IKKβ primes inflammasome formation by recruiting NLRP3 to the trans-Golgi network

Summary

The NLRP3 inflammasome plays a central role in antimicrobial defense as well as in the context of sterile inflammatory conditions. NLRP3 activity is governed by two independent signals: the first signal primes NLRP3, rendering it responsive to the second signal, which then triggers inflammasome formation. Our understanding of how NLRP3 priming contributes to inflammasome activation remains limited. Here, we show that IKKβ, a kinase activated during priming, induces recruitment of NLRP3 to phosphatidylinositol-4-phosphate (PI4P), a phospholipid enriched on the trans-Golgi network. NEK7, a mitotic spindle kinase that had previously been thought to be indispensable for NLRP3 activation, was redundant for inflammasome formation when IKKβ recruited NLRP3 to PI4P. Studying iPSC-derived human macrophages revealed that the IKKβ-mediated NEK7-independent pathway constitutes the predominant NLRP3 priming mechanism in human myeloid cells. Our results suggest that PI4P binding represents a primed state into which NLRP3 is brought by IKKβ activity.

Introduction

Cells of the innate immune system employ a repertoire of so- called pattern recognition receptors (PRRs) to discriminate self from non-self. Engagement of these PRRs triggers a broad array of effector functions geared toward eliminating a microbial threat. The inflammasome pathway constitutes a special class of this PRR system that is signified by the activation of the cysteine protease caspase-1 in a large supramolecular protein complex.¹ Activation of caspase-1 causes maturation of pro-inflammatory cytokines, most prominently IL-1β,² as well as the induction of a special type of cell death, known as pyroptosis.³ Among inflammasome sensors, NLRP3 plays a pivotal role in antimicrobial defense as well as sterile inflammatory diseases.⁴ This is owed to the fact that NLRP3 is a highly sensitive, yet non-specific PRR. In this regard, NLRP3 has been shown to respond to the perturbation of cellular homeostasis by a broad array of diverse stimuli, rather than being activated by a specific microbe-derived molecule.⁵ K⁺ efflux from the cytosol has been identified as a common denominator of many NLRP3 triggers.⁶ In this function, several types of lytic cell death have been shown to result in secondary engagement of the NLRP3 inflammasome pathway.⁷ However, K⁺ efflux-independent NLRP3 stimuli have also been described,^8,9 and a recent report has identified dispersal of the trans-Golgi network (TGN) as a common denominator of both K⁺ efflux-dependent and -independent NLRP3 triggers.¹⁰

Unlike other inflammasome sensors, NLRP3 critically depends on the engagement of a priming step.¹¹ This priming signal can be provided by different types of receptors, typically PRRs that trigger NF-kB activation. Lipopolysaccharide (LPS) activating TLR4 is commonly used to provide a priming signal preceding the actual NLRP3 activation step. Initially, the necessity of priming had been ascribed to the fact that NLRP3 is expressed at limiting amounts in murine macrophages. In this respect, it has been shown that in a process now also called “transcriptional priming,” NF-kB activating stimuli drive the expression of Nlrp3, thereby facilitating its activation.^12,13 In line with these findings, inhibition of transcription blocks this mode of NLRP3 priming, whereas transgenic expression of NLRP3 bypasses the requirement of transcriptional priming.^12,13 Extending this concept, NLRP3 can also be primed non-transcriptionally, e.g., by a short pulse of LPS treatment.^14–16 These modes of priming have been ascribed to a variety of post-translational modifications of NLRP3, including phosphorylation, de-phosphorylation, de-ubiquitination, and de-sumoylation.^17,18 Although being mechanistically unrelated, these events are commonly referred to as post-translational or non-transcriptional priming. The fact that many cells already express NLRP3 at sufficient amounts under steady-state conditions underscores the importance of non- transcriptional priming.¹⁹

Despite considerable insight into pathways that result in NLRP3 priming, the activation step of the NLRP3 inflammasome and its interconnection with priming have remained enigmatic. In this regard, we and others have identified the mitotic spindle kinase NEK7 (NIMA-related kinase 7) as a critical cofactor in NLRP3 activation in murine cells.^20–22 Notably, this role of NEK7 is distinct from its function in the cell cycle, as its kinase activity is not required for NLRP3 activation.^21,22 NEK7 has been suggested to interact with NLRP3 in a K⁺ efflux-dependent manner, and deletion of NEK7 does not affect transcriptional NLRP3 priming.^21,22 This, in combination with a study modeling a NEK7-containing NLRP3 pyroptosome based on a cryo-EM structure of the NLRP3/NEK7 complex,²³ has led to the conclusion that NEK7 is involved in NLRP3 activation downstream of K⁺ efflux.²⁴ Of note, studies identifying NEK7 as an indispensable factor for NLRP3 activation have mainly been conducted in murine models. Here, we report that reductionist genetic dissection of NLRP3 signaling in human cells revealed an additional pathway of NLRP3 priming that enables NLRP3 inflammasome activation independently of NEK7.

Results

Human iPSC-derived macrophages and human myeloid cell lines activate NLRP3 independently of NEK7

We and others have previously described NEK7 to be essential for the activation of the NLRP3 inflammasome in the murine system.^20–22 To study the role of NEK7 in the human system, we adopted a recently described iPSC-derived macrophage model, in which human iPS cells are differentiated into macrophages in vitro (hiPS-Macs).²⁵ hiPS-Macs are fully capable of inflammasome activation: after priming with LPS, activation of the NLRP3 inflammasome with the ionophore Nigericin or the NAIP-NLRC4 inflammasome with Needle Tox resulted in pyroptosis (LDH release) accompanied by the release of IL-1β and IL-18 (Figures S1A and S1B). Both cytokine and LDH release in response to Nigericin, but not Needle Tox, were sensitive to the NLRP3 inhibitor MCC950 (Figures S1A and S1B). To investigate the role of NEK7 in NLRP3 inflammasome activation in hiPS-Macs, we generated NEK7^-/- iPS cell clones via CRISPR-Cas9 genome editing. NEK7 deficiency neither affected macrophage differentiation nor did it lead to altered NLRP3 expression levels (Figure S1C). Contrasting previous reports from mouse cells,^20–22 NEK7-deficient hiPS-Macs showed no major impairment of their NLRP3 inflammasome response (Figures 1A and S1D). Cytokine and LDH release following Nigericin stimulation remained sensitive to MCC950 in NEK7^-/- hiPS- Macs, confirming that Nigericin-induced pyroptosis was still mediated by NLRP3 in these cells (Figure 1A). As expected, NAIP-NLRC4 activation and IL-6 release also proceeded unperturbed in NEK7^-/- hiPS-Macs (Figures 1A and S1D).

Figure 1. Human iPSC-derived macrophages and human myeloid cell lines activate the NLRP3 inflammasome independently of NEK7.

(A) Four clones per indicated genotype of human iPSCs were differentiated into macrophages (hiPS-Macs), primed with LPS for 4 h and then treated with the inflammasome activators Nigericin (NLRP3) or Needle Tox (NAIP-NLRC4) in the presence of the NLRP3 inhibitor MCC950 as indicated before release of LDH (left), IL-1β (middle), and IL-18 (right) was measured. Dots represent separately differentiated iPS cell clones of the indicated genotypes.

(B and C) BLaER1 monocytes of the indicated genotypes were primed with LPS for 4 h and subsequently stimulated with Nigericin or Needle Tox. LDH release (B) of one or two clones per genotype is depicted. (C) One representative immunoblot of three independent experiments is shown.

(D) Three clones of THP-1 cells of the indicated genotypes were primed with Pam3CSK4 for 4 h and subsequently stimulated with Nigericin for 2 h before release of LDH (left) and TNF (right) were measured. Two different sgRNAs against NEK7 were used (#1 and #2). Dots represent individual clones.

(E) THP-1 cells of the indicated genotypes were primed with Pam3CSK4 for 4 h and subsequently stimulated with Nigericin for 2 h before immunoblotting. One representative immunoblot of three independent experiments is shown.

(F) NLRP3^-/- BLaER1 cells expressing the indicated NLRP3 orthologs under the control of a doxycycline-inducible promoter were treated with doxycycline for the last 24 h of differentiation, primed with LPS for 4 h and subsequently stimulated with Nigericin (left) or Needle Tox (right) for 2 h. The same vector expressing mCherry instead of NLRP3 was used as a mock control.

(G) Western blot of cells treated as in (F), one representative of three independent experiments is shown.

Data are represented as mean ± SEM with dots representing biological replicates conducted on separate days unless indicated otherwise (one outlier in B is not depicted #). ***p < 0.001, **p < 0.01, *p < 0.05, ns p ≥ 0.05 calculated by two-way ANOVA followed by Tukey’s test (A, B, and D: TNF) or Šidák’s test (D: LDH).

See also Figures S1 and S2.

