Stoking inflammasome fires in the COVID-19 neighborhood

Abstract

The cellular and molecular sources of elevated IL-1β and IL-6 in COVID-19 remain unclear. In this issue of Cell Host and Microbe, Barnett et al. determine how immune cells sense SARS-CoV-2 infection in neighboring epithelial cells to trigger inflammasome signaling and IL-1β release, which in turn promotes IL-6 release.

The cellular and molecular sources of elevated IL-1β and IL-6 in COVID-19 remain unclear. In this issue of Cell Host and Microbe, Barnett et al. determine how immune cells sense SARS-CoV-2 infection in neighboring epithelial cells to trigger inflammasome signaling and IL-1β release, which in turn promotes IL-6 release.

Main text

Elevated pro-inflammatory cytokines are a clinical hallmark of severe COVID-19. Agents that block signaling by either interleukin (IL)-1β or IL-6 show clinical benefit in patients hospitalized with severe COVID-19, implicating these cytokines in disease pathogenesis.¹ However, the molecular pathways by which SARS-CoV-2 infection triggers excessive IL-1β and IL-6 release are still poorly understood. In this issue of Cell Host and Microbe, Barnett et al. elucidate how immune cells sense viral infection in neighboring epithelial cells to prompt inflammasome and IL-6 signaling in COVID-19.²

Epithelial and myeloid cells have pattern recognition receptors (PRRs) at strategic subcellular locations to sense active viral entry and replication within the cell. PRRs can also detect neighboring cell infection, damage, or death. PRR activation culminates in cellular pro-inflammatory and anti-viral host defense programs. Some cytosolic PRRs respond to cell-intrinsic and -extrinsic microbial threats by assembling inflammasomes. Inflammasomes are multimeric signaling platforms that cleave the potent pro-inflammatory cytokines IL-1β and IL-18 into their mature, active forms and induce pyroptotic cell death. Canonical inflammasomes generally comprise an inflammasome-nucleating PRR (e.g., NLRP3: NOD-like receptor [NLR] family, pyrin domain-containing protein 3; NLRP1), the adaptor protein ASC (apoptosis-associated speck-like protein containing a CARD), and the cysteine protease caspase-1. The PRR-ASC complex forms an activating scaffold that elicits caspase-1 protease activity.

Canonical inflammasome signaling is a two-step process. The first signal is often provided by a microbial stimulus that activates a cell surface PRR (e.g., Toll-like receptor 2; TLR2). This signal primes inflammasome activation by upregulating the expression of inflammasome components and substrates (e.g., NLRP3, pro-IL-1β) and boosting NLRP3 and adaptor function via post-translational modifications. The second signal triggers inflammasome assembly and caspase-1 activity. Cell-intrinsic, virus-induced cellular perturbations such as disrupted ion flux and lysosomal damage activate NLRP3, whereas extracellular ATP is a common cell-extrinsic NLRP3-activating signal. There are at least 7 distinct canonical inflammasomes in humans, and these are collectively expressed in cells of the innate and adaptive immune response as well as non-immune cells (e.g., epithelia). The specific inflammasome pathway(s) activated by SARS-CoV-2 and the participating cells remain important outstanding questions for understanding the molecular basis of pathological inflammation during COVID-19.

SARS-CoV-2 can infect ACE2 (angiotensin I converting enzyme 2)-expressing airway epithelial cells (AEC),³ monocytes⁴ and macrophages.⁵ Alternatively, antibodies promote SARS-CoV-2 uptake into monocytes independently of ACE2.⁶ Cytosolic SARS-CoV-2 entry and active viral replication then activates inflammasome signaling in AECs and monocytes.^{3 , 4 , 6} Barnett et al. first confirmed that airway epithelial cells (AECs), the primary cellular target of SARS-CoV-2, can signal via inflammasomes. While AECs can activate functional NLRP1 inflammasomes, these cells appear unable to elicit NLRP3 signaling, and NLRC4 and AIM2 signaling is minimally activated. The authors established that SARS-CoV-2 readily replicates in AECs and triggers lytic cell death without concomitant IL-1β release; this is consistent with a recent report that the SARS-CoV-2 protease activates the NLRP1 inflammasome, triggering AEC pyroptosis and selective IL-18 release without accompanying IL-1β.³

