Exacerbation of ulcerative colitis after rituximab salvage therapy

Abstract

Background

B-cells are considered to play a pathogenic role in human ulcerative colitis (UC) by producing autoantibodies that cause epithelial cell damage. Here we report on a patient with intractable UC who suffered from a severe exacerbation of UC after salvage therapy with rituximab, a B-cell-depleting anti-CD20-antibody.

Methods

A 58-year-old patient with active long-standing UC and unresponsiveness or adverse events to mesalamine, corticosteroids, azathioprine, methotrexate, infliximab, leukapheresis, mycophenolate mofetil, and adalimumab received 375 mg/m² rituximab.

Results

A severe exacerbation of UC activity was noted upon therapy that required hospitalization. Subsequent studies showed a complete depletion of CD20-positive mucosal B-cells associated with a suppression of local IL-10 production.

Conclusions

In contrast to rheumatoid arthritis patients, rituximab had deleterious effects in our UC patient by blocking IL-10 producing B-cells. Our data suggest an important anti- rather than proinflammatory role of B-cells in UC.

inflammatory bowel disease, ulcerative colitis, therapy, rituximab, CD20, interleukin-10

Insights into the pathogenesis of inflammatory bowel diseases (IBDs) have led to the introduction of targeted biological therapies into clinical practice. The majority of such novel therapies are directed against T-helper cell- or macrophage-derived proinflammatory cytokines.¹

IBDs seem to share some immunological features with rheumatoid arthritis (RA), as many immunomodulatory therapies (e.g., anti-TNF antibodies) appear to be effective in both disease entities. In RA, recent studies have addressed the role of B-cells in disease perpetuation. Small uncontrolled trials have suggested that B-cell depletion by rituximab, a chimeric monoclonal antibody against CD20, might have beneficial effects on clinical symptoms,² and these findings have recently been confirmed in a larger controlled trial.³ CD20 is expressed on the surface of pre-B and mature B-cells and is involved in B-cell activation, differentiation, and costimulation.

We report on the use of rituximab in a patient with refractory colitis despite extensive immunosuppressive therapy who experienced an exacerbation of disease after depletion of intestinal B-cells.

Case Report

Previous Medical History

In January 2005, a 58-year-old male Caucasian patient was admitted to the IBD outpatient clinic of the University of Mainz. The patient suffered from chronic bloody diarrhea (10 to 20 liquid stools per day) and intermittent abdominal cramping. Diarrhea had first started in 1986, and ulcerative pancolitis was diagnosed upon endoscopic and histopathologic examination. Since 1986 the patient had nonresponsiveness to treatment with high-dose corticosteroids (prednisolone up to 60 mg/d) and 5-aminosalicylic acid. Methotrexate had to be discontinued in 2002 due to the presence of alveolitis. Furthermore, azathioprine therapy was stopped after severe pancreatitis in 2003. Eight infusions of infliximab were administered in 2004 for treatment of colitis and enteropathy-associated arthropathy. However, infliximab therapy failed to ameliorate colitis activity, although a significant improvement of arthritic pain was noted.

The patient was then admitted to our outpatient clinic with the diagnosis of intractable steroid-refractory ulcerative colitis (UC) to undergo leukocytapheresis. The patient received a total of 6 leukocytapheresis courses with a Cellsorba column in 5 weeks, but no significant change in symptoms or endoscopic appearance was achieved. Nested PCR of colonic specimens after leukocytapheresis showed the presence of cytomegalovirus (CMV) DNA that had not been detected before leukocytapheresis. The patient was started on a 3-week course of valganciclovir 900 mg bid, although serum CMV early antigen was not present. Antiviral therapy did not influence the severity of diarrhea, although CMV DNA was cleared from the colonic mucosa. CMV was therefore not considered as a causative agent for the diarrhea. The patient was then put on immunosuppressive therapy with mycophenolat mofetil 1000 mg bid resulting in serum drug levels of 1.8–2 mg/L. Symptoms initially responded well but a severe relapse of colitis activity occurred after only 2 months. Adalimumab (160 mg, and 80 mg 2 weeks later) was then administered as a second anti-TNF-agent. No adverse events were recorded, but no improvement of clinical symptoms was noted. Since therapeutic strategies against granulocytes and T-cell-derived cytokines were not successful and the patient declined cyclosporin or proctocolectomy, anti-B-cell therapy was initiated.