We then sought to further characterize NEK7-independent NLRP3 activation in human cells. To this end, we used the BLaER1 transdifferentiation system that we have previously adopted to study innate immune sensing.^26,27 Mirroring hiPS- Macs, NEK7-deficiency showed no impact on NLRP3 inflammasome activation as assessed by release of LDH and IL- 1b (Figures 1B and S1E). To address whether the role of NEK7 for NLRP3 activation in human cells was overshadowed by a functional redundancy with its close homolog NEK6, we generated cells deficient for both NEK6 and NEK7. Analogous to NEK7-deficient cells, NEK6^-/- × NEK7^-/- BLaER1 cells displayed unimpaired activation of the NLRP3 inflammasome (Figures 1B and S1E). As expected, NLRP3^-/- BLaER1 cells showed no response to Nigericin stimulation, whereas they remained responsive to NAIP-NLRC4 inflammasome activation (Figures 1B and S1E). In line with these observations, caspase-1 maturation upon Nigericin treatment also proceeded independently of NEK7 (Figure 1C). Pretreatment with the NLRP3-specific inhibitor MCC950²⁸ or prevention of K⁺ efflux by increased extracellular K⁺ concentration⁶ blunted NLRP3 activation in wild type, NEK7^-/- and NEK6^-/- × NEK7^-/- cells stimulated with Nigericin, whereas it left the NAIP-NLRC4 inflammasome intact (Figures S1F and S2A–S2D), indicating that Nigericin still relied on inducing K⁺ efflux to trigger NLRP3 inflammasome activation in absence of NEK7. In line with the results obtained in BLaER1 cells, THP-1 cells deficient in NEK7 showed no attenuation of Nigericin-triggered inflammasome activation, whereas NLRP3^-/- THP-1 cells were completely defective (Figures 1D, 1E, and S2E).

NEK7-independent NLRP3 activation in human cells contrasts with NEK7-dependent NLRP3 activation in mouse cells published by us and others.^20–22 To investigate if this difference is caused by species-specific features of the human and mouse orthologs of NLRP3, we reconstituted NLRP3^-/- BLaER1 cells with different NLRP3 orthologs. Phenocopying the human ortholog, NEK7-deficient BLaER1 cells expressing mouse NLRP3 (mmNlrp3) mounted an unperturbed response to Nigericin (NLRP3) and Needle Tox (NAIP-NLRC4) (Figures 1F, 1G, and S2F). Taken together, these results demonstrate that unlike mouse cells, human cells are intrinsically capable of activating NLRP3 in a NEK6- and NEK7-independent manner.

Priming activates IKKβ to enable NEK7-independent NLRP3 inflammasome formation

Having established that human cells activate NLRP3 in absence of NEK7, we wondered whether the NEK7-independent pathway could be triggered in mouse cells where NLRP3 activation has been shown to depend on NEK7.²¹ Here, we used an immortalized mouse macrophage cell line constitutively expressing mmNlrp3 (mmMacs) in which we had initially discovered the requirement of NEK7 for NLRP3 activation through a forward genetic screen.²⁰ These cells do not require transcriptional priming of NLRP3 for inflammasome activation, and stimulation with Nigericin alone already activated NLRP3 in a fully NEK7-dependent manner (Figures 2A and 2B). When testing different priming modalities, we found that simultaneous treatment with LPS and Nigericin led to NLRP3 activation independently of NEK7, as determined by LDH release and caspase-1 maturation 4 h after stimulation (Figures 2A and 2B). Concurrent stimulation with Pam3CSK4 or R848 instead of LPS (Figures S3A–S3D) and with ATP instead of Nigericin (Figure 2C) similarly resulted in a NEK7-independent response. Of note, this NEK7 bypass triggered by concurrent priming and stimulation was only uncovered when studying the inflammasome response several hours after treatment (Figure S3E). Indeed, the NLRP3 inflammasome response 1 h following concurrent LPS + Nigericin treatment was still NEK7-dependent (Figure S3F). However, concomitant LPS treatment enhanced this early NEK7-dependent NLRP3 inflammasome response compared with Nigericin treatment alone. This is consistent with previous reports on rapid, non-transcriptional NLRP3 priming enabling accelerated inflammasome formation.^14,15,29 Taken together, these results suggest that NEK7-mediated priming and the LPS-induced NEK7 bypass pathway are not only functionally redundant but may also act synergistically to accelerate NLRP3 activation.

Figure 2. Priming activates IKKβ to bypass NEK7 via a translation-independent mechanism in mouse cells.

(A–C) Mouse macrophages constitutively expressing mmNlrp3 (mmMacs) of the indicated genotypes were stimulated with LPS + Nigericin simultaneously for 4 h, with DNA for 28 h or with LPS + ATP for 2 h. (B) One immunoblot representative of two clones from two independent experiments is shown.

(D) mmMacs were pretreated with cycloheximide (CHX) for 30 min and stimulated as in (C).

(E) Two mmMacs clones per genotype were stimulated as indicated. Release of IP-10 (left), TNF (middle), and LDH (right) of two clones (sub-columns) from three independent experiments (sub-rows) are depicted as heatmaps.

(F and G) mmMacs of the indicated genotypes were stimulated as in (A) before the release of LDH (F) and TNF (G) was measured.

(H) mmMacs of the indicated genotypes were stimulated as in (A) for 2 h. One representative of three independent biological replicates is shown.

Data are represented as mean ± SEM with dots representing biological replicates conducted on separate days. ***p < 0.001, **p < 0.01, *p < 0.05, ns p ≥ 0.05 calculated by two-way ANOVA followed by Tukey’s test (A, C, and F), Šidák’s test (D), or Dunnett’s test (G).

See also Figures S3 and S4 and Table S1.

LPS sensing initiates diverse transcriptional programs. However, NEK7-independent priming remained functional in the presence of translation-blocking concentrations of cycloheximide (CHX), indicating that it does not require de novo protein synthesis (Figures 2D and S3G). To elucidate the signaling cascade of NEK7-independent post-translational priming, we genetically perturbed TLR4 and its downstream signaling adaptors TRIF (Ticam1) and MyD88 in either unmodified or Nek7^-/- mmMac cells. NLRP3 activation in response to Nigericin treatment remained intact in Ticam1^-/- or Myd88^-/- cells (Figure 2E, right panel; Table S1), whereas these cells displayed a selective lack of antiviral (IP-10) or pro-inflammatory (TNF) gene expression, respectively (Figure 2E, left and middle panels). TLR4 deficiency abrogated LPS-dependent cytokine production altogether (Figure 2E, left and middle panels). Accordingly, unlike their TLR4-sufficient counterparts, Nek7^-/- × Tlr4^-/- cells were fully defective in NLRP3 activation (Figure 2E). In contrast, Nek7^-/- cells additionally deficient in either MyD88 or TRIF were still able to mount an NLRP3 inflammasome response after LPS + Nigericin treatment, albeit less effectively (Figure 2E). As expected, Myd88^-/- × Ticam1^-/- cells deficient in NEK7 were fully defective in NEK7-independent NLRP3 activation (Figure 2E). Altogether, these results indicate that the NEK7 bypass can be induced downstream of both Myd88 and TRIF signaling. To identify the common factor mediating the NEK7 bypass, we turned our attention to the TAK and IKK complexes that constitute the apical kinase complexes governing pro-inflammatory signal transduction downstream of both MyD88 and TRIF. When we used the small molecule Takinib to block the activity of TAK1, the key kinase of the TAK complex, we found that the NEK7 bypass was largely inhibited, whereas NLRP3 activation in response to Nigericin remained intact (Figure S4A). We obtained analogous results when we blocked IKKβ, a kinase in the IKK complex, using TPCA-1 (Figure S4B). Of note, for both inhibitors, the NEK7 bypass was not fully abrogated; however, it was attenuated to the same extent as the production of the NF-kB-dependent cytokine TNF (Figures S4A and S4B). The NEK7 bypass was blocked when we deleted Ikbkb, the gene coding for IKKβ, but remained unperturbed when we deleted Chuk, the gene coding for IKKα (Figures 2F–2H). Nek7^-/- × Ikbkb^-/- mmMac cells were almost completely defective in NLRP3 inflammasome activation, whereas AIM2 inflammasome activation in response to dsDNA transfection remained intact (Figures 2F–2H). Priming with R848 or NLRP3 activation with ATP similarly resulted in IKKβ-dependent NLRP3 inflammasome formation independently of NEK7 (Figures S4C and S4D). In conclusion, since IKKβ is activated downstream of the TAK1 complex, these findings suggest that IKKβ constitutes the critical kinase mediating NEK7-independent NLRP3 inflammasome formation.

RIPK1 and caspase-8 have been implicated in non-transcriptional NLRP3 priming.³⁰ Although the NEK7 bypass continued to function in Nek7^-/- × Ripk1^-/- mmMac cells (Figure S4E), Nek7^-/- × Casp8^-/- mmMacs were fully defective in activating the NLRP3 inflammasome despite Casp8^-/- cells displaying unperturbed NLRP3 activation (Figure S4F). ASC specking was also abrogated in Nek7^-/- × Casp8^-/- mmMacs in response to LPS + Nigericin (Figure S4G), showing that caspase-8 deficiency affects NEK7-independent NLRP3 priming upstream of inflammasome formation. Since we found IKKβ to be crucial for the NEK7 bypass, we checked whether caspase-8 deficiency had an effect on IKKβ activity.³¹ Indeed, we observed reduced IKKβ phosphorylation after LPS stimulation of Casp8^-/- mmMacs (Figure S4H), suggesting that reduced IKKβ activity, rather than a specific role of caspase-8, explains the inability of Nek7^-/- × Casp8^-/-mmMacs to activate NLRP3 in response to LPS + Nigericin.

