Research ArticleINNATE IMMUNITY

Targeting Piezo1 unleashes innate immunity against cancer and infectious disease

See allHide authors and affiliations

Science Immunology  21 Aug 2020:
Vol. 5, Issue 50, eabb5168
DOI: 10.1126/sciimmunol.abb5168

Myeloid cells feel the squeeze

The Piezo1 and Piezo2 proteins are excitatory ion channels used by mammalian cells to sense and respond to mechanical stimuli, but little is known about whether mechanosensors participate in immune responses. Aykut et al. investigated how modulation of Piezo1’s activity affects inflammation and immunity using agonist or antagonist compounds targeting Piezo1 along with mice lacking Piezo1 in myeloid cells. In isolated myeloid cells, Piezo1 served as the primary mechanosensory receptor. Loss of Piezo1 in myeloid cells or its global inhibition with a tarantula venom–derived peptide inhibited the differentiation of suppressive myeloid cells in mouse models of pancreatic adenocarcinoma and polymicrobial sepsis. These studies spotlight Piezo1-dependent mechanosensation as a checkpoint regulating differentiation of suppressive myeloid cells that stifle immune control of cancer and infection.

Abstract

Piezo1 is a mechanosensitive ion channel that has gained recognition for its role in regulating diverse physiological processes. However, the influence of Piezo1 in inflammatory disease, including infection and tumor immunity, is not well studied. We postulated that Piezo1 links physical forces to immune regulation in myeloid cells. We found signal transduction via Piezo1 in myeloid cells and established this channel as the primary sensor of mechanical stress in these cells. Global inhibition of Piezo1 with a peptide inhibitor was protective against both cancer and septic shock and resulted in a diminution in suppressive myeloid cells. Moreover, deletion of Piezo1 in myeloid cells protected against cancer and increased survival in polymicrobial sepsis. Mechanistically, we show that mechanical stimulation promotes Piezo1-dependent myeloid cell expansion by suppressing the retinoblastoma gene Rb1. We further show that Piezo1-mediated silencing of Rb1 is regulated via up-regulation of histone deacetylase 2. Collectively, our work uncovers Piezo1 as a targetable immune checkpoint that drives immunosuppressive myelopoiesis in cancer and infectious disease.

INTRODUCTION

Piezo1 belongs to a family of evolutionarily conserved nonselective cation channels that serve in mechanosensory transduction and have recently emerged as central in diverse biological processes (1, 2). Cells in many animals, plants, and other eukaryotic species contain a single Piezo gene; however, vertebrates have two members, Piezo1 and Piezo2 (3). Whereas Piezo2 is primarily expressed in dorsal root ganglia neurons and Merkel cells regulating mechanical nociception, Piezo1 is also expressed in non-neuronal tissues, implying a broader physiological role (47). Piezo1 forms homotrimeric assemblies in a three-blade propeller conformation with three peripheral blades, three intracellular beams, and a central cap, which allows for a lever-like gating mechanism (8, 9). Mechanical stimulation of Piezo1 at the cell surface evokes characteristic ionic currents (3). Piezo1 has been shown to be a regulator of key biological processes, including cell division, migration, and differentiation (6, 10, 11). Global genetic ablation of Piezo1 in mice is embryonically lethal, inducing defects in vascular remodeling that appear at midgestation (5). Human genetic studies have demonstrated that mutations in Piezo1 are associated with hereditary xerocytosis and congenital lymphatic dysplasia (12, 13). More recently, a gain-of-function Piezo1 allele in select African populations was associated with red blood cell dehydration and malaria resistance (14). However, the broader functions of Piezo1 in the hematopoietic and immune systems are not well understood.

Multiple lines of investigation highlight the emerging role of ion channels in leukocyte development and function. Mutations in Ca2+, Mg2+, and Cl channels have been linked to diverse immunodeficiencies (1517). We were therefore spurred to investigate the role of Piezo1 in basic leukocyte biology and inflammatory disease. To study the effects of Piezo1 in cancer immunology and sepsis, we generated mice that lacked Piezo1 in myeloid cell lineages. Our work suggests that deletion of Piezo1 augments immunity against cancer and polymicrobial sepsis.

