Fluorochrome‐labeled inhibitors of caspase‐1 require membrane permeabilization to efficiently access caspase‐1 in macrophages

Caspase-1 location in cells has been studied with fluorochrome-labeled inhibitors of caspase-1 (FLICA reagents). We report that FLICA reagents have limited cell-membrane permeability. This impacts experimental design as cells with intact membranes, including caspase-1 knockout cells, are not appropriate controls for cells with inflammasome-induced gasdermin D membrane pores.

pyrin domain that interacts with the initiator protein and is nucleated to form a filament [2], and a C-terminal caspase recruitment domain that both recruits procaspase-1, and mediates condensation of ASC filaments into a tight speck structure [3].Active caspase-1 cleaves and activates gasdermin D (GSDMD) which initiates pyroptotic cell death [4].The N-terminal cleavage fragment of GSDMD oligomerizes and forms pores in the plasma membrane with an inner diameter of ∼215 Å [5].Accumulation of GSDMD pores in the plasma membrane results in the release of mature IL-1β and IL-18, and ultimately cell lysis.
The first FLICA reagents were carboxyfluorescein (FAM) derivatives of caspase inhibitors, such as the pan-caspase inhibitor benzyloxycarbonyl-valine-alanine-aspartic acidfluoromethylketone (zVAD-FMK) that were developed to enable the detection of caspase activity in living cells for the study of apoptosis [6,7].FAM-VAD-FMK was reported to enter the cell and bind irreversibly to activated caspases [6].The peptide moiety of FLICA constructs conveys a degree of caspase specificity while the FMK leaving group facilitates covalent binding of the fluorescent inhibitor to the active site [7].Covalently bound FLICA reagent is retained in the cell after fixation and washing.FAM-YVAD-FMK and 660-YVAD-FMK are marketed as cellpermeable caspase-1-specific FLICA reagents.Both reagents use the YVAD tetrapeptide sequence based on the caspase-1 recognition site of IL-1β and are O-methylated on the aspartic acid residue to improve cell permeability; they differ only in the fluorescent moiety.These FLICA reagents are used to quantify cells with active caspase-1 and to assess the location of active caspase-1 within the cell.Our results suggest that FAM-YVAD-FMK and 660-YVAD-FMK are poorly permeable and require GSDMD pores to efficiently enter cells and bind active caspase-1.
Following inflammasome activation, FLICA-labeled caspase-1 co-localizes with ASC specks [8].We recreated a typical experiment that uses a FLICA reagent to visualize active caspase-1 in cells with an active NLRP3 inflammasome.LPS-primed WT bone marrow-derived macrophages (BMM) were incubated with FAM-YVAD-FMK prior to treatment with nigericin to trigger the NLRP3 inflammasome (Fig. 1A).Cells were fixed after 30 min and immunolabeled for ASC.We observed FLICA puncta that colocalized with the ASC specks (Fig. 1A, short exposure).However, Gsdmd −/− BMM, which do express and activate caspase-1 [4], lacked FLICA staining at the exposure used for the WT cells.ASCassociated FLICA puncta could be seen in Gsdmd −/− BMM using a longer exposure, although this gave a much stronger saturated signal in the WT cells (Fig. 1A, long exposure).This problem was not restricted to the FAM-YVAD-FMK.At short exposure, a signal from 660-YVAD-FMK could be seen in WT but not in Gsdmd −/− or Casp1 −/− cells (Fig. 1B).At much longer exposure, a signal could be detected in Gsdmd −/− but not Casp1 −/− cells.However, this exposure with WT cells gave a completely saturated white image.Considering both the quantification of signal strength and the length of imaging exposure allows us to estimate that labeling was approximately 30-fold stronger in WT cells than Gsdmd −/− cells for 660-YVAD-FMK, and sevenfold stronger for FAM-YVAD-FMK.
To confirm the presence of active caspase-1, the cleavage of caspase-1 and IL-1β was assessed in WT and Gsdmd −/− BMM (Fig. 2A).Caspase-1 p33, the active form of caspase-1 [8], its further cleaved inactivated product p20 [8], and cleaved IL-1β were present in Gsdmd −/− BMM at a similar amount as in WT BMM.This, along with the Gsdmd-dependence of readily detected FLICA signal, suggests that 660-YVAD-FMK and FAM-YVAD-FMK require GSDMD pores to efficiently enter the cell to bind caspase-1.Initial claims that FLICA reagents are cell-permeable were based on the ability of non-fluorochrome-containing inhibitors to enter live cells and inhibit apoptosis [7].However, early work showing inhibition of apoptosis by FAM-VAD-FMK was not reproducible [6,9].
We also tested three commonly used nonfluorescent inhibitors for their activity on Gsdmd −/− BMM.zVAD-FMK is a pan-caspase inhibitor, VX-765 is a prodrug of VRT-043198, an inhibitor of caspase-1 and caspase-4 [10] and MCC950 is an inhibitor of NLRP3 that directly interacts with the NLRP3 NACHT domain [11].These were each added to cells prior to NLRP3 inflammasome activation with nigericin.In WT cells, MCC950 prevented caspase-1 cleavage, and both caspase inhibitors trapped caspase-1 in the active p33 form (Fig. 2B).The latter is quanti- fied as an increase in caspase-1 p33 and decrease in caspase-1 p20 with inhibitors, compared with the nigericin alone sample.In Gsdmd −/− BMM, MCC950, and VX-765 showed similar inhibition as in WT cells, but zVAD-FMK only poorly inhibited caspase-1 processing to p20.Thus FMK-containing inhibitors may have limited cell permeability.Evidently, zVAD-FMK has sufficient permeability to inhibit cell death-dependent apoptotic caspases [12], but at 10 μM with 30 min preincubation, was not able to adequately penetrate an intact membrane to prevent rapid caspase-1 activation.
The permeability of FLICA reagents and the applicability of controls should be carefully considered when planning experiments and interpreting already existing data.Negative controls are important for FLICA staining experiments, as YVAD-FMK inhibitors cross-react with other cysteine proteases [13].Published protocols for FLICA staining of caspase-1 suggest untreated WT cells or activated caspase-1 knockout cells as negative controls [14,15].However, Casp1 −/− BMM do not have GSDMD pore formation and are thus not an appropriate control due to limited access of FLICA reagent to the cytosol.Indeed, any inflammasome-deficient controls where GSDMD activation is inhibited, or untreated cells, are inadequate controls for cells with normal GSDMD activation.For an experiment such as those shown in Fig. 1, Casp1 −/− BMM could be used as a negative control for staining of caspase-1 in Gsdmd −/− BMM, as these two cell types would be expected to have equal permeability to FLICA.However, the low permeability severely limits sensitivity, and will likely prevent analysis of the earliest sites of caspase-1 activation in the cell.While this study focused on caspase-1-specific FLICA reagents and GSDMD pores, the design of any experiment using FLICA reagents should consider the possibility of differential cell permeability to the reagent, particularly due to reduced pore formation as a result of genotype or treatment.