In contrast to ATP and Nigericin, which depend on K⁺ efflux to engage NLRP3, the TLR7 agonist Imiquimod (R837) has been shown to induce NEK7-dependent NLRP3 inflammasome formation independently of K⁺ efflux.⁹ In mmMac cells, Imiquimod strongly depended on NEK7 for NLRP3 activation even in combination with LPS (Figure 3A). Given that all K⁺ efflux-dependent stimuli tested here can engage the NEK7 bypass with concurrent IKKβ activation, we investigated whether K⁺ efflux might boost Imiquimod-driven NLRP3 activation in Nek7^-/- mmMacs. Indeed, under low extracellular K⁺ conditions that facilitate K⁺ efflux,¹⁰ Imiquimod stimulation together with LPS led to a NEK7-independent response that was significantly increased over LPS stimulation alone and not detectable with a physiological extracellular K⁺ concentration of 5 mM (Figures 3B and 3C). Although the relative contributions of LPS- or Imiquimod- induced IKKβ activity and K⁺ efflux- or Imiquimod-induced NLRP3 activation remain unclear, these data indicate that K⁺ efflux enhances the NEK7-bypassing effect of IKKβ activation.

Figure 3. K⁺ efflux works synergistically with IKKβ to bypass NEK7.

(A) mmMacs of the indicated genotypes were stimulated with the NLRP3 inflammasome inducers Nigericin or Imiquimod in the presence of LPS as indicated for 4 h (Nigericin) or 6 h (Imiquimod).

(B and C) mmMacs of the indicated genotypes were stimulated as indicated and imaged every 30 min (C) for 3 h in Hank’s balanced salt solution with or without potassium + 10% FCS and 5 mg/mL propidium iodide before LDH release was measured (B).

Data are represented as mean ± SEM with dots representing biological replicates conducted on separate days unless indicated otherwise. ***p < 0.001, **p < 0.01, *p < 0.05, ns p ≥ 0.05 calculated by two-way ANOVA followed by Tukey’s test.

Human myeloid cells use IKKβ instead of NEK7 to prime NLRP3 by default

Moving back into the human system, we wondered whether NLRP3 priming through IKKβ was also responsible for the NEK7-independence of NLRP3 activation in human cells. Using the hiPS- Mac model, we found that IKBKB^-/- cells showed a strong defect in NLRP3 inflammasome activation, whereas NAIP-NLRC4 activation proceeded normally, with IL-18 release being partially compromised (Figures S5A and S5B). However, we also observed a reduction in IL-6 amounts in IKKβ-deficient hiPS-Macs following LPS stimulation (Figure S5C). IKKβ, by governing NF-kB-dependent NLRP3 expression and also mediating the non-transcriptional NEK7 bypass, fulfills a dual role in NLRP3 priming. Hence, any effects on NLRP3 priming in IKBKB^-/- hiPS-Macs cannot unequivocally be ascribed to either transcriptional or non-transcriptional NLRP3 priming based on these experiments. Although these results establish that IKKβ is critical for NLRP3 priming in human cells, the relative contributions of transcriptional and non-transcriptional priming remain unclear in the hiPS-Mac model.

To clarify whether transcriptional or non-transcriptional NLRP3 priming is the predominant priming modality in the human system, we employed the BLaER1 model system. Given that hiPS-Macs express NLRP3 under steady-state conditions without transcriptional priming, we first sought to clarify if transcriptional priming was required for NLRP3 activation in BLaER1 cells. Although BLaER1 cells deficient in TAK1 (MAP3K7), in which NF-kB-mediated transcription after LPS sensing is completely abrogated, did indeed not produce pro-IL-1β upon LPS treatment anymore, they still expressed NLRP3 (Figure S5D). Congruently, blocking protein translation with CHX did not affect NLRP3 activation in these cells (Figure S5E). These data show that in BLaER1 cells, transcriptional priming is not required for NLRP3 inflammasome activation. Still, again mirroring hiPS-Macs, stimulation with Nigericin alone was not sufficient to activate NLRP3, but additional treatment with LPS was required to enable NLRP3 inflammasome formation in BLaER1 cells (Figure S5F). The NAIP-NLRC4 inflammasome formed in response to Needle Tox irrespectively of LPS priming as expected (Figure S5F). These data demonstrate that BLaER1 cells require non-transcriptional priming of NLRP3 for inflammasome activation. In line with these findings, a short pulse of concomitant LPS + Nigericin treatment led to robust NLRP3 activation in BLaER1 cells (Figure S5G). RIPK1, RIPK3, and caspase- 8 were dispensable for NLRP3 activation in response to Nigericin and NLRC4 activation, but in GSDMD^-/- BLaER1 cells, LDH release for both inflammasomes was blunted (Figure S5H).

Given that non-transcriptional priming was still dependent on TAK1 in BLaER1 cells and that TAK1 activates IKKβ, we then assessed NLRP3 activation in BLaER1 cells deficient for IKKβ. Corroborating our findings from hiPS-Macs and the murine system, LDH release and caspase-1 maturation following NLRP3 activation were blunted in IKBKB^-/- BlaER1 cells (Figures 4A, 4B, and S5I). In contrast, cells deficient in IKKα(CHUK), a close homolog of IKKβ, did not display a defect in inflammasome formation (Figure 4A). Cells deficient in both IKKα and IKKβ(CHUK^-- × IKBKB^-/-) phenocopied IKBKB^-/- cells (Figures 4A and 4B). As expected, given the steady-state expression of NLRP3 in BLaER1 cells, RELA^-/- × RELB^-/- cells displayed unperturbed NLRP3 activation (Figures 4A and S5I) despite strongly reduced pro-inflammatory cytokine transcription (Figure S5J). Reconstitution of IKBKB^-/- BlaER1 cells with wildtype IKKβ, but not IKKβ-K44M, a kinase-dead mutant of IKKβ,³² rescued NLRP3 activation, showing that the kinase activity of IKKβ was required for non-transcriptional NLRP3 priming (Figures 4C and 4D).

Figure 4. NLRP3 priming through IKKβ is required for inflammasome activation in human myeloid cell lines.

(A) LDH release from BLaER1 clones of the indicated genotypes primed with LPS for 2 h and subsequently treated with Nigericin or Needle Tox for 2 h.

(B) One representative of three immunoblots from cells treated as in (A). The asterisk denotes an unspecific band.

(C) Immunoblot of IKBKB^-/- BLaER1 cells expressing wild-type IKKβ or kinase-dead IKKβ-K44M under the control of a doxycycline-inducible promoter treated with doxycycline during the last 8 h of differentiation.

(D) LDH release from BLaER1 cells as in (C) primed with LPS for 2 h and subsequently treated with Nigericin or Needle Tox as indicated.

(E) LDH release from BLaER1 monocytes primed with LPS for 2 h before stimulation with Nigericin or Needle Tox. TPCA-1 was added at different time points as indicated.

Data are represented as mean ± SEM with dots representing biological replicates conducted on separate days. ***p < 0.001, **p < 0.01, *p < 0.05, ns p ≥ 0.05 calculated by two-way ANOVA followed by Dunnett’s test (A and E) or Šidák’s test (D).

See also Figure S5.

To investigate the kinetics of IKKβ-mediated non-transcriptional NLRP3 priming, we added the IKKβ inhibitor TPCA-1 to BLaER1 cells at different time points pre and post NLRP3 priming. Expectedly, adding TPCA-1 concurrently with LPS blocked all priming and abrogated NLRP3 activity (Figure 4E). However, adding TPCA-1 concurrently with or 30 min after Nigericin also blocked or strongly reduced NLRP3 activity, respectively (Figure 4E). Experiments with primary human monocytes corroborated these findings (Figure S5K). In summary, these data show that rapid, non-transcriptional priming by IKKβ is required for NLRP3 activation, further suggesting that human cells are NLRP3 inflammasome competent in the absence of NEK7 because they engage IKKβ by default.

Synergistically with IKKβ, NEK7 can accelerate NLRP3 activation human cells

Having demonstrated that IKKβ activation constitutes the predominant priming pathway in the human system, we wondered whether NEK7-mediated priming could be used by human cells at all. A hallmark of NEK7-mediated NLRP3 priming is the direct interaction of NEK7 and NLRP3.²¹ NLRP3 co-immunoprecipitated with NEK7 from THP-1 cells, indicating that the human NEK7 protein (hsNEK7) could in principle function to prime NLRP3 (Figures 5A and S6A). Of note, this interaction was independent of K⁺ efflux. We then reconstituted NLRP3 inflammasome signaling in HEK-293T cells, which normally do not express NLRP3 or ASC, the core signaling components of the NLRP3 inflammasome (Figures S6B and S6C). Notably, in this reconstitution system, inflammasome activation is driven by overexpression of NLRP3 and proceeds without stimulation by Nigericin. Hence, we consider inflammasome formation in this HEK-293T inflammasome assay to directly report the priming status of NLRP3. Here, we found that the mouse and human orthologs of NEK7 enhanced the activation of both NLRP3 orthologs, showing that hsNEK7 is capable of priming NLRP3 (Figures 5B and S6D). To investigate if NEK7 has a physiological role in NLRP3 priming, we went back to our hiPS-Mac system. Since we had found NLRP3 activation to require both NEK7 and LPS priming after concomitant LPS + Nigericin stimulation at early time points in mouse cells (Figure S3F), we tested the same condition in hiPS-Macs. Indeed, concomitant stimulation with LPS + Nigericin for 1 h resulted in NEK7-dependent release of LDH, whereas 4 h of LPS + Nigericin stimulation rendered NLRP3 activation NEK7-independent (Figures 5C and 5D).