RESULTS

Piezo1 signaling in inflammatory cells promotes cancer progression and myeloid-derived suppressor cell expansion

To assess the influence of global Piezo1 signaling on the progression of pancreatic ductal adenocarcinoma (PDA), we used GsMTx4 and Yoda1, respectively, to inhibit or activate Piezo1 (18, 19). Inhibition of Piezo1 conferred tumor protection in an orthotopic PDA model using tumor cells derived from Pdx1Cre;KrasG12D;Tp53R172H (KPC) mice (Fig. 1A). By contrast, activation of Piezo1 accelerated tumor growth (Fig. 1B). Similarly, in human PDA, low PIEZO1 expression was associated with extended survival (Fig. 1C). Whereas ~20% of patients that expressed low PIEZO1 were 5-year survivors, there were no long-term survivors in the high PIEZO1 group. PDA tumor cells expressed Piezo1 (fig. S1A); however, neither activation nor inhibition of Piezo1 affected tumor cell growth in vitro (fig. S1B), and knockdown of Piezo1 did not affect PDA growth in vivo (Fig. 1D and fig. S1C). We therefore postulated that inhibiting Piezo1 mitigates oncogenic progression by reprograming the inflammatory tumor microenvironment (TME). Using Piezo1P1-tdT mice that express a fluorescent Piezo1-tdTomato fusion protein, we confirmed that leukocytes robustly express Piezo1 (Fig. 1E). To directly test our hypothesis, we generated VavCre;Piezo1fl/fl mice, in which Piezo1 is deleted from all hematopoietic cells. VavCre;Piezo1fl/fl mice demonstrated attenuated tumor growth compared with littermate controls (Fig. 1F).

Fig. 1 Piezo1 signaling enables pancreatic cancer progression.

(A) WT mice bearing orthotopic KPC tumor were treated with GsMTx4 or vehicle (Veh.) and euthanized at 3 weeks. Representative images and quantitative analysis of tumor weights are shown. Scale bar, 1 cm. This experiment was repeated three times with similar results. (B) WT mice bearing orthotopic KPC tumor were treated with Yoda1 or vehicle and euthanized at 3 weeks. Representative images and quantitative analysis of tumor weights are shown. Scale bar, 1 cm. This experiment was repeated three times with similar results. (C) Five-year Kaplan-Meier survival curve of human PDA patients stratified by high (n = 42) versus low (n = 135) PIEZO1 expression based on the TCGA data. (D) WT mice were administered orthotopic KPC tumors cells treated with either short hairpin RNA (shRNA) directed against Piezo1 or control scrambled shRNA. Tumors were harvested on day 21 and weighed. This experiment was repeated twice with similar results. (E) Frozen sections of spleens from Piezo1P1-tdT reporter mice were costained with DAPI and αCD45. Representative immunofluorescent images are shown. (F) VavCre;Piezo1fl/fl mice and littermate controls bearing orthotopic KPC tumors were euthanized at 3 weeks. Representative images of tumors and quantitative analysis of tumor weights are shown. Scale bar, 1 cm. This experiment was repeated three times with similar results. (G) WT mice bearing orthotopic KPC tumor were treated with GsMTx4 or vehicle and euthanized at 3 weeks. The fraction of intratumoral Gr-1+CD11b+ MDSCs was determined by flow cytometry. This experiment was repeated three times with similar results. (H) VavCre;Piezo1fl/fl mice and littermate controls bearing orthotopic KPC tumor were euthanized at 3 weeks, and the fraction of intratumoral Gr-1+CD11b+ MDSCs was determined by flow cytometry. This experiment was repeated three times with similar results. (I) WT mice bearing orthotopic KPC tumor were treated with Yoda1 or vehicle and euthanized at 3 weeks. The fraction of intratumoral Gr-1+CD11b+ MDSCs was determined by flow cytometry. This experiment was repeated three times with similar results. (J and K) PDOTS from freshly resected human tumors (n = 5) were treated with GsMTx4 or vehicle for 3 days. (J) Spheroid sizes were measured. (K) Gr-1+CD11b+ MDSC cells were quantified. Data are presented as fold change compared with vehicle. (L and M) PDOTS from freshly resected human tumors (n = 5) were treated with Yoda1 or vehicle for 3 days. (L) Spheroid sizes were measured. (M) Gr-1+CD11b+ MDSC cells were quantified. Data are presented as fold change compared with vehicle. *P < 0.05, **P < 0.01, and ***P < 0.001. ns, not significant.

