|Home | About | Journals | Submit | Contact Us | Français|
Missense mutations of NLRP3 gene (CIAS1) are associated with autoinflammatory disorders characterized with excessive production of IL-1β. Here we analyzed the immune responses of knock-in mice carrying a point mutation of NLRP3 associated with Muckle-Wells Syndrome. We found that antigen presenting cells (APCs) from such mice produce massive amounts of IL-1β and IL-18 upon stimulation with TLR ligands in the absence of ATP. This is likely due to a diminished inflammasome activation threshold that allows a response to the small amount of TLR ligand entering the cell without ATP pulse. Moreover, the NLRP3 knock-in (KI) mice exhibited spontaneous and contactant-induced skin inflammation characterized by neutrophil infiltration and Th17-dominant response, which was originated from hematopoietic cells. The inflammation of KI mice was resulted from excess IL-1β production from APCs which augments Th17 differentitation. These results demonstrate that NLRP3 mutation leads to inflammasome hyper-activation and consequently Th17-dominant infammation in autoinflammatory diseases.
The NLRP3 (NLR family, pyrin domain containing 3, also called NALP3 or Cryopyrin)(Ting et al., 2008) inflammasome is an intra-cytoplasmic protein complex consisting of NLRP3, ASC and pro-caspase-1. Upon activation by microbial- or danger-associated molecular patterns this inflammasome converts pro-caspase-1 to active caspase-1, which then cleaves pro-IL-1β and pro-IL-18 to form mature IL-1β and IL-18, respectively (Petrilli et al., 2007a; Yu and Finlay, 2008).
Although normal activation of the NLRP3 inflammasome contributes to host-defense, excessive activation leads to inflammatory diseases. This is seen in patients bearing mutations in NLRP3 gene who exhibit autoinflammatory syndromes marked by skin, joint and eye inflammation as a result of greatly increased IL-1β production (Dinarello, 2007; Gattorno et al., 2007). Study of patients with these mutations or mice with similar mutations offer the opportunity to analyze the immunologic consequences of excessive inflammasome activation.
In the present study we generated a mouse strain expressing a R258W mutation (corresponding to the R260W mutation in humans) in the NLRP3 gene associated with one of the autoinflammatory syndromes, Muckle-Wells syndrome (Neven et al., 2004). These NLRP3-mutated mice exhibited inflammasome hyperactivation and developed autoinflammatory disease similar to that in humans. Importantly, inflammation in such mice was dominated by Th17 cytokine responses.
To analyze the immunologic response of inflammasome activation, we created a mouse model of the human Muckle-Wells Syndrome by generating knock-in (KI) mice bearing the R258W mutation in NLRP3 gene characteristic of patients with this autoinflammatory disease (Figure S1). As expected, we found that cells of KI mice express normal amounts of NLRP3 under both resting and stimulated conditions (Figure S1). However, while resting bone marrow-derived macrophages (BMDM) from KI mice did not secrete detectable amounts of IL-1β or IL-18, they did secrete large amounts of these cytokines upon TLR stimulation, both in the presence and absence of exogenous ATP; this pattern differed from that of Wt BMDM, which secreted substantial amounts of IL-1β or IL-18 in the presence but not in the absence of ATP upon TLR stimulation (Figures 1A and S2A). A similar result was obtained with bone marrow-derived dendritic cells (BMDC) and splenic CD11b+ cells (Figures S2B and S2C).
In further studies involving mRNA analysis, it's revealed that transcription of pro-IL-1β was comparable between LPS-stimulated KI and Wt macrophages (data not shown). However, in immunoblotting studies, we found that pro-IL-1β processing and mature IL-1β production from KI BMDM took place upon LPS stimulation both in the presence and absence of ATP, while that from Wt cells occured only with an ATP pulse (Figure S2D). Similarly, mature caspase-1 P20 and P10 were generated in stimulated KI cells when cultured in the presence or absence of ATP, whereas Wt BMDM produced these active fragments only in the presence of exogenous ATP (Figure 1B). These results were confirmed by anti-caspase-1 P10 immunoprecipitation studies which showed that stimulated BMDM from KI but not Wt mice secreted caspase-1 P10 (Figure S2E).
Finally, in contrast to production of IL-1β and IL-18, NLRP3 inflammasome-independent pro-inflammatory cytokine secretion (i.e., IL-12, IL-6, TNF-α and IL-10 secretion) by KI and Wt cells was entirely equivalent (Figure S3).
A possible mechanism of ATP-independent inflammasome activation of NLRP3 knock-in cells arises from recent work showing that NLRP3 inflammasome activation is dependent on intracellular potassium (K+) depletion triggered by exogenous ATP (Petrilli et al., 2007b). We therefore tested whether LPS-stimulated KI macrophages exhibited K+ efflux in the absence of ATP. While 50mM KCl incubation (sufficient to inhibit potassium efflux without affecting cell viability, data not shown) completely abolished secretion of IL-1β by Wt macrophages stimulated by LPS in the presence of ATP, it had no effect on LPS-stimulated KI macrophages (Figure 1C). Therefore, production of IL-1β from KI macrophage in the absence of ATP was not due to spontaneous potassium efflux.
Other possible mechanisms of ATP-independent inflammasome activation of NLRP3 knock-in cells is that the mutation affects the amount of endogenous ATP produced by cells and/or the ability of ATP to affect PAMPs (pathogen associated molecular patterns) entry into cells via pore formation (Kanneganti et al., 2007; Piccini et al., 2008). In initial studies we found that extracellular ATP concentration measured in culture fluids of KI and Wt macrophages at various times after LPS stimulation and LPS entry to KI or Wt macrophages were essentially equivalent (data not shown). Thus, the mutation does not confer ATP-independence by affecting the production of extra-cellular endogenous ATP.
To study the effect of the mutation on ATP-dependent pore formation we next determined the effect of a pannexin-1 (Panx1) inhibiting peptide (10panx1) on IL-1β secretion by Wt and KI macrophages. We found that incubation of LPS-stimulated Wt macrophages with either a low (1mM) or a high (4mM) concentration of 10panx1 for 30 minutes to 24 hours prior to an ATP pulse led to clearly reduced IL-1β production as compared to macrophages subjected to ATP addition in the absence of 10panx1 (Figure 1D). The reduced (but not absent) IL-1β production by Wt cells in the face of Panx1 inhibition is likely to be due to residual channel function or non-channel cellular entry of LPS that is still dependent on ATP. In contrast, incubation of LPS-stimulated KI macrophages with 10panx1 for 30 min to 4 hours prior to adding ATP had no effect on IL-1β production and only pre-incubation with 10panx1 for 24 hours led to reduced IL-1β production (Figure 1D); in addition, at all time points the effect of Panx1 inhibition on IL-1β secretion (or lack thereof) was independent of the presence of exogenous ATP, even at the 24 hour point that led to reduced IL-1β secretion. The latter was not due to inhibitor toxicity since IL-6 secretion was not affected by pre-incubation for any length of time (Figure S4A).