Rituximab as Salvage Therapy

At this point the patient had 15 loose bowel movements per day. Differential blood count showed mild hypochromic anemia with iron deficiency (Hb 11.1 g/dL) and neutrophilia (81.9%; total leukocytes 6.1/nl). C-reactive protein (CRP) levels were only mildly elevated (11 mg/L). Rituximab (Mabthera, Roche, Switzerland) was then administered intravenously (375 mg/m²; total dose 750 mg), and no immediate adverse reactions were recorded.

However, several days after rituximab administration the patient was admitted with 30 loose, bloody stools per day, weight loss of 4 kg, and fever up to 39°C. Clinical examination was not indicative of a fever site outside of the intestine. CRP levels were 140 mg/L and differential blood count showed mild lymphopenia (16.4%) and monocytosis (9.9%), while total leukocytes were within the normal range (6.4/nl). No eosinophilia or basophilia were recorded. Colitis activity index according to Rachmilevitz was 13 as compared to 9 before rituximab application. In the serum, CD20-positive cells were 0.02% of total white blood cell (WBC), as evaluated by FACS analysis. Stool samples were repeatedly negative for Shigella, Salmonella, Yersinia, Campylobacter, and Clostridium difficile toxin A and B. Abdominal x-ray did not show toxic megacolon, signs of perforation or ileus. Chest x-ray was normal. Sigmoidoscopy up to 30 cm ab ano revealed a severe continuous colitis from anus to the sigmoid with spontaneous bleeding, ulcerations, enhanced granularity and swelling of the mucosa, and loss of haustration (Fig. 1). Histopathology demonstrated a very dense mucosal inflammatory infiltration of mononuclear cells (Fig. 1) and CD3 immunohistochemistry revealed high mucosal infiltration of T-cells after rituximab therapy (Fig. 2). Furthermore, immunohistochemistry for CD20 showed complete depletion of CD20-positive B-cells in the intestinal mucosa after rituximab therapy, while before rituximab therapy, B-cell constituted a significant proportion of intramucosal mononuclear cells (Fig. 2). In subsequent functional studies, lamina propria mononuclear cells (LPMC) were isolated from colonic specimens as described previously⁴ and cultured for 48 h. Supernatants from cell culture were analyzed by ELISA for IL-10 content. In spite of an excellent cell viability, IL-10 levels in the LPMC supernatant from our patient were 0 pg/mL in contrast to 110 pg/mL (±51 pg/mL SEM) in 22 control patients with ulcerative colitis (Fig. 2). IFN-γ levels were comparable (Fig. 2).

Figure 1.

Endoscopic findings before (A) and after (C) rituximab salvage therapy in our UC patient. A marked deterioration of UC activity was noted after therapy with the presence of prominent ulcers, edema, and spontaneous bleedings. Hematoxylin and eosin-staining of colonic biopsies after rituximab therapy showed dense mucosal infiltrations of mononuclear cells (D), which was even more pronounced as compared to before initiation of therapy (B). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com]

Figure 2.