Myeloid inflammasome activation usually culminates in IL-1β release and cell death. Myeloid cells are likely to be a primary source of elevated IL-1β in severe COVID-19, yet they are rarely directly infected by SARS-Cov-2 due to limited ACE2 expression.⁵ Barnett et al. report no SARS-CoV-2 replication in peripheral blood mononuclear cells (PBMCs; containing lymphocytes [B, T, and NK cells; 70–90%], monocytes [10–20%], and dendritic cells [1–2%]), or mature IL-1β release from PBMCs, indicating that SARS-CoV-2 infection and cell-intrinsic inflammasome activation do not occur in monocytes or other leukocytes. This contrasts with reports of monocyte activation by Rodrigues et al.⁴ but is consistent with findings from Junqueira and co-workers that monocyte inflammasome activation cannot occur without virus-specific antibodies.⁶ Instead, the authors find that SARS-CoV-2 only triggers mature IL-1β release from co-cultures of epithelial cells (Veros or human AECs) and PBMCs. Thus, PBMCs only activate inflammasome signaling in the presence of infected epithelial cells. Barnett et al. confirmed this discovery in COVID-19 lung autopsy samples by staining for the inflammasome adaptor ASC. A hallmark of canonical inflammasome activation is that ASC coalesces into a single large “speck” in the cell upon activation. By co-staining for SARS-CoV-2 antigen and ASC, the authors discovered that primarily bystander, non-infected myeloid cells contain ASC specks, indicating that in vivo, cell-extrinsic stimuli provoke inflammasome activation in myeloid cells, generating a hitherto underappreciated source of elevated IL-1β in severe COVID-19.

Barnett et al. next examined how neighboring cell infection triggers myeloid inflammasome activation. Consistent with findings by Zheng et al.,⁷ the SARS-CoV-2 envelope I protein provides signal 1 in PBMCs, increasing pro-IL-1β, NLRP3, and AIM2 mRNA expression. In contrast, neither infectious SARS-CoV-2 virions nor SARS-CoV-2 E protein are sufficient to act as signal 2 to activate inflammasomes in pre-primed PBMCs. Infected, dying AECs will conceivably release newly synthesized virions, free viral E protein, and other host molecules, including host DNA or ATP, into the extracellular space. Indeed, Barnett et al. observed that infected epithelial cells released genomic DNA (gDNA) and mtDNA. Further, the authors showed that degrading extracellular DNA ablates IL-1β release from infected epithelial-PBMC co-cultures. Thus, gDNA or mtDNA is likely to be the epithelial molecule sensed by the PBMCs in co-culture settings. Future experiments are required to determine whether other molecules, such as ATP, are also involved.

How extracellular DNA triggers IL-1β release from PBMC is less clear. AIM2 is the primary dsDNA-sensing inflammasome in mice, whereas in human myeloid cells, dsDNA triggers the cytosolic cyclic GMP-AMP synthase (cGAS) stimulator of interferon genes (STING) pathway, which subsequently activates NLRP3.⁸ Consistent with this latter model for dsDNA-induced STING and inflammasome signaling, Barnett et al. show that STING inhibition ablates IL-1β release in the epithelial cell-PBMC co-culture; however, NLRP3 inhibition did not block IL-1β release within infected co-cultures, indicating the participation of a non-NLRP3 inflammasome. Further experiments with genetic knockouts are required to resolve the involvement of cGAS/STING with inflammasome pathway(s) in extracellular DNA sensing.

