD if normally distributed, and as median plus range if not. Non-normally distributed data were compared using the Mann-Whitney's test, and the Wilcoxon's test for paired data. A value of P < 0.05 was considered statistically significant.

Skin score, activity and severity indices

All nine SSc patients treated with rituximab experienced an improvement of the skin score, activity index, severity index, HAQ and GH during the follow up if compared to pre-treatment values (Table 2 and Figure 1). Neither infections nor infusion reactions were observed. The only serious adverse event was the development of an occult breast cancer, which was thought to be unrelated to the study medication. The mean follow up was 16.7 ± 12.6 months: all patients reached a six-month follow up, five patients reached a 12-month follow up, four patients reached an 18-month follow up, three patients reached a 24-month follow up and two patients reached a 36-month follow up.

Interestingly, in all nine patients treated with rituximab, the skin score improved gradually over time (Figure 1 and Table 2). After six months, the skin score improved in all the patients, decreasing from 21.1 ± 9.0 to 12.0 ± 6.1 (P = 0.001), with a median of improvement of 43.3% (range: 21.1 to 64.0%). Considering the last observation carried forward in each patient, the median skin improvement was 57.1% (range: 21.2 to 76.2).

After six months, the activity index decreased from 4.8 ± 1.3 to 1.2 ± 1.2 (P = 0.01) and the severity index from 10.5 ± 3.2 to 7.2 ± 2.8 (P = 0.01; Figure 1 and Table 2). All patients reported an improvement of their conditions as supported by the decrease in HAQ from 0.9 ± 0.7 to 0.4 ± 0.5 (P = 0.01) and an increase in GH from 59.4 ± 20.9 to 82.8 ± 16.6 (P = 0.01; Table 2). The only patient who did not present an improvement of the activity and severity indices, HAQ and GH, had a long disease duration.

Organ involvement

The FVC and DLCO values showed no significant differences at follow up (96.8 ± 18.9% and 58.4 ± 14.2% of predicted value, respectively) compared with baseline (91.6 ± 20.7% and 58.0 ± 15.8% of the predicted value, respectively; P = ns for both comparison). Four (44.4%) patients presented an improvement higher than 10% of FVC, (median increase 14.9% (range: 11.8% to 29.5%)). None of the patients presented a reduction in FVC considered clinically significant (>10%), but one patient showed a decrease in FVC values suggesting a trend to a progression of her restrictive lung disease [28, 29].

None of the patients showed signs of new or progressive cardiac disease, with stable ejection fractions and no modification on ECGs; none of the patients experienced renal crisis or symptoms suggesting progressive gastrointestinal disease.

Biological markers

At baseline, patients presented high levels of IL-6 (3.7 ± 5.3 pg/ml), that permanently decreased after six months (0.6 ± 0.9 pg/ml, P = 0.02; Table 2 and Figure 2a). Three months after the rituximab infusion, circulating B cells evaluated by flow-cytometry were depleted (peripheral blood CD19 <0.1%) in all but one patient, and between 6 and 12 months they begun to repopulate. Upon B-cell depletion, BAFF levels increased relative to baseline (baseline: 1233.5 ± 683.3 pg/ml vs six months: 3257.8 ± 1571.8 pg/ml), while in one patient the BAFF levels did not increase until the time of repopulation (Figure 2b).

The autoantibody titers and IgG and IgA levels did not vary over the study period, while IgM levels decreased from 133.7 ± 21.7 mg/dl to 90.9 ± 28.9 mg/dl at six months follow up (P = 0.008), and to 83.0 ± 18.7 after 12 months of follow up (P = 0.04; Table 2).

Before rituximab treatment, one patient presented myositis with high creatine kinase levels, which decreased significantly after anti-CD20 treatment (data not shown). Creatine kinase levels remained within the normal range during the 36 months of follow up. The only patient with a long disease duration who did present the less significant clinical improvement, showed an important amelioration of her arthritis, with a change of disease activity score (DAS) from 4.3 to 2.0.

Three (42.9%) out of the seven patients, who underwent skin biopsies before treatment, presented CD20-positive cells on biopsies of the clinically involved skin and uninvolved skin; only one patient of these three repeated the biopsy after 12 months and it showed a depletion of dermal B cells. The other two patients were treated only for six months and they did not agree to a repeat biopsy.

CD3 lymphocytes were found, predominantly, in a perivascular location in the mid and deeper portion of the dermis in all the involved and uninvolved skin biopsies of patients before and after treatment with anti-CD20. Figure 3 illustrates the presence of B cells (a) and T cells (b) in forearm biopsy in patient number three before therapy. The mean number of CD3-positive cells in skin biopsies of four healthy subjects was 8.0 ± 2.0 and none presented B cells (data not shown). Before treatment, the mean number of CD3-positive cells was 54.7 ± 27.9 in involved skin and 65.6 ± 39.7 in uninvolved skin (P = ns; Figure 3). After treatment, a similar number of CD3-positive cells was found in involved skin (44.3 ± 24.0) and in uninvolved skin (62.7 ± 23.4) of the five patients who underwent skin biopsies after 12 months.