Figure 5. NEK7 accelerates NLRP3 activation at early priming time points in iPSC-derived human macrophages.

(A) THP-1 cells were primed with Pam₃CSK₄ for 4 h and then stimulated with Nigericin for 30 min before lysates were immunoprecipitated with anti-NEK7 antibody or isotype control. One representative immunoblot of three independent experiments is shown.

(B) NEK7^-/- HEK293T cells were transiently transfected with plasmids driving expression of an ASC-RFP fusion protein and mouse or human orthologs of NLRP3 and NEK7 as indicated. ASC-RFP specks were imaged 24 h after transfection. Dots represent technical replicates from one representative of three independent experiments.

(C and D) Four clones per genotype of NEK7^-/- or wild-type human iPS cells were differentiated into hiPS-Macs and treated with Nigericin or LPS + Nigericin for 4 h or LPS + Nigericin for 1 h in the presence of the NLRP3 inhibitor MCC950 as indicated. Dots represent individual clones.

***p < 0.001, **p < 0.01, *p < 0.05, ns p ≥ 0.05 calculated by two-way ANOVA followed by Dunnett’s test (B) or Šidák’s test (C and D).

See also Figure S6.

From these data, we conclude that IKKβ, which is required to activate NLRP3 in all human cell lines tested here, operates in synergy with NEK7 to drive NLRP3 priming. NEK7 can accelerate NLRP3 priming at early time points, when IKKβ is not yet fully active. At later time points, IKKβ becomes redundant with NEK7.

Recruitment of NLRP3 to PtdIns4P induces NEK7- independent inflammasome activation Finally, we investigated how IKKβ activation enables NEK7-independent NLRP3 activation. As it has recently been reported that interaction of NLRP3 with phosphatidylinositol-4-phosphate (PI4P) on the TGN is an essential requirement for inflammasome formation,¹⁰ we investigated the subcellular localization of NLRP3 during priming. To this end we generated Pycard^-/- J774 mouse macrophages expressing a fusion protein of the PI4P-binding pleckstrin homology (PH)-domain of oxysterol- binding protein (OSBP) and mCherry (OSBP[PH]-mCherry). In these cells, we found LPS treatment to result in the accumulation of NLRP3 at PI4P-rich sites (Figures 6A and 6B). Of note, this translocation cannot be caused by NLRP3-mediated pyroptosis, since Pycard^-/- cells are incapable of NLRP3 inflammasome formation. The recruitment of NLRP3 to PI4P was markedly reduced by the IKKβ inhibitor TPCA-1 (Figures 6A and 6B). In line with our findings on LPS-dependent non-transcriptional priming in human and mouse cells, NLRP3 recruitment to PI4P occurred rapidly, generally within 30 min after LPS stimulation (Figure S6E). We did not observe NLRP3 translocation to mitochondria — in fact, PI4P-rich sites appeared mostly distinct from mitochondria (Figure S6F). To identify the cellular compartment that NLRP3 is recruited to, we fractionated lysates of Pycard^-/- J774 cells. Post-nuclear lysates were centrifuged at 5,000 × g to obtain a pellet (P5) and supernatant (S5) fraction. The S5 fraction was further subjected to centrifugation at 100,000 × g to yield a pellet (P100) and supernatant (S100) fraction. We found NLRP3 in all fractions irrespectively of LPS priming or concomitant IKKβ inhibition (Figure 6C). However, when we further fractionated P100 across a linear sucrose gradient, we found NLRP3 to become enriched in the top fractions upon LPS stimulation, where we also found the PI4P-binding OSBP(PH)-mCherry fusion protein (Figure 6D). This enrichment of NLRP3 was blocked in the presence of TPCA-1, and, in line with our imaging data, unstimulated cells showed some NLRP3 enrichment on both ends of the gradient. Of note, the mitochondrial membrane protein TOMM40 was also present in the P100 fraction, but at the opposite end of where the OSBP(PH)-mCherry construct was found. We then analyzed the organelles present in fractions #2 and #11 via mass spectrometry (Table S2). The TGN, but not the cis-Golgi network, was highly enriched in fraction #2 along with weakly PI4P⁺ organelles such as endosomes³³ (Figures 6E and S6G). Taken together, upon priming, NLRP3 translocates to PI4P-rich sites mostly on the TGN.

Figure 6. IKKβ-mediated recruitment of NLRP3 to PI4P enables NEK7-independent inflammasome formation.

(A) Nlrp3^-/- × Pycard^-/- J774 cells of the indicated Nek7 genotypes expressing mCherry tethered to phosphatidylinositol-4-phosphate (PI4P) via the PH domain of OSBP (OSBP(PH)-mCherry) and doxycycline-inducible mVenus-mmNlrp3 were treated with doxycycline for 24 h and TPCA-1 for 1 h before stimulation with LPS for 30 min. Scale bars represent 10 μm.

(B) Quantification of at least 10 randomly chosen fields of view per experimental condition from three independent experiments described in (A). Data are represented as mean ± SEM with dots representing biological replicates conducted on separate days.

(C) Lysates of J774 cells pretreated with TPCA-1 for 1 h and then stimulated with LPS for 30 min were depleted of nuclei (5 min 1,000 × g), and the supernatant was then centrifuged at 5,000 × g for 10 min (pellet P5) followed by 100,000 × g for 20 min (pellet P100, supernatant S100) before immunoblotting. One representative of three independent experiments is shown.

(D) P100 fractions from (C) were further fractionated across a linear sucrose gradient (20%–60%) into 12 fractions which were then immunoblotted. One representative of three independent biological replicates is shown.

(E) Enrichment of organelle-specific protein sets identified via mass spectrometry analysis of the protein content of fractions #2 and #11. p-values for set enrichment were calculated based on proteins differing between the two fractions (FC R 1.5, FDR < 0.05) using Fisher’s exact test with Benjamini-Hochberg correction.

(F) Nlrp3^-/- J774 cells of the indicated Nek7 genotypes expressing doxycycline-inducible variants of Nlrp3 as indicated were treated with doxycycline for 18 h followed by 2 h of Nigericin before immunoblotting.

***p < 0.001, **p < 0.01, *p < 0.05, ns ≥ 0.05 calculated by two-way ANOVA followed by Tukey’s test unless indicated otherwise.

See also Figure S6 and Table S2.

Based on these data, we hypothesized that the accumulation of NLRP3 on PI4P-rich sites induces NEK7-independent NLRP3 activation. To confirm this hypothesis, we directly tethered NLRP3 to PI4P by fusing it to the PH-domain of OSBP as reported before.¹⁰ Although a previously described K127A, K128A, K129A, and K130A quadruple mutant of mmNlrp3 (Nlrp3(4KA)) was incapable of localizing to the TGN in J774 cells, Nlrp3(4KA-OSBP(PH)) constitutively localized to the TGN as expected (Figure S6H). When we expressed wild-type Nlrp3, Nlrp3(4KA), and Nlrp3(4KA-OSBP(PH)) in Nlrp3^-/- J774 mouse macrophages, we found wild-type Nlrp3 to facilitate caspase-1 maturation in a NEK7-dependent manner and Nlrp3(4KA-OSBP(PH)) to activate caspase-1 independently of NEK7 (Figure 6F). Nlrp3(4KA) expectedly did not lead to any detectable caspase-1 processing (Figure 6F). Of note, these cells did not require priming with LPS, as they expressed Nlrp3 under the control of a doxycycline-inducible promoter, mirroring above results (Figure 2).

Together, these results demonstrate that IKKβ induces NEK7- independent NLRP3 priming by increasing the recruitment of NLRP3 to PI4P and establish PI4P-recruitment of NLRP3 as a priming modality of the inflammasome (Figure S6I).

Discussion

Since its first description in 2001,³⁴ NLRP3 has attracted much attention as a key driver of antimicrobial and sterile inflammation.⁷ Nonetheless, despite being in the focus for almost two decades, the molecular mechanism of NLRP3 activation has remained obscure. The two-step model of inflammasome priming and activation predates the discovery of NLRP3 and inflammasomes altogether, originating from the notion that both a pro-inflammatory and a cell-death inducing signal are required to release mature IL-1β from murine bone marrow-derived macrophages.³⁵ In retrospect, these early studies had assessed NLRP3 inflammasome activation employing a K⁺ efflux-inducing trigger. Subsequent studies have revealed that the pro-inflammatory signal indeed serves two independent functions in the context of NLRP3 inflammasome activation. Although this signal is critically required to induce pro-IL-1β expression, it is also necessary to render NLRP3 activatable in the first place. This became apparent when studying the maturation of caspase-1, the expression of which is independent of a pro-inflammatory signal, as a proxy of NLRP3 inflammasome activation. Here, it has been revealed that unprimed macrophages do not mature caspase-1 upon K⁺ efflux-inducing stimuli^13,36 but that additional priming by a pro-inflammatory signal was required to facilitate this step. Of note, this unique requirement of NLRP3 priming by a pro-inflammatory signal (referred to as signal 1 or priming in this manuscript) must not be confused with the signal that induces pro-IL-1β expression. Indeed, although both signals can be provided through the same PRR, they can also be separated, and the pro-IL-1β inducing stimulus is not necessary for NLRP3 inflammasome activation.