We analyzed the intratumoral immune phenotype in PDA-bearing mice treated with GsMTx4. We found that the Gr-1+CD11b+ myeloid-derived suppressor cell (MDSC) population was sharply diminished upon Piezo1 inhibition (Fig. 1G). Similarly, PDA tumors in VavCre;Piezo1fl/fl mice exhibited lower MDSC numbers (Fig. 1H). By contrast, activation of Piezo1 in vivo with Yoda1 increased the number of MDSCs in PDA (Fig. 1I). To evaluate the therapeutic efficacy of Piezo1 inhibition in human PDA, we treated patient-derived organotypic tumor spheroids (PDOTS) from freshly resected human tumors with GsMTx4 or vehicle using a three-dimensional (3D) microfluidic system that we recently validated as a platform for testing immune-based therapeutics (20, 21). Piezo1 inhibition attenuated spheroid growth in PDOTS and resulted in MDSC contraction (Fig. 1, J and K), whereas Piezo1 activation accelerated spheroid growth in PDOTS and increased MDSCs (Fig. 1, L and M).

Myeloid cells sense mechanical forces via Piezo1

Immune homeostasis depends on the input of various external stimuli recognized by inflammatory cells. Changes in hydrostatic pressure were recently shown to induce a proinflammatory response in lung monocytes in a Piezo1-dependent manner (22). Therefore, we sought to investigate the broader immunoregulatory role for Piezo1 signaling in myeloid cells. Human CD14+ myeloid cells expressed ~10-fold higher levels of PIEZO1 than the median expression in other tissues based on the BioGPS gene portal data (Fig. 2A). Similarly, we confirmed that mouse myeloid cells robustly express Piezo1 (Fig. 2B) (23). To directly test whether Piezo1 signals in myeloid cells, we used Piezo1−/− CD11b+ myeloid cells harvested from VavCre;Piezo1fl/fl mice in patch-clamp experiments. We found that mechanical stimulation of control Piezo1fl/fl myeloid cells in whole-cell patch-clamp configuration evoked currents characteristic of Piezo1 activation (Fig. 2, C and D) (24). By contrast, Piezo1−/− myeloid cells failed to respond to mechanical stimulation (Fig. 2, E and F). Collectively, these data indicate that myeloid cells express Piezo1 and sense mechanical forces via Piezo1, whereas in the context of Piezo1 deletion, myeloid cells fail to transduce pressure-mediated currents.

Fig. 2 Piezo1 senses pressure in myeloid cells.

(A) BioGPS gene portal data comparing PIEZO1 expression in human CD14+ monocytes compared with median expression in other tissues. (B) Frozen sections of spleens from Piezo1P1-tdT reporter mice were costained with DAPI and anti-CD11b. Representative immunofluorescent images are shown. (C to F) Mechanically activated currents were elicited from activated (C and E) Piezo1fl/fl (n = 3) and (D and F) Piezo1−/− (n = 5) myeloid cells in whole-cell patch-clamp configuration with pressure applied to the patch pipette (0 to −140 mmHg, Δ20 mmHg) at a holding potential of +70 mV.

Piezo1 signaling in myeloid cells promotes cancer progression and increases MDSC content of tumors

To specifically investigate the impact of Piezo1 signaling in myeloid cells in cancer, we generated Lyz2Cre;Piezo1fl/fl mice and challenged these mice and Piezo1fl/fl littermate controls with orthotopic KPC tumors. Lyz2Cre;Piezo1fl/fl mice were protected against PDA (Fig. 3A). Tumors in Lyz2Cre;Piezo1fl/fl mice exhibited a reduced MDSC infiltrate akin to pan-inhibition or deletion of Piezo1 (Fig. 3B). Moreover, consistent with our previous report that MDSCs suppress adaptive immunity in PDA (25), we found that targeting Piezo1 in myeloid cells resulted in enhanced intratumoral CD4+ and CD8+ T cell activation (Fig. 3, C and D). Control Piezo1fl/fl and Lyz2Cre;Piezo1fl/fl mice had similar frequencies of progenitor populations in the bone marrow and macrophage and dendritic cell (DC) populations in the periphery (Fig. 3, E and F). Further, Piezo1 deletion did not alter myeloid cell size (fig. S2, A and B).