Figure 1 .
Figure 1.FLICA reagents are not readily cell-permeable in the absence of GSDMD.BMM from WT, Casp1 −/− , and Gsdmd −/− mice were LPS-primed for 3 h and treated with FAM-YVAD-FMK (A) or 660-YVAD-FMK (B) for 15 min prior to 30 min incubation with 10 μM nigericin.Cells were washed, fixed, immunostained for ASC, mounted on microscope slides, and imaged at short and long exposures optimized to see the FLICA signal in WT and Gsdmd −/− cells respectively.Scale bars = 10 μm.Arrows indicate co-localized FLICA puncta and ASC specks.The 660-YVAD-FMK signal was completely saturated in WT cells with long exposure and was removed from the Merge + DAPI image.The intensity of FLICA puncta that co-localized with ASC specks was quantified using ImageJ.Bars indicate the median value from 10 cells per sample represented by individual shapes.The phenomenon of reduced FLICA labeling in BMM from Gsdmd −/− mice has been observed in at least four independent experiments.ASC, apoptosisassociated speck-like protein containing a caspase recruitment domain; BMM, bone marrow-derived macrophages; FAM, carboxyfluorescein; FLICA, fluorochrome-labeled inhibitors of caspases; FMK, fluoromethylketone; YVAD, tyrosine-valine-alanine-aspartic acid.

Figure 2 .
Figure 2. Caspase-1 activation in Gsdmd −/− BMM is not inhibited by zVAD-FMK.(A) BMM from WT and Gsdmd −/− mice were LPS-primed for 4 h and then incubated with 10 μM nigericin for the indicated times.Released proteins and cell lysates were pooled.GAPDH levels were assessed as a loading control.Results are representative of two independent experiments.Raw images are shown in Supporting information Fig. S1.(B) Upper panel -BMM from WT and Gsdmd −/− mice were primed with 100 ng/mL LPS for 4 h, with 10 μM MCC950, VX-765, or zVAD-FMK added for the last 30 min.Cells were then exposed to 5 μM nigericin for 1 h.Released proteins and cell lysates were pooled.GAPDH levels were assessed as a loading control.The blots shown are representative of three independent experiments.Raw images are shown in Supporting information Fig. S2.Lower panel -Levels of caspase-1 p33 and p20 were quantified using ImageJ and shown relative to the WT with nigericin sample.Results from three different experiments are designated by different shapes and bars display the mean.*P < 0.05, ns = not significant, paired ratio t-test.BMM, bone marrow-derived macrophages; FMK, fluoromethylketone; zVAD, benzyloxycarbonyl-valine-alanine-aspartic acid.