Although the two-step activation model constitutes an important conceptual framework for NLRP3 activation, it has proven to be an enormous conundrum because it is not trivial to allocate signaling events upstream of NLRP3 to either priming or activation. The fact that several pathways toward NLRP3 priming have been described³⁷ is likely attributable to stimulus-, cell type-, and species-dependent aspects as well as temporal dynamics playing an important role in this context. We conceptualize that priming serves the function to increase the cellular pool of NLRP3 molecules that are able to respond to an activating stimulus, either by upregulating production of the NLRP3 protein or by lowering the activation threshold of individual NLRP3 molecules. In this regard, we would interpret the existence of multiple redundant NLRP3 priming pathways as the possibility to integrate diverse pro-inflammatory inputs to achieve this activatable state. In fact, we consider this pleiotropy to be a key trait of NLRP3 priming, but not activation pathways. The mitotic spindle kinase NEK7 has been shown to be an essential cofactor of NLRP3 activation,^20–22 and it has been suggested that NEK7 facilitates inflammasome formation by mediating recognition of the second signal.^21,23 Studying the role of NEK7 in iPS-cell-derived human macrophages, we made the unexpected discovery that NLRP3 activation can be fully operational in the absence of NEK7. By genetically dissecting NLRP3 inflammasome signaling, we uncovered that these cells employ a NEK7-independent signaling cascade instead that drives IKKβ-dependent, post-translational priming of NLRP3. Although this IKKβ-dependent priming signal is the default pathway by which human cells engage the NLRP3 inflammasome, murine macrophages predominantly rely on NEK7 for NLRP3 priming. However, they can bypass NEK7 and switch to IKKβ-dependent priming under pro-inflammatory conditions signified by, for example, TLR activation. The NEK7-independence in human myeloid cells could not be attributed to species-specific constitutions of the NEK7 or NLRP3 molecules themselves: immunoprecipitation and reconstitution experiments showed that human NEK7 interacted with human NLRP3 and that NEK7 was able to facilitate NLRP3 activity. In line with this notion, iPSC-derived human macrophages also employ NEK7 to activate the NLRP3 inflammasome; however, this requires LPS priming and indicates a synergy between NEK7 and IKKβ only observed at an early time point, when the IKKβ post-translational priming mechanism is not yet fully operational. Indeed, in these cells, NEK7 becomes obsolete after prolonged LPS-priming when the IKKβ priming cascade is active. Mechanistically, IKKβ activity recruited NLRP3 to PI4P, a phospholipid enriched on the TGN. Tethering NLRP3 to PI4P led to inflammasome activation independently of NEK7, confirming that increased PI4P interaction serves to prime NLRP3 for inflammasome formation. Based on the redundancy between IKKβ and NEK7 in facilitating NLRP3 inflammasome formation, we conclude that NEK7 serves as a priming factor of the NLRP3 inflammasome.

NEK7 holds a unique position among NLRP3 priming pathways in that it is constitutively expressed and apparently uncoupled from upstream signals in its pro-inflammatory capacity. It has been suggested that NEK7 is employed for NLRP3 activation to avoid inflammasome formation during mitosis, when NEK7 is not available.²² Furthermore, it has been speculated that the cellular perturbation triggering NLRP3 commonly occurs during mitosis, and thus, the dependency on NEK7 prevents inadvertent inflammasome activation during cell division.²³ However, the here-uncovered redundancy of NEK7 priming with other cell cycle-independent priming pathways (e.g., IKKβ) advocates against a specific de-coupling of NLRP3 inflammasome activation and proliferation. This is also in line with the fact that many NLRP3 inflammasome-competent cells of the innate immune system are postmitotic. As such, despite detailed mechanistic insight into how NEK7 can accelerate NLRP3 inflammasome activation, the physiological role of NEK7 remains to be determined. The redundancy of NEK7 with a priming factor that acts by enhancing the interaction of NLRP3 and PI4P suggests that NEK7 itself might be involved in recruiting NLRP3 to PI4P at the TGN.

The role of K⁺ efflux is currently debated in the field: although it was recently shown that K⁺ efflux alone is not sufficient to drive inflammasome activation in primed BMDMs and consequently argued that K⁺ efflux only promotes recruitment of NLRP3 to the TGN,¹⁰ an older report demonstrated that K⁺ efflux does indeed suffice: inflammasome activation did occur in response to K⁺ efflux in primed BMDMs.⁶ Our study shows that recruitment of NLRP3 to PI4P can be induced by IKKβ activation independently of K⁺ efflux. In line with the latter report, K⁺ efflux was still required for inflammasome formation following IKKβ-mediated PI4P recruitment of NLRP3, hinting at a role of K⁺ efflux beyond recruiting NLRP3 to PI4P. Whether K⁺ efflux or dispersal of the TGN serves as the ultimate trigger of NLRP3 inflammasome formation remains to be investigated. From the fact that both IKKβ- and NEK7-mediated NLRP3 priming still require K⁺ efflux for inflammasome formation but that IKKβ-mediated priming can bypass NEK7, we conclude that NEK7 itself acts as a priming factor upstream of K⁺ efflux. Of note, K⁺ efflux-independent NLRP3 activators have also been described.^8,9 For one such agonist, Imiquimod, the NEK7 bypass was only activated in the presence of K⁺ efflux, suggesting that K⁺ efflux boosts NEK7-independent NLRP3 activation synergistically with IKKβ.

Another study recently implicated IKKβ in the recruitment of NLRP3 to the TGN.³⁸ In contrast with our findings, in their setting, Nigericin stimulation was still required for TGN recruitment of NLRP3, as reported previously.¹⁰ The authors concluded that IKKβ enhances Nigericin-dependent TGN dispersal, which they suggested to be the cause of increased NLRP3 activity.³⁸ However, whether increased TGN dispersal is a cause or an effect of increased cell death cannot be concluded from their work. In our study, we observed that IKKβ activation recruited NLRP3 to PI4P on an undispersed TGN independently of Nigericin stimulation or TGN dispersal in pyroptosis-deficient Pycard^-/-cells. We showed that recruiting NLRP3 to an intact TGN was sufficient for subsequent inflammasome formation independently of an additional priming stimulus. Hence, it is unlikely that the increased TGN dispersal observed by Nanda and colleagues would explain the priming effect of IKKβ that we describe here. Rather, given that we observe K⁺ efflux to act synergistically with IKKβ and NEK7, increased recruitment of NLRP3 to the TGN might explain the previously reported effects.³⁸

This study establishes NEK7 as a priming rather than an activation signal for NLRP3. Moreover, in its capacity as a priming factor NEK7 does not constitute an absolute requirement for NLRP3 inflammasome activation. Instead, a priming signal emanating from IKKβ can fully compensate for NEK7 by enhancing the interaction of NLRP3 and PI4P. This signal supersedes the NEK7 requirement in human myeloid cell lines and also represents the dominant priming entity in iPSC-derived human macrophages.

Limitations of the study

We have shown that NEK7-independent priming of NLRP3 depends on the kinase activity of IKKβ but does not require de novo translation. However, the target that is phosphorylated by IKKβ remains to be determined in future studies. To confirm that recruitment of NLRP3 to PI4P is sufficient for NEK7-independent inflammasome activation, we overexpressed an engineered fusion protein of NLRP3(4KA) and the PI4P-interacting PH domain of the protein OSBP that constitutively interacts with PI4P, as reported previously.¹⁰ Although we controlled for unspecific NLRP3 activation by expressing NLRP3 fused to a non-PI4P-binding point- mutated version of the same PH domain, we cannot exclude that engineering NLRP3 influenced its dependency on NEK7. Finally, owing to the fact that Nek7^-/- mice are not viable,³⁹ this study does not include an experiment showing that IKKβ activation bypasses NEK7 in vivo.