Fig. 3 Piezo1 deletion on myeloid cells confers antitumor immunity.

(A to D) Cohorts of Lyz2Cre;Piezo1fl/fl mice and littermate controls were administered orthotopic KPC-derived tumor cells. Scale bar, 1 cm. (A) Mice were euthanized at 3 weeks. Representative pictures of tumors and quantitative analysis of tumor weights are shown. (B) The prevalence of intratumoral Gr-1+CD11b+ MDSCs was determined. (C) Intratumoral CD4+ T cells were assessed for expression of CD44, CD69, and T-bet. (D) Intratumoral CD8+ T cells were assessed for expression of CD69, LFA-1, LAG3, and granzyme B. Tumor experiments in Lyz2Cre;Piezo1fl/fl mice were performed four times. (E) Bone marrow (BM) cells from Lyz2Cre;Piezo1fl/fl mice and control Piezo1fl/fl mice were analyzed for the frequency of CD117+CD115+CD135+Ly6CCD11b myeloid progenitor cells (MPCs) and CD117+CD115+CD135Ly6C+CD11b common monocyte progenitors (cMoP) in Lin (CD3CD19NK1.1Ly6G) cells. Data are representative of experiments performed three times. (F) Splenocytes from Lyz2Cre;Piezo1fl/fl mice and control Piezo1fl/fl mice were analyzed for the frequency of splenic CD11b+CD68+F4/80+ macrophages and CD11b+CD11c+MHCII+ DCs. Data are representative of experiments performed three times. *P < 0.05 and ***P < 0.001.

Deletion of Piezo1 in myeloid cells protects against polymicrobial sepsis

We postulated that Piezo1 may also influence outcome in bacterial sepsis, which is regulated by myeloid cell expansion and characterized by increased interstitial pressure (26, 27). Lyz2Cre;Piezo1fl/fl mice and Piezo1fl/fl littermate controls were subjected to polymicrobial sepsis via cecal ligation and puncture (CLP). Lyz2Cre;Piezo1fl/fl animals were protected from sepsis, exhibiting minimal core temperature loss, reduced clinical sepsis scores, decreased interstitial edema, and marked protection from sepsis-induced death (Fig. 4, A to D). Serum levels of proinflammatory cytokines were also substantially lower in CLP-treated Lyz2Cre;Piezo1fl/fl mice (Fig. 4E). Further, consistent with our data in tumor-bearing hosts, Lyz2Cre;Piezo1fl/fl mice exhibited a decreased frequency of MDSCs in peripheral blood mononuclear cells (PBMCs) and reduced MDSC abundance in the peritoneum (Fig. 4, F and G). Accordingly, bacteremia was markedly reduced in Lyz2Cre;Piezo1fl/fl mice (Fig. 4H). Similarly, culture of peritoneal contents yielded ~104-fold lower bacterial colonies in Lyz2Cre;Piezo1fl/fl mice compared with controls (Fig. 4I). Consistent with these data, GsMTx4 treatment of wild-type (WT) mice also protected against CLP, resulting in preservation of core body temperature, lower clinical sepsis scores, and reduced serum levels of tumor necrosis factor–α (TNF-α) (fig. S3).

Fig. 4 Piezo1 deletion on myeloid cells confers protection against polymicrobial sepsis.

(A to I) Cohorts of Lyz2Cre;Piezo1fl/fl mice and littermate controls were subjected to CLP. (A) Rectal temperature was serially measured, and (B) mice were assigned a clinical sepsis score over the first 24 hours (n = 5 mice per group). (C) Pulmonary edema was measured at 24 hours. (D) Additional cohorts of mice were subjected to survival analysis using the Kaplan-Meier estimator (n = 10 mice per group). (E) Serum levels of select inflammatory mediators were measured at 24 hours. MCP-1, monocyte chemoattractant protein-1. (F) The prevalence of Gr-1+CD11b+ MDSCs in PBMCs and (G) the number of peritoneal MDSCs were determined at 24 hours. (H) Bacterial titers were measured in the blood and (I) peritoneal cavity. CLP experiments in Lyz2Cre;Piezo1fl/fl mice were performed four times with similar results. *P < 0.05, **P < 0.01, and ***P < 0.001.