Instead, it is again likely to be due to residual activity of the pannexin-1 channel or non-channel entry which in the case of KI cells (but not Wt cells) can be mediated by the low amounts of endogenous ATP released upon LPS stimulation. This is supported by the finding that KI macrophages stimulated with LPS (in the absence of exogenous ATP) in the presence of an ATP-hydrolysing enzyme, apyrase or antagonists of P2X7 receptors such as KN-62 or 2′, 3′-O-(2,4,6-trinitrophenyl)-ATP (TNP-ATP) (Atarashi et al., 2008; Piccini et al., 2008) exhibit reduced IL-1β secretion, suggesting that endogenous ATP does indeed play a role in entry of PAMPs into KI macrophages (Figure S4B). Overall, we conclude that since inhibition of the Panx1 channel does not change the independence of KI cells from exogenous ATP, the latter is not due to an effect of the mutation on Panx1 mediated PAMPs entry into the cell. In addition, since inflammasome activation is decreased upon maximal pannexin-1 inhibition in KI cells, it is likely that this mechanism does play a role in inflammasome activation in these cells, possibly via Panx1 activation and entry of small amounts of PAMPs induced by low levels of endogenous ATP.
The above results suggested that cellular entry of low amount of PAMPs was sufficient to activate the KI inflammasome but not the Wt inflammasome, consistent with recent work suggesting that the mutation might have resulted in a conformation change in NLRP3 (Aksentijevich et al., 2007) that lowers its activation threshold. To test this possibility, we determined the ability of various concentrations of LPS to stimulate BMDMs from KI or Wt mice to produce IL-1β. As expected, LPS-stimulated KI BMDM produced substantial amounts of IL-1β in the absence of ATP even at a low LPS dose (Figure 1E). In contrast, in the absence of ATP, Wt cells produced IL-1β only when stimulated with high concentrations of LPS (≥500 ng/ml) and the amount of IL-1β produced by Wt BMDMs at these concentrations was much less than that produced by KI BMDMs (Figure 1E). A similar result was obtained with an immunoblotting study, wherein in the absence of ATP, KI cells secreted mature IL-1β upon stimulation with either high (5ug/ml) or low (0.1ug/ml) amount of LPS, whereas Wt cells responded only to high amounts of LPS, with the secretion of relatively low amount of IL-1β (lane 1) (Figure S4C). In addition, caspase-1 immunoblotting studies also demonstrated that LPS alone either at high or low concentration was sufficient to trigger caspase-1 activation in KI macrophages in the presence or absence of ATP, while in the case of Wt cells, low concentration of LPS could not induce caspase-1 activation unless ATP was added, even though high amount of LPS was able to activate caspase-1 from Wt cells in the absence of ATP (lane 1) (Figure 1F). We therefore conclude that in the absence of exogenous ATP, the inflammasome of KI cells respond to low amounts of ligand that cannot activate the inflammasome of Wt cells due to a lower threshold of NLRP3 inflammasome activation.
Most knock-in mice raised under pathogen-free conditions displayed decreased linear growth and weight gain as well as hair loss (Figure S5A, S5B and data not shown). Some of the KI mice died before weaning and only few of those that matured generated offspring. This lack of reproductive capacity was associated with development of dermatitis affecting the ears, top of the head and/or the tail base areas at 8 to 16 weeks of age (Figure 2A) that was accompanied by the development of enlarged spleen, cervical, axillary and/or inguinal nodes draining areas of dermatitis as well as hepatomegaly (Figure S5C).
Histologic study of the affected skin of KI mice showed marked thickening and heavy neutrophil infiltration in both epidermis and dermis (Figure 2B). In addition, the lymph nodes showed loss of germinal center architecture, poorly developed follicles, expanded interfollicular regions and a diffuse neutrophil infiltration, whereas the spleen contained expanded red pulp areas filled with neutrophils and islands of tri-lineage hematopoietic cells (Figure 2C). Finally, the liver of KI mice with skin inflammation exhibited a marked peri-ductal leukocyte infiltration (Figure 2D). Flow cytometric study of whole cell populations from lymph node and spleen tissues with neutrophil specific antibodies (7/4Ag and Ly6G) revealed that spleen and to a lesser extent lymph node from inflamed KI mice contained increased numbers of neutrophils (Figure 2E); similarly, staining of tissues with anti-Gr-1 and anti-F4/80 showed that KI spleen and to a lesser extent lymph node and liver contained increased numbers of granulocytes and macrophages (Figure S5D-F). These observations indicate that the spontaneous skin inflammation and its associated lymphoid component is characterized by a prominent “acute” inflammatory infiltration of neutrophil, granulocytes and macrophages.
Complete blood cell counts of inflamed KI mice were elevated and thus compatible with the presence of inflammation. As KI mice without inflammation have comparable number of cells with Wt control, the myeloid cell number increase was most likely secondary to the inflammatory status of KI mice (Table S1). In addition, autoantibody levels (anti-dsDNA) was increased in the blood of KI mice as compared with Wt mice, but the level of increase did not correlate with the extent of inflammation: young KI mice without skin inflmmation showed similar levels of anti-dsDNA as KI mice with severe inflammation (Figure S6). Interestingly, kidney, lung and gut of KI mice were free of inflammation; serum inflammatory cytokine levels of KI mice was equivalent to those of Wt mice and body temperature of inflamed KI mice was almost comparable to that of Wt control (data not shown).
To understand further the spontaneous inflammation occurring in NLRP3-mutated KI mice, we determined the cytokine profile of skin tissues from KI and Wt mice by real-time RT-PCR. Normal (uninflamed) skin of KI mice contained low levels of transcript for cytokines and factors such as IL-12p40, IL-1β, IL-6, IL-17A, IFNγ, RORγt and T-bet which were equivalent to that in Wt skin (data not shown). In contrast, in the skin tissue of inflamed KI mice, transcript levels of pro-inflammatory cytokines such as IL-12p40, IL-12p35, IL-23p19, IL-1β, IL-6 and TNF-α were significantly higher than in Wt skin (Figure 3A); in addition, expression of various Th17-related cytokines and factors such as IL-17A, IL-17F, IL-21, TGF-β1, IL-23 receptor (IL-23R), RORγt and IL-22, (the latter a Th17 cytokine involved in psoriatic skin inflammation) (Zheng et al., 2007), were all substantially increased (Figure 3A). However, the inflamed skin of KI mice manifested a poor Th1 response: the expression level of KI tissue IFNγ was much lower than IL-17A or even undetectable in some experiments, albeit on average still higher than IFNγ in Wt tissue. In addition, it should also be noted that while tissue IL-12Rβ1 levels of KI and Wt mice were similar, expression of T-bet and IL-12Rβ2 were significantly decreased in KI lesional tissue, and the decrease of IL-12Rβ2 was even more significant than that of T-bet (Figure 3A). Finally, although the IL-4 expression was increased in inflammed skin of KI mice, that of IL-5 and GATA-3 was not significantly elevated than in Wt skin, suggesting that a typical Th2 response was not present and the increased amount of IL-4 in KI tissue was originated most likely from mast cells or basophils rather than T cells (Pouliot et al., 2005). Overall, these data indicate that the spontaneous skin inflammation of KI mice is associated with a Th17-skewed cytokine response.