Methods

Key Resources Table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
anti-Caspase-1 (p20) (human), mAb (Bally-1)	AdipoGen, San Diego, CA	Cat# AG-20B-0048-C100
anti-Caspase-1 (p20) (mouse), mAb (Casper-1)	AdipoGen	Cat# AG-20B-0042-C100
anti-NEK7	Abcam, Cambridge, UK	Cat# ab133514
anti-NLRP3/NALP3, mAb (Cryo-2)	AdipoGen	Cat# AG-20B-0014-C100
anti-Human IL-1 beta /IL-1F2	R&D Systems Inc,	Cat# AF-201-NA
	Minneapolis, MN
Chemicals, peptides, and recombinant proteins
1-Thioglycerol (MTG)	Sigma-Aldrich, St. Louis, MO	Cat# M6145
Accutase	Stemcell Technologies,	Cat# 07920
	Vancouver, Canada
Adenosine 5’-triphosphate disodium salt hydrate	Sigma-Aldrich	Cat# A6419
Ascorbic Acid	Sigma-Aldrich	Cat# A4544-100G
B-27 supplement	Thermo Fisher Scientific,	Cat# 17504-001
	Waltham, MA
Blasticidin S HCl (10 mg/ml)	Thermo Fisher Scientific	Cat# A1113903
BSA	GE Healthcare, Chicago, IL	Cat# SH30574.01
CHIR99021	Miltenyi Biotec, Bergisch Gladbach, Germany	Cat# 130-103-926
Cycloheximide	Carl Roth, Karlsruhe, Germany	Cat# 8682.1
Doxycycline hyclate	Sigma-Aldrich	Cat# D9891-1G
Geltrex	Thermo Fisher Scientific	Cat# A1413302
GeneJuice	Merck, Darmstadt, Germany	Cat# 70967-3
Ham’s F12 nutrient mix	Thermo Fisher Scientific	Cat# 21765029
Herring Testis(HT)-DNA sodium salt	Sigma Aldrich	Cat# D6898
Hoechst-33342	Sigma-Aldrich	Cat# B2261-25MG
Human CSF-1 (M-CSF) (iPSC differentiation)	R&D Systems	Cat# 216-MC-005
Human Transferrin	Roche, Basel, Switzerland	Cat# 10-652-202-001
IMDM with GlutaMAX	Thermo Fisher Scientific	Cat# 31980022
Imiquimod (R837)	Invivogen	Cat# tlrl-imq
L-Glutamine	Thermo Fisher Scientific	Cat# 25030024
LFn-YscF	Rauch et al.40	N/A
Lipofectamine 2000 Transfection Reagent	Thermo Fisher Scientific	Cat# 11668019
LPS-EB Ultrapure	Invivogen, San Diego, CA	Cat# tlrl-3pelps
LysC	Wako	Cat# 12902541
MCC950	Sigma-Aldrich	Cat# PZ0280
MitoTracker DeepRed	Thermo Fisher Scientific	Cat# M22426
mTeSR1	Stemcell Technologies	Cat# 85850
N-2 Supplement	Thermo Fisher Scientific	Cat# 17502048
Nigericin sodium salt	Sigma-Aldrich	Cat# N7143
Pam3CSK4	Invivogen	Cat# tlrl-pms
Phorbol 12-myristate 13-acetate	ENZO Life Sciences, Farmingdale, NY	Cat# BML-PE160-0005
Protective antigen (pA)	Biotrend, Cologne, Germany	Cat# LL-171E
Puromycin Dihydrochloride	Carl Roth	Cat# 0240.4
R848	Invivogen	Cat# tlrl-r848-5
Recombinant Human BMP-4	R&D Systems	Cat# 314-BP-010
Recombinant Human CSF-1 (M-CSF) (BlaER1 differentiation)	Recombinantly produced	N/A
Recombinant Human DKK-1	R&D Systems	Cat# 5439-DK-010
Recombinant Human FGF2	R&D Systems	Cat# 233-FB-025
Recombinant Human IL-3	R&D Systems	Cat# 203-IL-010
Recombinant Human IL-3 (BLaER1 differentiation)	Recombinantly produced	N/A
Recombinant Human IL-6	R&D Systems	Cat# 206-IL-010
Recombinant Human SCF	R&D Systems	Cat# 255-SC-010
Recombinant Human VEGF	R&D Systems	Cat# 293-VE-010
ROCK Inhibitor Y-27632	Stemcell Technologies	Cat# 72302
Stempro-34 SFM	Thermo Fisher Scientific	Cat# 10639-011
Takinib	Selleck Chemicals, Houston, TX	Cat# S8663
TPCA-1	R&D Systems	Cat# 2559/10
Trypsin	Sigma-Aldrich	Cat# T6567
b-Estradiol	Sigma-Aldrich	Cat# E8875
Critical commercial assays
Human IL-1β ELISA Set II	BD Biosciences, San José, CA	Cat# 557953
Human Total IL-18 DuoSet ELISA	R&D Systems	Cat# DY318-05
MiSeq Reagent Kit v2, 300 Cycles	Illumina, San Diego, CA	Cat# MS-102-2002
Mouse CXCL10/IP-10/CRG-2 DuoSet ELISA	R&D Systems	Cat# DY466
Mouse TNF (Mono/Mono) ELISA Set II	BD Biosciences	Cat# 558534
OptEIA Human IL-6 ELISA Set	BD Biosciences	Cat# 555220
OptEIA Mouse IL-1β Elisa Set	BD Biosciences	Cat# 559603
Pierce LDH Cytotoxicity Assay Kit	Thermo Fisher Scientific	Cat# 88954
Deposited data
Mass spectrometry data of Figure 6E	This study	PRIDE: PXD035302
Immunoblot source data and raw numerical data used to plot the figures	This study	Mendeley data: https://doi.org/10.17632/h7vc8hnb7j.1
Experimental models: Cell lines
BLaER1	Rapino et al.26	N/A
HEK-293T	Cavlar et al.41	N/A
iPSC	Camargo Ortega et al.42	N/A
Mouse Macrophages, Nlrp3, Asc-CFP, Cas9-expressing	Franklin et al.43	N/A
THP-1	ATCC, Manassas, VA	Cat# TIB-202
Target sites of sgRNAs used in this study
hsMAP3K7	GTAAACACCAACTCATTGCGTGG
hsNEK6	GTCTTTTCGCTGCTCGCTGGCGG
hsNEK7	ATTACAGAAGGCCTTACGACCGG
hsNLRP3	GCTAATGATCGACTTCAATGGGG
hsIKBKB	ATGAAGGTATCTAAGCGCAGAGG
mmMyd88	GGTTCAAGAACAGCGATAGGCGG
mmNek7	GTCTCTTGGATGGAGTGCCGG
mmNlrp3	CCTCTCTGCTCATAACGACGAGG
mmTicam1	GTACAGGCGAGCCACCGTCCAGG
mmTlr4	GATCTACTCGAGTCAGAATGAGG
mmPycard	GTGCAACTGCGAGAAGGCTATGG
Recombinant DNA
LentiCas9-Blast	Sanjana et al.44	N/A
LentiGuide-Puro	Sanjana et al.44	N/A
pBabe-U6-sgRNA-Cas9	Schmidt et al.45	N/A
pBlast-hsNEK7	This study	N/A
pBlast-mCherry-OSBP(PH)	This study	N/A
pBlast-mmNek7	This study	N/A
pLIX-hsNLRP3	This study	N/A
pLIX-mmNlrp3	This study	N/A
pLIX-mVenus-mmNlrp3	This study	N/A
pLIX-mVenus-mmNlrp3(4KA)	This study	N/A
pLIX-mVenus-mmNlrp3(4KA-OSBP(PH))	This study	N/A
pLK0.1-sgRNA-CMV-GFP	Schmid-Burgk et al.46	N/A
pRP-Asc-RFP	This study	N/A
prZ-CMV-Cas9	Schmid-Burgk et al.45	N/A
Software and algorithms
CellProfiler 3.1.5	Carpenter et al.47	https://cellprofiler.org
CHOPCHOP	Labu et al.48	https://chopchop.cbu.uib.no
MaxQuant 2.0.3	Cox and Mann49	https://maxquant.org
Outknocker	Schmid-Burgk et al.46	http://www.outknocker.org
Perseus	Tyanova et al.50	https://maxquant.org/perseus/
Prism 9.0	GraphPad, San Diego, CA	https://www.graphpad.com/scientific-software/prism/

Experimental Model and Subject Details

BLaER1 cells

BLaER1 cells (female) were cultivated in RPMI supplemented with 10 % FCS, 1 mM pyruvate, 100 U/ml penicillin and 100 mg/ml streptomycin at 37 °C and 5 % CO₂. BLaER1 cells were differentiated in medium containing 10 ng/ml hrIL-3, 10 ng/ml hrCSF-1 (MCSF) and 100 nM b-estradiol for 5-6 days. In the course of these studies, we serendipitously identified that BLaER1 cells express transcripts of SMRV (squirrel monkey retrovirus) and subsequent experiments confirmed that BLaER1 cells harbor the SMRV proviral genome. Testing early passages of BLaER1 cells by Dr. Thomas Graf (personal communication) confirmed that the parental BLaER1 cell line²⁶ is positive for SMRV. Of note, extensive characterization of BLaER1 monocytes in comparison to other human myeloid cells has not provided any indication that SMRV positivity would impact on the functionality of these cells as myeloid cells. Samples of other cell lines used in this work were confirmed to be free of SMRV by PCR. All BLaER1 cell experiments were conducted on a CASP4^-/- background (herein referred to as control).