Interferons and inflammasomes often counter-regulate each other. Di Domizio et al.⁹ recently showed that mtDNA from infected epithelial cells activates cGAS/STING signaling in macrophages to drive aberrant Type I interferon expression. It is, however, unclear whether an individual myeloid cell exposed to extracellular DNA can trigger both inflammasome and interferon signaling. It is also unclear how extracellular gDNA or mtDNA may enter cells to access myeloid cytosolic sensors (e.g., cGAS, NLRP3, or AIM2). Answering these outstanding questions could identify new therapeutic strategies for selectively inhibiting pathological inflammasome signaling without affecting anti-viral defense.

IL-6 is another potent inflammatory cytokine that drives COVID-19 pathogenesis. Like IL-1β, IL-6 was released from SARS-CoV-2-infected epithelial-PBMC co-cultures but not monocultures. Given that IL-1 signaling can trigger IL-6 release, Barnett et al. blocked IL-1 signaling with the IL-1 receptor antagonist anakinra in co-culture experiments. Indeed, anakinra ablated IL-6 release from infected epithelial-PBMC co-cultures, indicating that IL-6 is elicited by IL-1 in this system. Further, in vivo autopsy staining showed an IL-6-positive epithelial cell adjacent to an IL-1β-positive macrophage in the COVID-19 lung. A model for the epithelial-immune cell response to SARS-CoV-2 is shown in Figure 1 . Further experiments should explore whether IL-6 release from infected epithelial-PBMC co-cultures can be blocked by STING and inflammasome inhibitors or by digestion of extracellular gDNA and mtDNA. Such experiments would further establish the signaling hierarchy in this IL-1β–IL-6 feedback cycle and could provide the basis for specific therapeutic targeting to interrupt this pathological inflammatory amplification loop.

Figure 1.

Bystander myeloid cells activate the inflammasome in response to neighboring SARS-CoV-2 infected epithelial cells

SARS-CoV-2 infects airway epithelial cells, triggering pyroptotic cell death, likely via NLRP1.³ Newly released virions or free viral envelope proteins prime inflammasome signaling via TLR2. Epithelial gDNA and mtDNA are released into the extracellular space. Neighboring, uninfected myeloid cells sense the extracellular epithelial DNA and trigger STING-dependent inflammasome activation and resultant IL-1β release. IL-1β signaling promotes subsequent IL-6 release from epithelial and likely other cells.

Inflammasome signaling in other viral infections, such as influenza, can be both protective and pathogenic.¹⁰ In influenza, this may depend on the timing of inflammasome activation, whereby inflammasome signaling early during infection is critical for effective immune responses and prolonged signaling drives pathology.¹⁰ Alternatively, the magnitude of inflammasome signaling may determine whether inflammasome activation is protective or pathogenic. Barnett et al. compared BALF from mild and severe COVID-19 patients with healthy controls using scRNA sequencing to measure inflammasome priming status. They determined that patients with severe COVID-19 had higher expression of inflammasome pathway proteins than mild or healthy controls, indicating that their inflammasomes were “primed” and that the threshold for inflammasome activation is lower in these patients. This suggests that, like timing, the magnitude of inflammasome activation in COVID-19 may determine disease outcome. Future analysis should determine whether increased inflammasome priming observed in severe COVID-19 patients is a consequence of severe infection or pre-priming of inflammasomes before infection (e.g., due to inflammation associated with risk factors such as age or metabolic syndrome). Understanding the answers to these questions may help identify who is at risk of severe COVID-19 and how best to stratify patients for prevention or treatment.

This study identifies an essential role for bystander, non-infected myeloid cells in amplifying the inflammatory response to SARS-CoV-2. By studying epithelial and immune cells in concert, Barnett et al. reveal a critical feedback loop whereby epithelial cells promote immune cell IL-1β release, which in turn promotes IL-6 release. Such cell-extrinsic, myeloid-sensing pathways may be targeted to selectively inhibit pathogenic inflammatory signaling during viral infection without compromising anti-viral defense.