Immunohistochemical analyses for B-cells before and after rituximab therapy (Clone L26, Dako, Glostrup, Denmark; dilution 1:25). Prior to rituximab therapy, a significant amount of lamina propria mononuclear cells were CD20-positive (A). After rituximab infusion, CD20+ cells could no longer be visualized (B), indicating that complete mucosal B-cell depletion was achieved. In contrast to LPMCs from 22 control patients with UC, isolated LPMCs from our patient produced no IL-10 after rituximab therapy (D). Immunohistochemistry with an anti-CD3 antibody (CI597R06, DCS, Hamburg, Germany; dilution 1:150) demonstrated numerous mucosal T-cells after therapy, however (C), and comparable levels of IFN-γ in the supernatant from LPMC in our patient and control patients with UC (E). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com]

The patient was put on therapy with 5-ASA, corticosteroids, and ciprofloxacin (400 mg bid IV). Fever was attributed to exacerbated colitis, since no other source was identified, and resolved after 2 days. At present, the patient continues to suffer from 10–15 loose stools per day and considers proctocolectomy. A second infusion of rituximab was not administered.

Discussion

Targeting of T-cell-derived proinflammatory cytokines such as TNF has significantly enlarged our therapeutic armamentarium in patients with both severe Crohn's disease⁵ and UC.⁶ It has also impressively supported the pathogenetic concept arising from animal models that IBDs are at least partially due to a dysregulated mucosal immune response.⁷ B-cells have only recently been appreciated as important players in this mucosal imbalance, although the mucosa of the intestine and exocrine glands harbors the largest amount of activated B-cells within the organism.⁸ Animal models of IBD that have addressed the role of B-cells yielded contradictory results: In the adoptive transfer model in SAMP1/YitFc mice, cotransfer of mesenteric B-cells with CD4+ T-cells aggravated intestinal inflammation in SCID recipients as compared to transfer of CD4+ T-cells alone, and B-cell number in mesenteric lymph nodes correlated with disease severity.⁹ The authors suggested that mesenteric lymph node-derived B-cells might inhibit the regulatory effects of CD25+ T-cells. In patients with IBD, an exaggerated B-cell response has been reported with antibodies directed against cytoplasmic proteins of bacteria that harbor within the intestinal lumen.¹⁰ This finding and other reports have led to the notion that the contribution of B-cells to the immunological dysregulation of the mucosal balance might be underestimated. The proinflammatory role of B-cells in chronic inflammatory diseases was further supported by a recent randomized trial on the use of rituximab in patients with RA.³ Clinical response rates were significantly higher in patients receiving rituximab than in control patients. Peripheral blood B-cells were significantly reduced in rituximab patients and correlated with the positive effect of rituximab.

In our patient with UC, rituximab therapy led to almost complete depletion of intestinal and systemic CD20+ B-cells, as shown by immunohistochemistry and FACS analysis, but did not ameliorate colitis activity. Instead, severe clinical aggravation was noted upon rituximab therapy, and no beneficial effects occurred at later stages. These findings are consistent with data from Mizoguchi et al¹¹ on a protective effect of regulatory B-cells in the TCRα^−/− model of colitis. This observation could be due to effects of B-cells on the mucosal clearance of apoptotic cells¹² or to B-cell-dependent regulation of CD4 T-cell activity.¹³ In fact, recent studies using the Gαi2^−/− mouse model showed an enhanced production of IL-10 by B-cells in response to lipopolysaccharide (LPS) stimulation that was required to induce regulatory T-cells. The protective action of such B-cells was dependent on the presence of Gαi2, which is involved in IL-10 regulation.^14,15

In our patient with UC unresponsive to anti-cytokine and anti-granulocyte therapy, exacerbation of colitis after rituximab therapy was paralleled by suppressed IL-10 levels and B-cell numbers. This patient has received extensive prior therapy, so the influence on B-cells by these regimens cannot be fully excluded. Nevertheless, our findings argue against a major proinflammatory contribution of B-cells in UC in this patient and rather suggest a model in which B-cells might be an important source for production of the antiinflammatory cytokine IL-10. Therefore, the case reported here does not support the application of rituximab as salvage therapy for UC.

Author notes

Reprints: Laboratory of Immunology, I. Medical Clinic, University of Mainz, Langenbeckstr. 1, 55131 Mainz, Germany Email: )

The first 2 authors contributed equally.