The Th17 dominant cytokine profile in inflamed skin tissue of NLRP3-knock-in mice prompted us to examine responses of T cells from their lymphoid tissues. We found that CD4+ T cells isolated from lymph nodes and spleens of “inflamed” KI mice (mice with active skin lesions) exhibited an activated phenotype (CD25hi, CD69hi, CD44hi, and CD62Llo), (Figure S7) and produced considerably more IL-17 and to a lesser extent IFNγ and IL-4 than Wt cells (Figures 3B and S8). These data correlated with the fact that RORγt expressison was significantly upregulated in KI CD4+ T cells while T-bet and GATA-3 expression was only slightly increased as compared to Wt cells (Figure 3C). Thus, CD4+ T cells from inflamed KI mice were pre-activated in vivo and exhibited a Th17 phenotype.
Previous studies revealed that NLRP3 inflammasome was activated during DTH, a T cell mediated cellular immune response to repeated epicutaneous exposure to contact antigens (Sutterwala et al., 2006; Watanabe et al., 2007). In accord with those results, we found that in 1-chloro-2, 4-dinitrobenzene (DNCB)-induced contact hypersensitivity, young KI mice without spontaneous inflammation showed greater DTH responses than Wt mice characterized by greater ear swelling, thickening of epidermis and dermis, as well as heavier tissue infiltration of neutrophils and monocytes (Figure 4A and data not shown). Importantly, the cytokine profile in this induced inflammation was similar to the spontaneous dermatitis developed in KI mice. In particular, mRNA encoding Th17-related cytokines and factors such as IL-17A, IL-17F, IL-23p19, IL-23R, RORγt and IL-22 were all elevated in skin tissue of KI mice as compared to that of Wt mice that did develop DTH inflammation (Figure 4B). In contrast, transcription of Th1-related cytokines and factors such as IFNγ, IL12Rβ1, T-bet and IL-12p40 in DTH tissues of KI and Wt mice were equivalent. In addition, although IL-12p35 expression was higher in KI tissue, its receptor, IL-12Rβ2 was strongly down-regulated (Figure 4B). Finally, expression of Th2-related cytokines and factors, IL-4, IL-5 and GATA-3, was weaker in DTH tissue of KI mice than in Wt mice.
In view of the predominance of the Th17 response in the DTH inflammation in KI mice, we also determined levels of Th17-inducing cytokines. It is noteworthy that although IL-1β and to a lesser extent TGF-β1 and TNF-α were elevated in KI mice as compared to Wt mice, IL-6 and IL-21 levels were almost equal (Figure 4B). These data suggest that IL-1β was in fact the main factor resulting in the elevated IL-17 production in the contact hypersensitivity from KI mice. Finally, CD4+ T cells in lymph nodes draining the DTH site of KI mice produced twice as much IL-17 than controls but secreted similar amount of IFNγ (Figure 4C). Thus, consistent with spontaneous inflammation, KI CD4+ T cells were also skewed to Th17 differentiation in DNCB-induced inflammation.
The above results suggested that the inflammation of KI mice was most likely due to hematopoietic cells. To confirm this possibility, we created bone marrow chimeric mice by transfer of bone marrow cells from either KI or Wt mice to lethally irradiated Wt recipient mice. Eight weeks after cell transfer, BMDMs from KI-Wt chimeric mice produced IL-1β upon stimulation with various TLR ligands stimulation in the absence of ATP as did KI BMDMs in earlier experiments (Figures 5A-B and and1A).1A). This correlated with the fact that 40% of KI-Wt chimeric mice (but none of the WT-KI mice) developed skin inflammation similar to that observed in the KI mice mentioned above and consisting of dermatitis affecting the ears or the tail base areas (Figure S9). This was accompanied by the development of enlarged lymph nodes draining the area of dermatitis as well as splenomegaly and hepatomegaly (Figure S9A-B). As in the KI mice, the affected skin of KI-WT chimeric mice exhibited marked thickening of epidermis and dermis accompanied by a heavy infiltration of neutrophils; in addition, the liver, spleen and lymph node architecture of these mice was similar to that in the KI mice (Figure S9C-E). Finally, cytokine analysis by RT-PCR revealed that lesional skin and draining lymph nodes of KI-Wt chimeric mice contained cell populations producing elevated levels of IL-17A and IL-1β, whereas production of IFNγ was comparable to that in control Wt-Wt chimeric mice (Figure 5C-D). These results thus provide strong evidence that the inflammation of KI mice was originated from hematopoietic cells.
In further studies we focused on the cellular basis of the Th17 bias of the inflammation in NLRP3-mutated KI mice. In initial studies to determine whether the Th17 dominant phenotype of CD4+ T cells from KI mice was due to an intrinsic abnormality in T cell differentiation, we determined cytokine production in cultured CD4+ T cells from KI or Wt mice stimulated in the absence of antigen presenting cells under Th1, Th2 and Th17 polarizing conditions. We found that similar amounts of IL-17, IFNγ or IL-4 was produced by KI and Wt cells under each condition (Figure S10) and thus concluded that the Th17 dominant phenotype in KI mice was most likely due to an abnormality in the antigen-presenting cells (APCs).
To test this latter possibility, we then determined cytokine production by anti-CD3/anti-CD28-stimulated Wt CD4+ T cells co-cultured with splenic CD11b+ cells from KI (inflamed) or Wt mice in the presence or absence of TLR ligands under Th17 polarizing conditions (TGF-β+IL-6). The percentage of IL-17 producing cells generated in cultures containing KI CD11b+ cells was substantially higher than in cultures with Wt CD11b+ cells (Figure 6A, left panel). Conversely, the percentage of cells producing IFN-γ was clearly lower (Figure 6A, left panel) and the ratio of IL-17/IFN-γ was significantly higher in cultures containing KI CD11b+ cells (Figure 6A, right panel). Finally, it should be noted that the same set of stimulated T cells co-cultured with splenic CD11b+ cells from either KI or Wt mice under Th0, Th1 and Th2 conditions, gave rise to very low numbers of IL-17 secreting cells. In addition, the number of IFNγ– and/or IL-4-producing cells was decreased in cultures of Wt T cells incubated with KI CD11b+ cells in comparison with T cells co-cultured with Wt CD11b+ cells (Figure S11). These results suggest that KI APCs on the one hand support Th17 differentiation following initial induction of the Th17 program by TGF-β and IL-6; and on the other hand inhibit other forms of T cell differentiation under lineage specific conditions.
In further studies we determined whether the above capacity of KI macrophages to enhance IL-17 responses was mediated by a secreted factor. Accordingly, Wt CD4+ T cells were cultured under Th17 conditions in the presence of supernatants derived from TLR ligand-stimulated and -unstimulated BMDMs. We found that T cells incubated with KI BMDM supernatants produced significantly higher amounts of IL-17 and lower amounts of IFNγ (Figure S12A). It was thus clear that soluble factor(s) produced by KI BMDMs can substitute for macrophages in the enhancement of Th17 responses.
