Low bead counts were more common with the VersaMAP kit in our hands (> 90% of samples on some runs and up to 1 in 3 standard/control wells). In contrast for the Bio-Plex and MILLIPLEX kits, low bead counts were not observed in any selleck monoclonal humanized antibody standard/control wells and in 11% and 1% of samples respectively. This may have been a result of greater median bead aggregation observed with this type of sample for the VersaMAP kit than for the Bio-Plex and MILLIPLEX kits (29% vs 11% and 12% respectively). Even though each kit performed as specified and intended by the manufacturers, our aim was to quantify low concentrations of both IL-17

and IFNγ in tissue samples. Given our findings for sensitivity, standard curves and technical performance, only the Bio-Plex and MILLIPLEX kits were evaluated further. Spiked cytokine recovery was used to measure the ability of each kit to accurately quantify recombinant cytokines in tissue homogenates.

Nine biopsies each from three patients were individually prepared by manual disruption in extraction buffer (A). Supernatants from each patient were combined and split into aliquots. For each set of aliquots from a single patient, one PCI32765 was spiked with extraction buffer alone (“unspiked”) and two were spiked with known concentrations of both recombinant human IL-17 and IFNγ. Therefore we evaluated the ability of each of the kits to accurately measure cytokine spikes in mucosal tissue homogenates at lower and higher concentrations (1.5, 6, 50, 100 and 1000 pg/mL; for range of standard curves see Table 1). Observed IL-17 values were lower than expected for both the Bio-Plex kit (≥ 6 pg/mL: 38% ± 8% [mean ± SD], 29–47% [range]) and the MILLIPLEX

kit (≥ 6 pg/mL: 36% ± 12%, 21–49%) Thymidine kinase — see Fig. 1A. Neither kit adequately measured IL-17 spike recovery at 1.5 pg/mL. The background levels in unspiked samples from the three patients were 0.0, 0.0 and 1.8 pg/mL for the Bio-Plex kit and slightly higher at 0.0, 2.4 and 2.5 pg/mL for the MILLIPLEX kit. The IFNγ spikes were recovered with generally lower than expected accuracy using the MILLIPLEX kit (≥ 50 pg/mL: 32% ± 12%, 19–42%) and overall with higher than expected accuracy with the Bio-Plex kit (≥ 50 pg/mL: 218% ± 235%, 57–487%) — see Fig. 1B. Neither kit adequately measured IFNγ spike recovery at 1.5 pg/mL and only the MILLIPLEX kit performed as expected at 6 pg/mL (121%). High levels of IFNγ background were detected in the unspiked samples using the Bio-Plex kit (49.2, 264.0 and 1193.7 pg/mL) compared with background levels of 0.3, 4.5 and 6.7 pg/mL with the MILLIPLEX kit. Note that a control containing only the RPMI-1640 and FCS extraction buffer (A) yielded an IFNγ reading of 1177.7 pg/mL with the Bio-Plex kit compared with 0.0 pg/mL for the PBS-based extraction buffers (B) and (C).