Infliximab for Crohn’s disease

Authors


SIRS, The treatment of steroid-refractory/dependent Crohn’s disease has greatly improved with the development of humanized monoclonal antibodies to the pro-inflammatory cytokine tumour necrosis factor-α (TNF-α). Recently, Nikolaus et al.1 and Cornillie et al.2 have assessed the clinical and immunological mechanisms of response and failure of treatment with the humanized anti-TNF-α antibody infliximab.

Nikolaus et al.1 followed 24 steroid-refractory active Crohn’s disease patients for 16 weeks after a single infusion of infliximab. The antibody was found to down-regulate the lipopolysaccharide-induced secretion of TNF-α by peripheral blood cells, and to decrease the nuclear concentration of NFκB transcription factor p65 in mucosal biopsies. Relapsers were characterized by an early rise in TNF-α production, before reactivation of the clinical symptoms. The authors ascribed these findings to the antibody-induced death of TNF-α-producing cells, and an early reactivation of the immune response was expected to be caused by the different genetic make-up of the non-responding patients. Cornillie et al.2 studied the pharmacokinetics of infliximab and its effects on the systemic and local immune functions in a large number of Crohn’s disease patients. They found no differences in the serum levels of infliximab between responders and non-responders, but showed that the antibody decreased the T-helper-1 cytokine production of lamina propria cells. They speculated that the latter observation was due to increased in situ lysis and reduced homing of these cells by infliximab.

The execution of the study by Nikolaus et al.1 has one major shortcoming, i.e. the use of a whole-blood assay to determine the TNF-α-producing capacity of cells, which has major consequences for the interpretation of their results. The enzyme-linked immunoabsorbent assay (ELISA) they used (R & D Systems, Minneapolis, USA) shows a major interference by infliximab, as established using similar culture and assay conditions, i.e. 5% CO2, 37 °C and 0.1 μg lipopolysaccharide/mL blood for 24 h, on blood samples from seven Crohn’s disease patients and two healthy controls.

The pre-infusion whole-blood median TNF-α secretion level of these Crohn’s disease patients was 381 pg/mL (range, 205–5056 pg/mL), and blood samples taken 2 h after the end of the infliximab infusion (5 mg/kg body weight) showed hardly any detectable TNF-α secretion: 10 pg/mL (range, 0–31 pg/mL). The addition of 75 μg/mL infliximab to four whole-blood samples of two controls, followed by 24 h lipopolysaccharide stimulation, was also found to almost completely inhibit TNF-α levels: median 0 pg/mL (0–28 pg/mL) compared to 3223 pg/mL (394–5945 pg/mL) without infliximab. The addition of autologous infliximab plasma to the pre-infliximab plasma, both after lipopolysaccharide stimulation, of the four individuals with the highest TNF-α level resulted in a 97.3% (96.3–98.6%) reduction from 3868 pg/mL (1537–5945 pg/mL) to 111 pg/mL (22–221 pg/mL). In addition, incubating plasma samples of the pre-infliximab lipopolysaccharide-stimulated whole blood of the patients and controls and a 750 pg/mL TNF-α ELISA standard with a similar infliximab concentration for 2 h in vitro resulted in undetectable TNF-α levels of all samples, except for the highest patient and control plasma, which decreased from 5056 to 31 pg/mL and 5945 to 70 pg/mL, respectively.

Finally, we assessed whether infliximab had lysed the TNF-α-producing cells by adding pre-infusion plasma to the whole-blood cell pellet of thoroughly washed (four times) blood obtained 2 h after the end of the infliximab infusion of the same patient, followed by 24 h lipopolysaccharide stimulation, in three patients. The addition of pre-infusion plasma to the infliximab-exposed cell pellet increased the TNF-α levels from 11 pg/mL (11–15 pg/mL), with infliximab, to a median of 245 pg/mL (174–280 pg/mL), which turned out to be 97.6% (56.7–127.9%) of the level (287 pg/mL; 136–432 pg/mL) obtained with washed pre-infusion whole-blood samples.

These experiments reveal that the presence of infliximab in plasma samples interferes in this ELISA determination of TNF-α, rendering it undetectable. In addition, (blood) cells after in vivo exposure to this antibody are not lysed and are still capable of normal TNF-α production and secretion; this was also noted by Cornillie et al.,2 in contrast to speculations by both groups. We believe that an early reactivation of the immune response after infliximab treatment of Crohn’s disease patients, with blood cell TNF-α production and subsequent mucosal NFκB p65 activation as read-out parameters, is not a genetically determined immune mechanism, but merely a consequence of differences in the clearance of infliximab from the circulation, and thus a pharmacokinetic phenomenon. How then does this reconcile with the observation of Cornillie et al.2 that there is no difference over time in the infliximab levels between responders and non-responders? Most probably, the answer lies in the observations by Plevy et al.,3 which reveal that responders show a decrease in TNF-α and interferon-γ production by lamina propria cells, whereas non-responders show an increased production of these cytokines. Cornillie et al.2 have a unique opportunity to assess similar differences in their large number of data on inflamed and non-inflamed cytokine-producing lamina propria cells and to relate these to the infliximab levels. These results would provide a strong indication as to whether the response and failure mechanisms of infliximab are related to the homing of T-helper-1 cells to the mucosa or to pharmacokinetics, or whether other genetically determined immunoregulatory and mucosal pharmacodynamic mechanisms are involved.

Ancillary