In additional studies we determined if the enhancing effect of KI BMDM supernatants required stimulation of T cells under Th17 conditions. We found that under Th0 conditions (culture of cells in the absence of added cytokines), Wt CD4+ T cells cultured with supernatants from either KI or Wt BMDMs did not produce IL-17 (Figure S13A, upper 2 rows). On the other hand, culture of Wt CD4+ T cells with KI BMDM supernatants in the presence of TGF-β led to a greater IL-17 response than Wt BMDM supernatants under the same conditions, although the overall level of the IL-17 response was much lower compared to the response under full Th17 conditions (TGF-β+IL-6)(Figure S13A, lower 2 rows). The same experiments using BMDC instead of BMDM gave similar results (Figure S13B). These data made it clear that the enhancing effect of KI APC supernatants requires Th17 condition, e.g., the supernatants are acting on T cells that have undergone initial Th17 differentiation.
Previous work showing that IL-1β supports Th17 differentiation (Kryczek et al., 2007; Sutton et al., 2006) plus the above findings with macrophage supernatants led us to hypothesize that IL-1β produced by APCs from KI mice might be a soluble factor causing increased IL-17 production. To explore this possibility we co-cultured IL-1 receptor deficient (IL-1RI-/-) CD4+ T cells with KI CD11b+ cells under the same conditions as those described above for Wt CD4+ T cells. We found that the IL-17 production was substantially decreased compared with Wt CD4+ T cells under the same conditions, while IFN-γ secretion was increased (Figure 6B, left panel, upper 2 rows). This result is also reflected in the ratio of IL-17/IFN-γ producing cells, wherein the Th17 skewing is considerably reduced in cultures of IL1RI-/- CD4+ T cells (Figure 6B, upper right panel). Similar results were obtained in cultures of IL-1RI-deficient CD4+ T cells and knock-in BMDM supernatants (instead of CD11b+ cells) comparing to Wt CD4+ T cells under same condition (Figure S12B, left panel, upper 2 rows).
Finally, it should be noted that co-culture of KI CD11b+ cells with CD4+ T cells from IL1RI-/- mice did not result in a decrease in the induction of IL-17 (and increase in IFN-γ) to the level seen in co-cultures of Wt CD11b+ cells with IL-1RI-deficient CD4+ T cells (Figure 6B, left panel, the 2nd and 4th rows). In some contrast, culture of KI BMDM supernatants with IL-1RI-deficient CD4+ T cells did result in a decrease in IL-17 induction almost to the level seen with IL-1RI-deficient CD4+ T cells cultured with Wt BMDM supernatant under most conditions (Figure S12B, left panel, the 2nd and 4th rows). These results suggest that soluble IL-1β may not be the only factor involved in the shift toward IL-17 predominant T cell responses induced by KI APCs. This factor, however, is not one of the previous Th17 supporting cytokines since addition of IL-6, IL-21 or IL-23 to cultures did not cause increased IL-17 expression (data not shown). Membrane-bound factors or cytokines induced by the NLRP3 inflammasome that present at low concentrations in APC supernatants and work synergistically with IL-1β seem to be a more likely possibility.
To confirm that the IL-1 induction of Th17 pro-inflammatory response was driving inflammation in KI mice, we determined the effect of both anti-IL-1RI and anti-IL-17 on pre-existing dermatitis in KI mice.
In the anti-IL-1R1 study, we applied IL-1 receptor-1 (IL-1R1) blocking antibody or isotype control Ig to KI mice with skin inflammation using a previously described protocol with modification (Amadi-Obi et al., 2007). We found that after 10 days of antibody treatment the skin lesions of inflamed KI mice were markedly improved as compared with mice administered with isotype control. In particular, the lesions of such treated mice exhibited less epidermal/dermal thickening and decreased infiltration of neutrophils and/or macrophages (Figure 7A). In addition, real time RT-PCR analysis to assess cytokine production in the skin disclosed that anti-IL-1R1 treatment led to decreased levels of IL-17 as well as other pro-inflammatory cytokines such as IL-1β, IL-12p35 and IFNγ (Figure 7B). However, IL-12p40 and IL-12Rβ1 was unchanged and IL-12Rβ2 was increased (Figure 7B). We attribute the decreases in cytokines other than IL-17 to the general decrease in inflammatory cells in the skin lesions and the fact that IL-1β itself can induce pro-inflammatory cytokines. On the other hand, we attribute the increase in IL-12Rβ2 expression to repressive influence (probably excerted by IL-1β) which we consistently observed in the skin inflammations in KI mice (see Figures 3A and and4B).4B). Finally, a very similar set of results obtained when we assessed the effect of anti-IL-1-R1 antibody on induced skin inflammation associated with contact hypersenstivity in young KI mice yet free of spontaneous skin inflammation. The DTH reaction on tne ears of KI mice administered with anti-IL-1R was greatly reduced compared to that in KI mice received control Ig and, correspondingly, the skin histology and cytokine profile were consistent with an attentuated inflammatory reaction (Figures 7C-D). Once again, IL-1β and IL-17 was greatly reduced and IL-12Rβ2 chain was increased (Figures 7D).
In further studies to verify that IL-1β-induced IL-17 was a critical feature of the inflammation developing in KI mice, we administered anti-IL-17A neutralizing antibody to mice with skin inflammation as described above. Similar to anti-IL-1R1 treatment, the skin lesions in KI mice received anti-IL-17A were again markedly improved as compared to those administered with isotype control with respect to skin thickness and cytokine profile (Figure 7E-F). The decreases in cytokines other than IL-17 was probably due to the general decrease in inflammatory cells in the skin lesions. Once again, while IL-12Rβ1 was decreased, IL-12Rβ2 was increased (Figure 7F). Taken together, these anti-cytokine studies provide confirmatory evidence that IL-1β is the main cytokine associated with the skin inflammation developing in KI mice and this cytokine is acting through the induction of Th17 response.
In this study we analyzed a “knock-in” (KI) mouse strain engineered to harbor a point mutation in the NLRP3 gene equivalent to that in patients with Muckle-Wells Syndrome. The KI mice exhibited increased NLRP3 inflammasome activity characterized by ATP-independent IL-1β as well as IL-18 secretion and developed dermatitis as do patients with this autoinflammatory disorder (Gattorno et al., 2007). Importantly, the spontaneous and contactant-induced skin inflammation was characterized by a cytokine response strongly skewed to Th17 cytokines that was mediated at least in part by dysregulated antigen-presenting cells (APCs) producing IL-1β. This conclusion was highlighted by the fact that administration of either anti-IL-1R1 or anti-IL-17A antibodies led to marked resolution of the skin inflammation. The robust IL-17 responses are likely to contribute to the neutrophilic infiltration characterizing this inflammation as well as its intensification via neutrophil production of IL-1β.