THP-1 cells

THP-1 cells (male) were obtained from ATCC and cultivated in RPMI supplemented with 10 % FCS, 1 mM pyruvate, 100 U/ml penicillin and 100 mg/ml streptomycin at 37 °C and 5 % CO₂. THP-1 cells were differentiated by adding 100 ng/ml PMA to the medium for 18 hours, rinsed off with ice-cold PBS and replated for experiments.

mmMacs and J774 mouse macrophages

Mouse macrophages were cultivated in DMEM supplemented with 10 % FCS, 1 mM pyruvate 100 U/ml penicillin and 100 mg/ml streptomycin at 37 °C and 5 % CO₂. mmMacs were detached for passaging with 0.05 % Trypsin at 37 °C for 15 minutes after one PBS wash and then rinsed off with DMEM. J774 cells were passaged by scraping in 5 ml fresh DMEM and transferred to new flasks.

hiPSC, hiPS-Macs cell culture

Human induced pluripotent stem cells (hiPSCs) used to make NEK7^-/- hiPSCs were kindly provided by Adam O’Neill and Magdalena Götz.⁴² hiPSCs for IKBKB^-/- were purchased from XCell Science. hiPSCs were cultivated on Geltrex-coated plates in complete mTeSR1 Medium at 37 °C and 5 % CO₂ and detached for passaging using 1.5 ml Accutase for 5 minutes at 37 °C after a PBS wash. After passaging, cells were cultivated in the presence of 5 mM ROCK-Inhibitor overnight.

Differentiation of hiPSCs into hiPS-Macs

Differentiation into iPS-Macs was achieved as described previously.²⁵ Briefly, 150,000 hiPSC were plated into a one well of a Geltrex- coated 6-well plate and differentiated in StemPro base medium with StemPro Supplement, 200 mg/ml human transferrin, 2 mM gluta- mine, 0.45 mM MTG and 0.5 mM ascorbic acid (= StemPro medium, ascorbic acid was added just before use) by stimulation with 50 ng/ml VEGF, 5 ng/ml BMP-4 and 2 mM CHIR99021 at 5 % oxygen for two days, followed by two days of stimulation with 50 ng/ml VEGF, 5 ng/ml BMP-4 and 20 ng/ml FGF2. From day four, StemPro medium was supplemented with 15 ng/ml VEGF and 5 ng/ml FGF2. Starting at day six, 10 ng/ml VEGF, 10 ng/ml FGF2, 50 ng/ml SCF, 30 ng/ml DKK-1, 10 ng/ml IL-6 and 20 ng/ml IL-3 were added to StemPro medium until day ten. From day eight, cells were cultivated under normoxic conditions. From day twelve, 10 ng/ml FGF2, 50 ng/ml SCF, 10 ng/ml IL-6 and 20 ng/ml IL-3 were added to StemPro medium. Starting at day sixteen, cells were cultivated in 75 % IMDM with 25 % F12 supplement, N2 supplement, B-27 supplement, 0.05 % BSA and 100 U/ml penicillin and 100 mg/ml streptomycin (= SF-Diff medium) supplemented with 50 ng/ml rhCSF-1 (M-CSF) at least until day 28. Culture medium was exchanged as necessary, but at least every two days. After differentiation, hiPS-Macs were carefully harvested from the supernatant, spun down and replated in RPMI with 10 % FCS, 1 mM Pyruvate, 100 U/ml Penicillin and 100 mg/ml Streptomycin for experiments.

HEK-293T cells

HEK-293T cells were cultivated in DMEM with 10 % FCS, 1 mM pyruvate, 100 U/ml penicillin and 100 mg/ml streptomycin at 37 °C and 5 % CO₂. For passaging, cells were washed with PBS once and then incubated with 0.05 % Trypsin at 37° for 5 minutes. Cells were then rinsed off with DMEM.

Trans-Golgi network imaging

J774 macrophages expressing mVenus-mmNlrp3 and the PH-domain of hsOSBP (OSBP-PH) fused to mCherry were plated in ibidi 8-well slides (100,000 per well in 200 ml of DMEM) and imaged on a Nikon Eclipse Ti spinning disk confocal microscope with 100× magnification on the following day. Results were manually quantified from at least 10 randomly selected areas per condition per replicate using FIJI.⁵¹ For nuclear staining, Hoechst-33342 was diluted to a final concentration of 10 mg/ml.

ASC speck imaging

ASC specks in transiently transfected HEK-293T cells were imaged 24 hours after transfection on a Leica Hi8 epifluorescence microscope using 10× magnification. Specks were quantified with CellProfiler.⁴⁷

Immunoblotting

Cells were lysed at approximately 5 Mio/ml in 1× Lämmli Buffer and boiled for 5 minutes at 95 °C. For precipitation of total protein from supernatants, stimulations were done in medium containing 3% FCS. Precipitation of total protein from supernatants was achieved by combining 700 ml of supernatant with 700 ml MeOH and 150 ml of CHCl₃. Samples were spun down at 20.000 g for 20 minutes, and the upper phase was discarded. Again, 700 ml MeOH were added and samples were centrifuged at 20.000 g for 20 minutes. The pellet was then dried and resuspended in 100 ml 1× Lämmli buffer and boiled at 95 °C for 5 minutes. Samples were run on 12% SDS-PAGE gels at 150 V for 85 minutes and were subsequently transferred onto a nitrocellulose membrane at 100 V for 75 minutes at 4 °C. Membranes were then blocked in 5 % milk for 1 hour at room temperature. Primary and secondary antibodies were diluted in 1-5 % milk.

ELISA and LDH assay

LDH assays were done on supernatants immediately after experiments. Results are presented relative to a lysis control from the same experiment with the values of unstimulated controls subtracted as background. ELISAs were done according to manufacturer’s instructions on supernatants stored at -20 °C.

Stimulation of immune signaling

NLRP3 was primed as indicated with 1 mg/ml Pam₃CSK₄ or 200 ng/ml LPS. NLRP3 was activated with 5 mM ATP or Nigericin at

6.5 mM (BLaER1 cells) or 10 mM (all other cells) as indicated. To activate the AIM2 inflammasome 400 ng HT-DNA were transfected into a 96-well with 0.5 ml Lipofectamine in 50 ml OptiMEM by incubating OptiMEM and Lipofectamine for 5 minutes followed by 20 minutes of incubation of the Lipofectamine-DNA mix in OptiMEM and dropwise addition of the mix to the cells. For immunoblots, transfections were done in a 12-well format. The amount of Lipofectamine and HT-DNA was scaled accordingly by well area. The NAIP- NLRC4 inflammasome was activated with an anthrax toxin lethal factor fused to the Burkholderia T3SS needle protein (LFn-YscF, 0.025 mg/ml), which was delivered into cells with protective antigen (pA, 0.25 mg/ml).⁴⁰ If not otherwise indicated, cells were stimulated with this construct (herein referred to as Needle Tox) for 2 hours.

Inhibition of translation

For mmMacs, cycloheximide (CHX) was added to the medium 30 minutes before stimulation to a final concentration of 10 mg/ml. For BLaER1 cells, CHX was added to the medium simultaneously with LPS at the indicated concentrations in the range of 1-10 mg/ml.

Doxycyclin-inducible gene expression

In BLaER1 cells and J774 Pycard^-/- cells transduced with pLIX-Puro derived vectors, gene expression was induced by adding medium to a final concentration of 1 mg/ml doxycycline for the last 24 hours of differentiation. J774 cells transduced with Nlrp3 variants for analysis of caspase-1 processing were stimulated with 1 mg/ml doxycycline for 18 hours before stimulation for inflammasome activation.

Inhibition of TAK1, IKKβ and NLRP3

Takinib was added to a final concentration of 50 mM as indicated. TPCA-1 was used at 5 mM final concentration as indicated. MCC950 was added as indicated to a final concentration of 10 mM.

Inhibition and induction of K⁺ efflux

To block K⁺ efflux, Potassium chloride (KCl) was added to medium together with the priming stimulus to the indicated final concentrations. The osmolarity of the medium was kept constant over all conditions. To induce K⁺ efflux, cells were stimulated in sterile Hank’s balanced salt solution with (140 mM NaCl, 5 mM KCl, 1.3 mM CaCl₂, 1.0 mM MgSO₄, 10 mM HEPES (pH 7.5), 5.5 mM glucose) or without potassium (145 mM NaCl, 1.3 mM CaCl₂, 1.0 mM MgSO₄, 10 mM HEPES (pH 7.5), 5.5 mM glucose) with 10% FCS as described before.¹⁰

Sucrose gradient fractionation

For the fractionation experiment, Nlrp3^-/- × Pycard^-/- J774 cells stably transduced with pLI-mVenus-mmNLRP3, pBlast-AUG- OSBP(PH)-mCherry were used.

Two days prior to stimulation, 1310⁷ cells were plated per 15 cm dish, using 2 dishes per condition (unstimulated, LPS, TPCA- 1 > LPS). 18-20 hours prior to stimulation, doxycycline was added to a final concentration of 1 mg/ml to induce expression of mVenus-mmNLRP3. As indicated, cells were pre-treated with 5 mM TPCA-1 for 30 minutes. Subsequently, cells were stimulated with 200 ng/ml LPS for 30 minutes. Cells were washed once with PBS and scraped using 500 mL of ice-cold isotonic buffer (0.25 M sucrose, 10 mM Tris-HCl (pH 7.5), 10 mM KCl, 1.5 mM MgCl₂, 1 mM DTT) supplemented with protease inhibitor. Then, cells were homogenized by performing 30 strokes with a 29G needle (VWR, BDAM324891). Lysates were centrifuged at 10003g for 5 minutes at 4°C to remove nuclei and any remaining cells. The resulting supernatant was centrifuged at 50003g for 10 minutes at 4°C to obtain the heavy membrane fraction (pellet, P5). The resulting supernatant was centrifuged at 100,0003g for 20 minutes at 4°C in a TLA 120.2 rotor (Beckman Coulter) to obtain the light membrane fraction (pellet, P100) and the cytosolic fraction (supernatant, S100). The fractions P5 and P100 were washed once with isotonic buffer, pelleted repeating the centrifugation step at 50003g and 100,0003g, respectively, and resuspended in 500 mL isotonic buffer.