Piezo1 governs myeloid cell expansion via regulation of Rb1

Because Rb1 can suppress MDSC expansion (28), we postulated that Piezo1 controls MDSC levels in disease by regulation of Rb1. We observed an increase in peritoneal MDSCs in Lyz2Cre;Rb1fl/fl mice compared with littermate controls after CLP (Fig. 5A). Accordingly, Lyz2Cre;Rb1fl/fl mice exhibited exacerbation in sepsis-related inflammation, morbidity, and death (Fig. 5,B to D). Consistent with our hypothesis, we further found that increasing mechanical stress in myeloid cells lowered expression of Rb1 (Fig. 5E). Similarly, activating Piezo1 using Yoda1 reduced Rb1 expression (Fig. 5F). By contrast, GsMTx4-mediated inhibition of Piezo1 (Fig. 5G) or deletion of Piezo1 in myeloid cells (Fig. 5H) up-regulated Rb1. Piezo1−/− and Piezo1fl/fl myeloid cells expressed similar levels of Rb1 by quantitative polymerase chain reaction (qPCR) after an overnight culture on plates of low stiffness (fig. S4A). Collectively, these data indicate that whereas Piezo1 signaling or mechanical stimulation suppresses Rb1, ablation of Piezo1 up-regulates Rb1. Moreover, concomitant Piezo1 and Rb1 deletion in myeloid cells using Lyz2Cre;Piezo1fl/fl;Rb1fl/fl mice abrogated the tumor protection observed with Piezo1 deletion alone (Fig. 5, I and J).

Fig. 5 Piezo1 governs myeloid cell expansion via epigenetic regulation of Rb1.