While the NLRP3 inflammasome of APCs from knock-in (KI) mice was hyperactive it still required TLR stimulation for activation. Either TLR signaling blockade with neutralizing antibody or caspase-1 inactivation with specific inhibitors abolished IL-1β production from KI cells (data not shown). Nevertheless, NLRP3 activation was more easily triggered in the KI cells than in normal cells because it could occur in the absence of exogenous ATP, a molecular that is ordinarily necessary for the cell membrane pore formation through which TLR ligands can gain entry into the cell (Kanneganti et al., 2007). We found that such ATP-independent activation was not because of abnormalities of ligand entry into knock-in cells, but rather due to a lowered threshold for activation, whereby the inflammasome is triggered by low concentrations of ligand in the absence of exogenous ATP. This conforms to a model in which the inflammasome containing a mutated NLRP3 that undergoes easier activation because of a protein conformational change resulting from the mutation (Aksentijevich et al., 2007). In addition, the inflammasome of KI macrophages was activated independent from K+ efflux while still required marginal Panx1 channel activity. This indicated that low intracellular K+ concentration may be necessary but not sufficient for NLRP3 inflammasome activation while Panx1 should be a more critical component as also implied from recent studies (Pelegrin and Surprenant, 2006, 2007). Parenthetically, it is important to mention that the NLRP3 inflammasome is triggered by a broad number of TLR ligands and other substances such as uric acid (Yu and Finlay, 2008). Thus, it is likely that the recognition unit of the inflammasome (the LRR domain of NLRP3) is responding to a common intra-cellular molecule activated by such stimulants rather than to the original stimulus.
The lower threshold of NLRP3 activation in the KI mice could explain the fact that they were susceptible to spontaneous skin inflammation, since the latter was most likely triggered by local trauma from scratching or fighting that results in exposure of cells to low concentrations of TLR ligands that are sufficient to activate the inflammasome of KI mice but not of normal mice. Inflammasome activation and resultant IL-1β/IL-18 secretion may then lead to continued inflammasome activation by the production of endogenous inflammasome stimulants.
As evidenced by the fact that NLRP3 knock-in bone marrow transfer to Wt recipient mice resulted in the development of skin inflammation in recipient mice, we can conclude that the spontaneous skin inflammation that develop in KI mice was largely due to abnormalities in cells derived from the hematopoietic system. However, while NLRP3 is predominantly expressed in APCs such as macrophages, dendritic cells and neutrophils, it is also expressed in other cells including keratinocytes (Sutterwala et al., 2006; Watanabe et al., 2007). It is thus possible that the skin inflammation observed in KI mice was also due, in part, to release of inflammasome related cytokines from keratinocytes. In addition, it is possible that other inflammasome expressing cells contributed to the inflammation depending on the site of the inflammation.
An important outcome of this study was the observation that in both spontaneous and contactant induced skin inflammation, knock-in mice exhibited Th17-predominant cytokine profiles in affected skin as well as lymphoid tissues. In extensive in vitro co-culture studies to determine the basis of this profile we established that the abnormality lies in the antigen presenting cell rather than in the T cell. In addition, the co-culture experiments revealed that the skewing could at least in part be attributed to the excess production of IL-1β in accordance with previous work on the properties of this cytokine (Kryczek et al., 2007; Sutton et al., 2006; van Beelen et al., 2007; Yang et al., 2008). It should be noted, however, that in studies of human cells, augmentation for IL-17 production by IL-1β alone or in association with IL-23 was seen in relation to the differentiation of memory T cells, but not in naïve T cells which were the major cell population studied here (van Beelen et al., 2007; Yang et al., 2008). In addition, in studies of murine cells, IL-1β appeared to be playing a role as a substitute for IL-6 (Kryczek et al., 2007), or needed IL-23 to induce IL-17 expression (Sutton et al., 2006), whereas in the present study TGF-β and IL-6 incubation were necessary antecedents to the effect of IL-1β on Th17 differentiation. Thus, the present studies do not correlate with previous studies of IL-1β activity, except perhaps with those of Veldhoen et.al., who found that IL-1β together with TNF-α could augment Th17 cells in association with TGF-β and IL-6 (Veldhoen et al., 2006). This suggests that in this context the IL-1β was acting like IL-23, since it is now known that the major role of IL-23 is to sustain already differentiated IL-17 secreting cells rather than to induce these Th17 cells (McGeachy et al., 2007; Veldhoen et al., 2006). These studies therefore indicate that IL-1β plays an hitherto unexpected role in Th17 differentiation that will require further studies at the molecular level to fully understand.
In our studies of the ability of APCs or APC supernatants from KI mice to influence the differentiation of IL-1RI-deficient T cells it became evident that while IL-1β could account for much of the skewing toward Th17 differentiation, it probably does not account for all of the skewing. The latter was evident from the fact that while macrophage supernatants induced little more than baseline IL-17 responses in IL-1-R1-deficient cells, responses induced by CD11b+ cells were considerably higher than baseline responses. We speculate that the residual activity obtained with APCs is due to membrane-bound factors or cytokines produced at low concentration that are induced by inflammasome activation, such as IL-1 family members that act through unique receptors or work synergistically with IL-1β.
In both the tissue inflammation and co-culture studies, the enhancement of Th17 responses by APCs derived IL-1β was largely accompanied by decreased IFN-γ production. A possible explanation of this finding derives from previous work showing that IL-1β selectively inhibits IL-6 activated STAT-1 phosphorylation and thereby regulates T-bet transcription (Shen et al., 2000), the latter a key factor in Th1 differentiation. A second explanation derives from our observation that lesional tissue of KI mice exhibited down regulation of the IL-12Rβ2 chain and that treatment of mice with anti-IL-1R1 reversed this effect. It is thus possible that while inflammasome activation may lead to secondary induction of IL-12p70 and other pro-inflammatory cytokines (as was indeed observed in lesional tissue), such IL-12p70 cannot induce IFNγ production. On a related point, while lymph node and splenic CD4+ T cells from inflammed KI mice also manifested a Th17 skewing, the skewing in this case was rather modest and the cells also produced increased amounts of IFN-γ and IL-4 compared with control cells. We attribute this “global” increase in cytokine production to the fact that these peripheral cells were exposed to a mixed cytokine microenvironment, consisted not only by IL-1β but also other cytokines such as IL-18 and IL-33, the latter two being involved in stimulating production of IFNγ and IL-4 from naive T cells respectively (Nakanishi et al., 2001; Schmitz et al., 2005).