The fraction P100 was then loaded onto a 20%-60% continuous sucrose density gradient (10 mM Tris-HCl (pH 7.5), 100 mM KCl, 1.5 mM MgCl₂, 1 mM DTT, and protease inhibitor cocktail). The gradients were centrifuged in an SW40Ti rotor (Beckman Coulter) at 170,0853g for two hours at 4°C and 13 fractions of 0.93 ml each were collected using a BioComp Gradient Station. 30 mL of each fraction were used for SDS-PAGE followed by immunoblotting.

Furthermore, to analyze the distribution of various organelle markers, the fractions P5, P100 and S100 were subjected to SDS- PAGE followed by immunoblotting. Protein concentrations were determined by BCA assay and adjusted between samples (unstimulated, LPS, TPCA-1 > LPS) for each of the fractions separately.

Mass spectrometry sample preparation

Sucrose gradient fractions #2 and #11 were lysed in 1% SDC with 100mM Tris-HCl. Protein amounts from each sample were adjusted to 30 mg with a BCA protein assay kit. Samples were reduced with 10mM tris(2-carboxy(ethyl)phosphine) (TCEP), alkylated with 40mM 2-chloroacetamide (CAA), and digested with trypsin and lysC (1:50, enzyme/protein, w/w) overnight. Digested peptides were desalted using SDB-RPS-stage tips. Desalted peptides were resolubilized in 5ml 2% ACN and 0.3% TFA and about 200 ng of peptides were injected into the mass spectrometer.

Samples were loaded onto 50-cm columns packed in-house with C18 1.9mM ReproSil particles (Dr. Maisch GmbH), with an EASY- nLC 1200 system (Thermo Fisher Scientific) coupled to the MS (Orbitrap Exploris 480, Thermo Fisher Scientific). A homemade column oven maintained the column temperature at 60°C. Peptides were introduced onto the column with buffer A (0.1% formic acid) and were eluted with a 120-min gradient starting at 5% buffer B (80% ACN, 0.1% formic acid) followed by a stepwise increase to 30% in 95 min, 65% in 5 min, 95% in 235 min and 5% in 235 min at a flow rate of 300 nL/min. Samples were measured in data-dependent acquisition with a TopN MS method in which one full scan (300–1650 m/z, R=60,000 at 200m/z) at an Automatic Gain Control (AGC) target of 3310e6 ions was first performed, followed by 15 data-dependent MS/MS scans with higher-energy collisional dissociation (AGC target 1×10e5 ions, maximum injection time at 25ms, isolation window 1.4 m/z, normalized collision energy 30%, and R=15,000 at 200 m/z). Dynamic exclusion of 30 s was enabled.

Analysis of MS samples

The MS raw files were processed in MaxQuant version 2.0.3.0⁴⁹ and fragment lists were queried against the mouse UniProt FASTA database (25,320 entries, 10/2020) with cysteine carbamidomethylation as a fixed modification and N-terminal acetylation and methionine oxidations as variable modifications. Enzyme specificity was set as C-terminal to arginine and lysine as expected using trypsin and lysC as proteases and a maximum of two missed cleavages.

Bioinformatics analysis of the MS data was performed using the Perseus software suite (version 1.6.7.0).⁵⁰ After filtering to remove potential contaminants, reverse hits, and proteins only identified by modification sites, the remaining summed intensities were log₂-transformed. Quantified proteins were filtered for at least 2 valid values in one fraction across three biological replicates. Missing values were imputed by sampling from a normal distribution (width 0.3, downshift 1.8) and significantly up- or downregulated proteins were determined by two-sided Student’s t-test (FDR < 0.05, S0 R 1.5). To determine the systematic enrichment or de-enrichment of a select list of GOCC annotated organelles in each fraction a Fisher’s exact test was performed on the significantly differentially regulated proteins between the two fractions.

Transient Transfection of HEK-293T cells

HEK-293T cells were transiently transfected with 400 ng plasmid DNA in 50 ml OptiMEM with 1 ml GeneJuice by incubating GeneJuice with OptiMEM for 5 minutes followed by 15 minutes of incubation of the DNA-GeneJuice mix in OptiMEM. DNA concentrations were kept constant across all conditions using pBluescript as stuffer DNA.

Plasmid DNA purification

Plasmid DNA was purified from E.Coli DH5a using a Thermo HiPure Maxiprep Kit according to manufacturer’s instructions.

Preparation of lentiviral particles

Lentiviral particles were prepared according to.⁵² Briefly, HEK-293T cells were transfected with 20 mg transfer plasmid, 15 mg pCMVD8.91 packaging plasmid and 6 mg pMD2.G VSV-G pseudotyping plasmid dish by diluting the plasmids in 1 ml 1× HBS, adding 50 ml 2.5 M Calcium chloride and gently pipetting the mix onto a 10-cm dish with approximately 6 Mio. HEK-293T cells in fresh medium. Alternatively, pMDLg/pRRE and pRSV-REV were used as packaging plasmids.⁵³ After 8 hours the medium was exchanged. Supernatants were harvested 48 hours later, spun down and filtered before being used to transduce target cells. Successfully transduced cells were selected with 2.5 - 5 mg/ml puromycin or 10 mg/ml blasticidin S for 48 hours, or FACSorted for fluorescence markers.

Genome editing and overexpression

sgRNA oligos were designed using CHOPCHOP⁴⁸ and cloned into expression plasmids as described previously.^44,46 BLaER1 cells were electroporated in OptiMEM with 5 mg of plasmids driving expression of Cas9 and an sgRNA on a BioRad GenePulser XCell as described previously.⁴⁵ THP-1 cells and murine macrophages were transduced with lentiviral particles driving expression of Cas9 (Lenti-Cas9-Blast⁴⁴) or an sgRNA (LentiGuide-Puro⁴⁴). HEK-293T cells were transiently transfected with plasmids driving expression of Cas9 or an sgRNA.

hiPS cells conditioned to grow as single clones were electroporated with Cas9-crRNA-trRNA complexes (RNPs) targeting NEK7 on a 4D-Nucleofector (Lonza Bioscience). Grown single clones were duplicated, lysed and out-of-frame editing in NEK7 was analyzed via deep sequencing as described previously.⁴⁶ Several NEK7^-/- and NEK7^+/+ clones were expanded and used for experiments. To generate IKBKB^-/- hiPSCs, XCL1 hiPS cells were electroporated with 0.5 mg of plasmids driving expression of Cas9 and an sgRNA targeting IKBKB with a 4D-Nucleofector (Lonza Bioscience). Grown single clones were picked and the sequence of the targeted IKBKB region was confirmed by Sanger sequencing.

Plasmids

Cloning of genes of interest into pLIX, pRP and pFUGW backbones was performed by conventional restriction enzyme cloning. pMDLg/pRRE was a gift from Didier Trono (Addgene plasmid #12251; http://n2t.net/addgene:12251; RRID:Addgene_12251), pRSV-Rev was a gift from Didier Trono (Addgene plasmid #12253; http://n2t.net/addgene:12253; RRID:Addgene_12253), pLIX_403 (herein referred to as pLIX) was a gift from David Root (Addgene plasmid #41395; http://n2t.net/addgene:41395; RRID:Addgene_41395). LentiGuide-Puro (Addgene plasmid # 52963; http://n2t.net/addgene:52963; RRID:Addgene_52963) and lentiCas9-Blast (Addgene plasmid #52962; http://n2t.net/addgene:52962; RRID:Addgene_52962) were a gift from Feng Zhang. pTY-zeo-NLRP3(127-128-129-130 4KA)-GFP) and pTY-zeo-Flag-NLRP3(DKKKK OSBPPH)) were a gift from Zhijian J. Chen.¹⁰

Quantification And Statistical Analysis

Numbers of independent replicates (n) are reported in the respective figure legends. p-values were calculated based on two-way ANOVAs followed by Šidák’s multiple comparisons test for groups containing two elements, or Tukey’s test for larger groups. Dunnett’s test was used wherever comparing all experimental conditions to one control instead of all other conditions was appropriate as indicated in the respective figure legends. All statistical analyses were done using GraphPad Prism 9. *p < 0.05, **p < 0.01, ***p < 0.001, ns p ≥ 0.05.

Data and code availability

Mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository (PRIDE: PXD035302) and are publicly available as of the date of publication.

Immunoblot source data and raw numerical data used to plot the figures were deposited on Mendeley Data: https://doi.org/10.17632/h7vc8hnb7j.1

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. This paper does not report original code.