(A to D) Cohorts of Lyz2Cre;Rb1fl/fl mice and littermate controls were subjected to CLP. (A) The prevalence of MDSCs in the peritoneal cavity, (B) serum levels of inflammatory mediators, and (C) rectal temperatures were measured at 24 hours. (D) Additional cohorts of mice were subjected to survival analysis (n = 11). (E) Myeloid cells harvested from WT bone marrow were cultured on 0.2- and 50-kPa Matrigen-Softwell plates overnight and tested for expression of Rb1 by qPCR. (F) Myeloid cells harvested from WT bone marrow were treated overnight with Yoda1 or vehicle and tested for expression of Rb1 by qPCR. (G) Myeloid cells harvested from WT bone marrow were treated overnight with GsMTx4 or vehicle and assayed for expression of Rb1 by qPCR. (H) Myeloid cells harvested from bone marrow of Lyz2Cre;Piezo1fl/fl mice and littermate Piezo1fl/fl controls were assayed for expression of Rb1 by qPCR. All qPCR experiments were performed in biologic replicates of three. (I and J) Lyz2Cre;Piezo1fl/fl;Rb1fl/fl mice and littermate controls were challenged with orthotopic PDA and euthanized at 3 weeks. (I) Tumor weight and (J) the frequency of tumor-infiltrating MDSCs were recorded. (K) Myeloid cells harvested from bone marrow of Lyz2Cre;Piezo1fl/fl mice and controls were tested for expression of HDAC2 by Western blotting. This experiment was repeated twice with similar results. (L to O) Myeloid cells harvested from the bone marrow of Lyz2Cre;Piezo1fl/fl and littermate control mice were analyzed by scRNA-seq. (L) 3D t-SNE plot of myeloid gene clusters and expression analysis based on genes that showed a ≥2-fold difference are shown. (M) Expression of Rb1 and Hdac2 in GMPs and MDPs from Lyz2Cre;Piezo1fl/fl mice normalized to control GMPs and MDPs. (N) Ingenuity Pathway Analysis indicating top 5 canonical pathway alterations based on the z score in GMPs and (O) MDPs from Lyz2Cre;Piezo1fl/fl mice versus control. NER, nucleotide excision repair; mTOR, mammalian target of rapamycin. (P) ChIP-seq analysis in Piezo1+/+ and Piezo1−/− bone marrow cells demonstrating binding of acetyl-histone H4 to the Rb1 promoter. IP, immunoprecipitation. (Q to S) Lyz2Cre;Piezo1fl/fl mice and littermate Piezo1fl/fl controls were treated with LBH589 and subjected to CLP. (Q) Rectal temperature was serially measured, (R) mice were assigned a clinical sepsis score over the first 24 hours, and (S) the number of peritoneal MDSCs was measured at 24 hours (n = 4 per group). *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Because Rb1 expression is inhibited by histone deacetylase 2 (HDAC2) (29), we hypothesized that Piezo1 signaling expands MDSCs by suppressing Rb1 via epigenetic silencing. Consistent with our hypothesis, we observed reduced HDAC2 expression in Lyz2Cre;Piezo1fl/fl myeloid cells (Fig. 5K). We performed single-cell RNA sequencing (scRNA-seq) of the bone marrow of Lyz2Cre;Piezo1fl/fl and littermate control mice to investigate whether Piezo1 promotes MDSC expansion via the HDAC2-Rb1 axis (Fig. 5L). We found diminished Hdac2 expression but increased Rb1 expression in Piezo1−/− granulocyte-monocyte progenitors (GMPs) and macrophage and DC progenitor (MDP), each of which gives rise to MDSCs (Fig. 5M) (30). Further, Ingenuity Pathway Analysis indicated that multiple pathways that drive MDSC expansion, including eukaryotic initiation factor 2 (eIF2) signaling and Sirtuin signaling, were among the most down-regulated pathways in Piezo1-deficient GMPs and MDPs (Fig. 5, N and O) (31, 32). Using chromatin immunoprecipitation sequencing (ChIP-seq), we confirmed enhanced histone H4 acetylation of the Rb1 promoter in Lyz2Cre;Piezo1fl/fl bone marrow cells (Fig. 5P and fig. S4, B and C). Similarly, HDAC2 inhibition eliminated the protection against polymicrobial sepsis and the associated reduced MDSC expansion in Lyz2Cre;Piezo1fl/fl hosts (Fig. 5, Q to S). In aggregate, these data indicate that deletion of Piezo1 in myeloid cells is protective against neoplastic and infectious diseases in an Rb1-dependent manner.

DISCUSSION

More than 600 different ion channels, transporters, and pumps have been reported to mediate the transfer of charged ions across the membrane, but only a minority have been described in immunocompetent cells (33, 34). Divergent leukocyte populations can express distinct repertoires of ion channels. Nevertheless, these ion channels share two common features: (i) the expression of monovalent cation channels that regulate membrane potentials and thereby indirectly control Ca2+ influx and (ii) the expression of divalent cation channels that enable intracellular signal transduction (17). Ca2+ influx governs fundamental physiological functions in inflammatory cells, including cell proliferation, differentiation, migration, and apoptosis. Ca2+ influx also regulates the expression and function of a diversity of enzymes and transcription factors, which regulate inflammatory cell physiology (16, 3537). However, the molecular mechanisms by which ion channels regulate immune function remain largely unknown.

Piezo1 is a nonselective cation channel that serves in mechanosensory transduction. Its physiologic role has been delineated in red blood cells, vascular endothelium, and neuronal function (5, 14, 38). However, Piezo1 has not been intensively studied in inflammatory disease or in cancer. Considering the emerging appreciation for the role of ion channels in immune cell function and inflammation, we hypothesized a critical function for Piezo1 in leukocyte biology. Our results provide direct demonstration that Piezo1 is a vital regulator of innate immune responses with implications for the balance of tumor immunity and clearance of infectious viral and bacterial diseases. Moreover, we demonstrate that myeloid cells sense mechanical forces via Piezo1 because Piezo1-deficient leukocytes were unable to transduce mechanosensory signals. Thus, Piezo1 links physical forces to immune regulation.