A major difference between Th17- and Th1-mediated inflammations is that the former is associated with neutrophil infiltration and the latter with macrophage infiltration (Kroenke et al., 2008; Steinman, 2008). Indeed, in both psoriasis and airway disease models, IL-17 is required for neutrophil recruitment (Miyamoto et al., 2003; Nograles et al., 2008). Thus, the heavy neutrophil infiltration in the inflamed skin tissue of NLRP3 KI mice appears to be directly related to the local predominance of Th17 cytokines, and previous studies showing that IL-1β was required for neutrophil recruitment can now be attributed to its induction of IL-17 (Hornung et al., 2008; Miller et al., 2007). On the other hand, IL-17 signaling is also required for optimal IL-1 production and function (Chabaud et al., 1998; Cheung et al., 2008; Granet et al., 2004; Koenders et al., 2005). Thus, it is possible that in the autoinflammation ensuing from a NLRP3 mutation, the hyperactive inflammasome leads to high IL-1β secretion as well as Th17 inflammation, the latter reciprocally leading to additional IL-1β production through the recruitment of inflammasome-containing neutrophils. This results in a positive feedback loop that ensures an inflammation with high level of neutrophil infiltration and tissue damage. This possibility is supported by the fact that administration of either anti-IL-1R1 or anti-IL-17 to KI mice with skin inflammation led to similar improvement of the inflammation. And this finding is consistent with the presence of an in vivo network of interacting cytokines, wherein IL-1β acts to expand IL-17 secreting cells and IL-17, in turn, recruits IL-1β-producing cells. On this basis, we would causiously suggest that anti-IL-17 may be an effective alternative approach to the treatment of inflammasome-related disorders.
Two other features of the KI mice also deserve mention as possible mechanisms of disease. One is that the mice exhibited elevated levels of anti-dsDNA antibody, which raised the question that they prone to the development of autoimmune disease. While such autoimmunity doesn't seem to be contributing to the skin inflammation, it's presence in KI mice cannot be ruled out in view of recent work showing that autoantibodies can induce macrophage activation and inflammatory cytokine production in rheumatoid arthritis (Monach et al., 2004). Another is that the KI mice displayed increased T cell activation suggesting that the inflammasome abnormality affects T cell homeostasis. While the activation seen most likely reflects the on-going inflammation (it was not seen in mice without inflammation), it remains possible that homeostatic disturbances are present in view of recent studies showing that activation of the inflammasome is associated with necrotic cell death. It thus becomes possible that endogenous proteins resulting from such cell death trigger autoimmune responses (Willingham et al., 2007).
In conclusion, we have generated a mouse model mimicking human Muckle-Wells syndrome and found that NLRP3 inflammasome hyperactivation leads to excessive IL-1β secretion and secondarily Th17-dominated inflammation. This model therefore predicts that humans with autoinflammatory disease may also have IL-17-dominated inflammation. This findings may thus lead to new ways of controlling autoinflammatory diseases.
NLRP3 R258W Knock-In mice were generated as described in Supplemental Data according to reported mouse NLRP3 gene sequence (Anderson et al., 2004). Knock-in mice as well as age and gender matched C57BL/6 wild type mice (Jax664) and IL-1 receptor 1 deficient mice (on C57BL/6 background, Jax3245) were purchased from Jackson Laboratories (Bar Harbor, ME) and housed in pathogen-free facility. Animal use adhered to National Institutes of Health Animal Care Guidelines.
Ear, liver, spleen and cervical lymph nodes were harvested from KI or Wt mice, part of them was paraffin embedded and sectioned for H&E staining for inflammation and architecture study, the rest were embedded in O.C.T. (Optimal Cutting Temperature, SAKURA) and kept frozen at -80°C followed with sectioning for immunohistochemical staining with specific antibodies.
Recombinant murine M-CSF, GM-CSF and IL-4 were from Peprotech. Unless otherwise described, the doses of TLR ligands applied for stimulation were as follows: PGN (TLR2 ligand, 10μg/ml; InvivoGen); P3CSK4 (TLR2 ligand, 0.2 g/ml; InvivoGen); PolyI:C (TLR3 ligand, 25μM; InvivoGen); LPS (TLR4 ligand, 0.2μg/ml; Sigma-Aldrich); Flagellin (TLR5 ligand, 0.5μg/ml; InvivoGen); Loxoribine (TLR7 ligand, 100μM; InvivoGen); Imiquimod (R837, TLR7 ligand, 5μg/ml). Panx1 blocking peptide (10panx1, WRQAAFVDSY, Sigma Genosys). ATP, Apyrase, KN-62, TNP-ATP were all from Sigma-Aldrich.
Concentrations of cytokines from mouse cell culture supernatants were determined by BD Biosciences ELISA kits for assay of IL-12p40, TNF-α, IL-10, IL-6 and IFNγ; eBioScience kits for assay of IL-1β, IL-17A, IL-23p19; R&D systems kits for assay of IL-12p70, IL-18 and IL-21. Mouse Anti-ds DNA ELISA kit was from Alpha Diagnostic Intl. Inc..
For NLRP3 expression analysis, BMDM lysis and blotting was performed as described previously (Meng et al., 2003). The membrane was blotted with NLRP3 Ab (M-12, sc-34410) (Santa Cruz) and β–Actin Ab (Cell Signaling). For caspase-1 and mature IL-1β analysis, BMDM was pretreated for 4 h with LPS, 5mM ATP (Sigma) was added and 30 minutes later, NP40 (final 1%), DTT (final 10mM), as well as protease inhibitors were added to culture medium. Cells as well as supernatant were collected, nuclei and membranes were removed via centrifugation, SDS sample buffer was added before boiling and analyzing by immunoblotting for caspase-1 as well as cleaved IL-1β with Abs as follow: Caspase-1 P20 (kind gift from Dr. Nunez), Caspase-1 P10 (sc-514) and IL-1β (H-153) (Santa Cruz) (Franchi et al., 2007; Pelegrin and Surprenant, 2006).
Immunoprecipitation (IP) followed by immunoblotting (IB) of caspase-1 P10 was carried out with BMDM culture supernatant by using the same antibody mentioned above.
BMDM were prepared as described previously (Spiller et al., 2007), briefly, bone marrow (BM) cells were cultured in Petri dishes in 10 ml of complete RPMI medium in the presence of recombinant M-CSF (40 ng/ml). After 3 days of culture, half of the medium was replaced by fresh medium. On day 6, adherent cells were harvested and used as indicated.
For making BMDC, BM cells were prepared from NLRP3 KI as well as control mice and cultured in Petri dishes (1 × 106/ml) in 10 ml of complete RPMI medium supplemented with recombinant GM-CSF (20 ng/ml) and IL-4 (20 ng/ml). After 3 days of culture, half of the medium was replaced by fresh medium and on day 6 cells were harvested and sorted by anti-mouse CD11c magnetic beads (Miltenyi Biotech) prior to use in studies of cytokine secretion or co-culture experiments (Watanabe et al., 2008).
Total lymph node and spleen cell staining for infiltrated neutrophils was carried out with 2 different antibodies, anti 7/4 antigen (CL8993B, Cedarlane labs) and Ly6G (clone 1A8, 551459, BD Biosciences). Splenic macrophages were purified via positive selection with anti-CD11b microbeads (Miltenyi Biotech) from total splenocytes and applied for IL-1β secretion study or co-culture with T cells. CD4+ T cells were purified from lymph nodes and spleen with microbeads (L3T4, Milteny, Auburn, CA) via magnetic sorting. These T cells were either stimulated with plate-bound anti-CD3 (10μg/ml) and soluble anti-CD28 (2 μg/ml) for 2 days for ELISA study or activated for 4 days with anti CD3/28 followed with re-activation (PMA/ionomycin, 0.25 μg/ml each, 5h) and intracellular staining for IL-17A, IFNγ as well as IL-4. Golgi stop (0.67 ul/ml, BD, 51-2092KZ, 554724) was added 3 hrs before cells were harvested for intracellular staining. CD4+ T cell differentiation was done as described previously and Flow Cytometry was using Becton Dickinson FACS Caliber with Cell Quest programme (Zhang and Boothby, 2006).