We show that mechanotransduction in myeloid cells requires Piezo1. Increased interstitial pressures have been reported in PDA and in inflamed tissues (39, 40). In PDA, extensive desmoplasia and accumulation of extracellular matrix proteins that maintain intratumoral tensile stress have been correlated with decreased survival (41). These observations suggest that targeting Piezo1 may be an attractive therapeutic approach to PDA treatment.

We demonstrate that targeting Piezo1 in myeloid cells is protective against cancer and polymicrobial sepsis and is associated with a reduced infiltrate of immature myeloid cells. Regulation of sepsis by ion channels has existing basis in clinical medicine because the use of calcium channel blockers was recently reported to attenuate pathogenic cytokine release and decrease mortality in patients with sepsis (42). In addition, leukocytes in patients with severe sepsis exhibit up to 2.5-fold greater membrane stiffness compared with controls (43). A persistence of cell stiffening is associated with dismal prognosis. Membrane mechanics have been linked to tuning the sensitivity of mechano-gated channels, including Piezo1 (44). Piezo1 signaling in monocytes has recently been linked to neutrophil recruitment and clearance of Pseudomonas aeruginosa infection via a hypoxia-inducible factor 1α–endothelin-1 axis (22). Piezo1 also drove lung pathology and pulmonary fibrosis, suggesting that Piezo1 is a critical regulator of homeostasis. Hence, these and our findings indicate that the effects of Piezo1 blockade are disease context dependent and may be beneficial in fibrosis and abdominal sepsis but detrimental in localized lung infection.

In cancer, we have previously shown that MDSCs are recruited to PDA by diverse biochemical signals including tumor cell secretion of granulocyte-macrophage colony-stimulating factor (GM-CSF), chemokine signals consequent of Toll-like receptor 9 activation in pancreatic stellate cells, and soluble inflammatory mediators generated by the distinctive tumor-associated microbiome (25, 4547). Collectively, the recruited MDSCs serve to dampen CD8+ T cell responses in the TME. However, mechanisms of cell intrinsic regulation of MDSC homeostasis in PDA or sepsis have not been described. Recent findings suggest that Piezo1 can activate HDACs (48). HDAC2, specifically, has recently emerged as a master regulator of MDSC function and differentiation (29, 49, 50). Furthermore, HDAC2 is highly expressed in PDA and mediates therapeutic resistance (51). These findings, together with our observation of reduced immature myeloid cell numbers in Lyz2Cre;Piezo1fl/fl mice and in animals treated with the pharmacological Piezo1 inhibitor GsMTx4, suggest a possible role of HDAC2 in Piezo1 signaling in myeloid cells. We demonstrate that HDAC2 is down-regulated in myeloid cells lacking Piezo1, and stimulation of Piezo1 is associated with increased MDSC numbers via epigenetic silencing of Rb1 expression. These results also parallel with previous studies showing that HDAC2 expression inversely correlates with Rb1 in myeloid cells (29). Hence, inhibition of HDAC2 abrogated protection against sepsis and led to MDSC contraction in mice with targeted Piezo1 deletion in myeloid cells, although it is possible that other signaling molecules downstream of Piezo1 contributed to the result observed in our study. In aggregate, we show that Piezo1 serves as an ion channel checkpoint that can markedly suppress immunological responses via enhancing myeloid cell tolerance. Moreover, targeting of Piezo1 in the immune compartment may have therapeutic implications for cancer immunotherapy and treatment of select bacterial infections.

MATERIALS AND METHODS

Study design

We performed phenotypic, transcriptomic, and functional analyses of myeloid cells to investigate the role of signaling via Piezo1. We used transgenic mice as well as chemical activators and inhibitors to assess the functional outcomes of mechanical stimulation of myeloid cells. Experiments were conducted in replicates as indicated in the figure legends. Experiments were not conducted in a blinded or randomized manner. No outliers were excluded.