Total RNA was extracted from mice skin or cultured cells with Trizol (Invitrogen) and cDNA was synthesized using Reverse Transcription Kits (Applied Biosystems) according to manufacturer's instructions. RT-PCR of exon 3 of NLRP3 gene from either KI BMDM RNA or Wt Control with the following primers: 5′ GTGAGAGTGTGGACCTCAACAG 3′ and 5′ GTCCAGGAGATGCTGCAGTTTC 3′. Equal loading was confirmed by simultaneous HPRT amplification. Cytokine RT-PCR was performed using Taqman Gene expression assays as well as gene-specific primer/probe settings (Supplemental Data).
Female mice (6-10 weeks old) were sensitized by topical applicatiion of 20 μl 5% DNCB (Sigma) dissolved in acetone/oliver oil solution (4:1)(v/v) to the shaved middle back skin (~ 2 cm2) on day 0. On day 5, all mice were challenged by topical application of 20 μl of 0.5% DNCB on both sides of both ears. 24 hours after the irritation, ear samples were taken for either histologic study upon H&E staining or extraction of RNA for real time RT-PCR analysis. Spleen and lymph node T cells were isolated from these DNCB treated mice for intracellular staining of IL-17 and IFN-γ.
Bone marrow chimeras were generated as described with modification (Zhang et al., 2009). Briefly, recipient C57BL/6 mice were irradiated with one dose of 940 rads at 70 rads/min from a 137Cs soruce. 2mg/ml Neomycine sulfate was given in drinking water 1 day prior to till 2 weeks after irradiation. 24 h after irradiation, 2 × 107 total bone marrow cells from gender matched KI or Wt donor mice were injected intravenously into recipients. 4 weeks after injection, total genomic DNA from 100ul blood of recipient mice was subjected to knock-in specific PCR analysis to confirm successful repopulation of donor cells. 8 weeks after bone marrow transfer, recipient mice were used for experiments.
For IL-17 neutralization, Anti-IL-17 Ab (MAB421) was applied with Rat IgG2A (MAB006) as Isotype Control. For IL-1R1 blockade, Anti-L-1RI Ab (MAB7711) was applied with Ms IgG1 (MAB002) as control Ig (all these antibodies were from R&D systems). For intervention of spontaneous inflammation from KI mice, antibody was injected intraperitoneally with 100ug in 200ul PBS every other day for 5 injections. 2 days after the last injection, mice were sacrificed for analysis. For blockade of DNCB induced hypersensitivity, the 1st Ab injection was 2 days before DNCB sensitization, then inject on day 0 (DNCB sensitization), day 1, day 3 and day 5 (DNCB irritation), mice were used for experiment on day 6.
Two-tailed Student's t test was used to evaluate the significance of differences and P < 0.05 was regarded as statistically significant.
Table S1: Complete Blood Cell counts from Wild type (Wt), Knock-in without inflammation (KI fine) and inflamed Knock-in (KI inflamed) mice. Although KI mice without obvious inflammation exhibited only slightly increased number of neutropils and monocytes as compared to littermate control, those with skin inflammation displayed much greater increase in levels of granulochytes, monocytes, and platelets.
Figure S1. Generation of NLRP3 R258W Knock-In mice.
(A) Schematic representation of the NLRP3 genomic locus, gene-targeting construct and the knock-In (KI) targeted NLRP3 allele. Dashed lines indicate homologie arms for recombination. In the targeting vector, exon 3 was flanked by LoxP sites to provide the option of converting the knock-in mice to conditional knock-out mice. A neomycin gene (Neo) flanked by FRT sites was inserted in an intron, with Neo providing selection resistance and the FRT sites offering the opportunity to remove the Neo cassette from targeted genome via Flipase expression. 5′ and 3′ probes are also indicated.
(B) Genomic DNA from offspring mice of chimera X C57BL/6 mating parents was digested with EcoRV and Southern blotted using the 3′ probe. The expected band size for Wt is 9.4 kb and for KI is 6.0 kb. The presence of a 6.0 kb fragment in the KI sample due to the introduction of a new EcoRV site confirmed successful integration of the R258W mutant from targeting vector.
(C) Total RNA extracted from bone marrow-derived macrophages (BMDM) of both Wt and KI mice was reverse transcribed with random primers, the cDNA was applied as template for PCR with primers spanning R258W mutation. Wt and KI samples yield identical bands (upper panel), indicating that transcription of exon 3 was normal in the KI cells. Equal loading was confirmed with HPRT amplification (lower panel).
(D) Sequencing of PCR products from (C) revealed the successful integration of R258W (CGA/TGG) mutation in NLRP3 gene.
(E) mRNA from resting or LPS-stimulated BMDM of Wt or KI mice were subjected to real time RT-PCR using primers and probes downstream of the R258W mutation, both Wt and KI mice express the same amount of NLRP3 following stimulation, which revealed that the mutated gene is normally expressed.
(F) Western blotting analysis of NLRP3 expression in BMDM from either NLRP3 KI or Wild type mice with or without LPS stimulation overnight (upper panel). The same blot was reprobed with Actin antibody as control for equal protein loading (lower panel).
Data shown are representative of two (C) or three (E,F) independent experiments.
Figure S2. Inflammasome of antigen presenting cells of NLRP3 KI mice was activated upon TLR stimulation in the absence of ATP.
(A) As in Figure 1A. Mature IL-18 from culture supernatants was measure with ELISA.
(B) As in (A), except that BMDC instead of BMDM from KI and Wt mice were analyzed for mature IL-1β production.
(C) KI and Wt splenic CD11b+ cells were analyzed as in (B) upon indicated PAMPs stimulation.
(D) As in Figure 1B, pro- and mature IL-1β were analyzed with actin expression served as loading control.
(E) Immunopreciptation of caspase-1 P10 from cell culture supernatants of either Wt or KI BMDM with or without LPS stimulation for 4 h followed by immunoblotting with the same anti-caspase-1 P10 antibody. Two repeats of this experiment was indicated as Exp1 and Exp2.
Data shown are representative of three independent experiments.
Figure S3. Inflammasome independent cytokine production from knock-in macrophages was normal.
(A and B) As in Figure 1A. Listed cytokines were measured with ELISA. Data shown are representative of three independent experiments.
Figure S4. Endogenous ATP contributes to Inflammasome activation of NLRP3 KI BMDM that has a lower threshold.
(A) ELISA measurement of IL-6 from samples used in Figure 1D, confirming lack of toxicity from 10panx1 incubation.
(B) ELISA measurement of IL-1β release from supernatants of BMDM from KI mice stimulated with 10ng/ml of LPS in the presence of Apyrase, KN-62 or TNP-ATP for 24 h.
(C) Western blot analysis of cell lysates for detection of pro- and mature IL-1β upon stimulation with high (5ug/ml) or low (0.1ug/ml) amount of LPS overnight in the presence or absence of ATP pulse.
Data shown are representative of two independent experiments.
Figure S5. Sub-optimal linear growth and spontaneous inflammation of NLRP3 KI mice.
(A) Photos of NLRP3 KI and littermate control (Wt) mice at age of 6 weeks, indicating ruffled coats, hair loss (bare skin) and smaller size of KI mice.
(B) H&E staining of bare skin of NLRP3 KI mice and normal skin of Wt mice. Arrow, hair follicles, those from knock in are abnormally thick and contain hyper-proliferating epithelial cells.
(C) Liver, spleen and cervical lymph nodes from inflamed knock-in mice are enlarged in comparison to Wt control.
(D) Immunohistochemical staining of frozen sections from livers of NLRP3 KI or control mice with granulocyte (top panel) and macrophage (bottom panel) specific antibodies. Arrow heads illustrate cell infiltration in inflamed KI mice. pv, protal vein; bd, bile duct.
(E and F) Granulocyte (E) and macrophage (F) specific immunohistochemical staining with anti Gr-1 and anti F4/80 antibody for frozen sections from cervical lymph node and spleen of NLRP3 KI or Wt control mice.
Data are representative of two (A, B, D-F) and five (C) independent experiments.
Figure S6. KI mice produced elevated amount of autoantibody.
ELISA measurement of anti-dsDNA auto-antibody from fresh serum of littermate Wt or KI (not inflamed, “KI fine”) mice as wells as older KI mice (KI inflamed) displaying spontaneous inflammation.
Data are representative of two independent experiments.
Figure S7. T helper cells from inflammed KI mice are pre-activated.
Flow cytometry of the surface expression of indicated cell activation markers from CD4+ T cells of spleen (SPL) and lymph node (LN) of KI mice with or without inflammation as well as Wt mice. Filled peak, Wt control; dashed line, KI mice without inflammation; solid line, inflamed KI mice.
Data are representative of three independent experiments.
Figure S8. CD4+ T cells from KI mice with skin inflammation produced elevated amounts of IL-17, IFNγ and IL-4 upon TCR activation in vitro.
Purified CD4+ T cells from spleen (SPL) or lymph nodes (LN) of inflamed KI mice or Wt mice were activated with plate-bound anti-CD3 (10ug/ml) and soluble anti-CD28 (1ug/ml) for 2 days; culture supernatants were harvested for IL-17 (A), IFNγ (B) and IL-4 (C) ELISAs. Asterisk indicates statistical significance in comparison to Wt as follow: *, p<0.05; **, p<0.01; ***, p<0.001.
Data shown are mean ± S.D. from a representative of three (A and B) or two (C) independent experiments.
Figure S9. Knock-in Bone Marrow Reconstitution of Wt mice resulted in skin inflammation.
(A) Photos of recipient mice that received KI or Wt bone marrow cell transfer, showing inflammation in ear skin of KI-Wt bone marrow chimeric mice.
(B) Liver, spleen and cervical lymph nodes from inflamed KI-Wt bone marrow chimeric mice are enlarged in comparison to Wt-Wt control.
(C) H&E staining of skin and liver of inflamed KI-Wt or Wt-Wt bone marrow chimeric mice. Upper panel, ear epidermis and dermis of KI-Wt bone marrow chimeric mice are thickened accompanied with leukocytes infiltration. Lower panel, livers from KI-Wt bone marrow chimeric mice showed marked leukocyte infiltration (arrow head) in the protal areas (pv, protal vein) that resulted in loss of typical bile duct (bd) structure.
(D) H&E staining of cervical lymph node and spleen tissues of KI-Wt or Wt-Wt bone marrow chimeric mice. GC, germinal center. RP, red pulp; WP, white pulp.
(E) As in Figure 2E. Filled peak, Wt-Wt; solid line, KI-Wt.
Data are representative of two independent experiments.
Figure S10. Differentiation of CD4+ T cells from inflammed KI mice into Th1, Th2 or Th17 cells is normal.
Flow cytometric study of the intracellular expression of IL-17, IFNγ and IL-4 from CD4+ T cells of KI and Wt mice 4 days after culture under indicated polarizing conditions upon activation with PMA/ionomycin, numbers indicate frequences of IL-17, IFNγ and/or IL-4 positive cells.
Data are representative of two independent experiments.
Figure S11. CD11b+ cells from KI mice supress IFNγ and IL-4 production under respective T helper conditions.
Flow cytometry of the intracellular expression of IL-17, IFNγ and IL-4 from co-culture of Wt CD4+ T cells and CD11b+ cells from either inflamed KI or normal Wt mice as in Figure 6A, except that in this experiment the cells were cultured with Th0, Th1 and Th2 polarizing cytokines instead of Th17 condition.
Data are representative of three independent experiments.
Figure S12. IL-1 from supernatant of KI BMDM augments normal Th17 differentiation in the presence of TGFβ and IL-6.
(A) (left) Flow cytometry of the intracellular expression of IL-17 and IFNγ from CD4+ T cells of Wt mice 4 days after culture in the presence of supernatant from either KI or Wt BMDM without stimulation or stimulated with PGN for 24 h or primed with LPS for 5 h at 1:1 ratio with regular medium. (right) Quantification of results from flow cytometry at left depicted as the ratio between percentage of IL-17 and IFNγ producing cells in each sample, asterisk indicates statistical significance of IL-17/IFNγ between T cells incubated with supernatants from KI and Wt BMDM as follow: *, p<0.05.
(B) (left) As in (A), except that either KI (upper two rows) or Wt (lower two rows) BMDM derived supernatants were incubated with either Wt (1st and 3rd rows) or IL-1 receptor I deficient (IL1RI-/-) (2nd and 4th rows) CD4+ T cells as indicated. (right) Quantification of results from left panel depicted as IL-17/IFNγ as in (A), asterisk indicates statistical significance of IL-17/IFNγ between Wt and IL1RI-/- CD4+ T cells incubated with supernatants from either KI (upper panel) or Wt (lower panel) BMDM as follow: *, p<0.05.
Data shown are mean ± S.D. from a representative of two independent experiments.
Figure S13. IL-1 does not support Th17 differentiation in the absence of Th17 polarizing cytokines (TGFβ+IL-6).
Flow cytometry of the intracellular expression of IL-17 and IFNγ from Wt CD4+ T cells 4 days after culture in the absence (upper 2 rows) or presence of TGFβ (lower 2 rows) as indicated. Supernatant from either KI or Wt BMDM (A) or BMDC (B) without stimulation or stimulated with PGN for 24 h or primed with LPS for 5 h was added at 1:1 ratio with regular medium for T cell culture.
Data are representative of two independent experiments.
We thank Dr. Ailing Lu for help with targeting vector construction and Immunohistochemical staining, Drs. Martha Quezado and Helgi Valdimarsson for help with histological analysis, Dr. Gabriel Nunez for providing Caspase-1 P20 antibody.
Competing interestes statement: The authors declare that they have no competing financial interestes.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.