Animals and in vivo models

C57BL/6, Lyz2Cre, VavCre, Piezo1fl/fl, Piezo1P1-tdT, and Rb1fl/fl mice were purchased from the Jackson laboratory (Bar Harbor, ME) and bred in-house. All genetically modified strains were maintained on a C57BL/6 background. Age-matched 8- to 10-week-old mice were used in the experiments. Both male and female mice were used, but animals were gender-matched within each experiment. For orthotopic pancreatic tumor challenge, mice were administered intrapancreatic injections of FC1242 tumor cells (1 × 105) derived from KPC mice, as previously described (46). Cells were suspended in phosphate-buffered saline (PBS) with 50% Matrigel (BD Biosciences, Franklin Lakes, NJ) before administration. In select experiments, sublethally irradiated (600 centigrays) mice were adoptively transferred intravenously with fluorescence-activated cell sorting (FACS)–purified CD3+ T cells (3 × 104) 3 days after orthotopic tumor injections. In other experiments, mice were serially treated with GsMTx4 (0.8 mg/kg; Alomone Labs, Jerusalem, Israel) or Yoda1 (2.6 mg/kg; Tocris, Bristol, UK). In vivo doses of GsMTx4 and Yoda1 were based on previous reports and pharmacokinetic studies (52, 53). In select experiments, mice were treated daily with the HDAC inhibitor LBH589 (20 mg/kg; Selleckchem, Houston, TX) as previously described (54). CLP experiments were performed as described previously (55). Briefly, mice underwent laparotomy, and the cecum was exteriorized and ligated at ~50% distal to the ileocecal valve using an absorbable 4-0 suture. The cecum was then perforated using a 23-gauge needle and returned to the peritoneal cavity, after which the peritoneum was closed. Temperature was serially monitored using a MicroTherma 2 rectal probe (ThermoWorks, American Fork, UT). Clinical sepsis score was assessed as previously described (55). Briefly, points were assigned in the following categories: appearance: normal (0), lack of grooming (1), piloerection (2), hunched up (3), and above and eyes half closed (4); behavior (unprovoked): normal (0), minor changes (1), less mobile and isolated (2), and restless or very still (3); behavior (provoked): responsive and alert (0) and unresponsive and not alert (3); clinical signs: normal respiratory rate (0), slight changes (1), decreased rate with abdominal breathing (2), and marked abdominal breathing and cyanosis (3); hydration status: normal (0) and dehydrated (5). Serum cytokine levels were analyzed using the LEGENDplex arrays, as per the manufacturer’s protocol (BioLegend, San Diego, CA). All procedures performed on these animals were in accordance with the regulations and established guidelines and were reviewed and approved by the New York University (NYU) School of Medicine Institutional Animal Care and Use Committee or through an ethical review process.

Cellular preparation and flow cytometry

For tumor studies, single-cell suspensions of PDA tumors were prepared for flow cytometry as described previously (56). Briefly, pancreata were placed in cold 2% FACS [PBS with 2% fetal bovine serum (FBS)] with collagenase IV (1 mg/ml; Worthington Biochemical, Lakewood, NJ), trypsin inhibitor (1 mg/ml; EMD Millipore, Billerica, MA), and deoxyribonuclease I (2 U/ml; Promega, Madison, WI) and minced with scissors to submillimeter pieces. Tissues were then incubated at 37°C for 20 min with gentle shaking every 5 min. Specimens were passed through a 70-μm mesh and centrifuged at 350g for 5 min. Cell pellets were resuspended, and cell labeling was performed after blocking FcγRIII/II with an anti-CD16/CD32 monoclonal antibody (mAb) (eBioscience, San Diego, CA) by incubating 1 × 106 cells with 1 μg of fluorescently conjugated mAbs directed against mouse CD3 (17A2), CD4 (GK1.5), CD8 (53.-6.7), CD44 (IM7), CD69 (H1.2F3), LFA-1 (H155-78), PD1 (29F.1A12), LAG-3 (C9B7W), BCL6 (7B1), CXCR5 (L138D7), Gr-1 (RB6-8C5), CD11b (M1/70), TNF-α (MP6-XT22), interferon-γ (IFN-γ) (XMG1.2), interleukin-6 (IL-6) (MP5-20F3), IL-2 (JES65-H4), IL-10 (JES5-16E3), F4/80 (BM8), I-A/I-E (M5/114.15.2), CD117 (2B8), CD115 (AFS98), CD135 (A2F10), Ly6G (1A8), Ly6C (HK1.4), CD11c (N418), CD19 (1D3/CD19), NK1.1 (PK136), B220 (RA3-6B2), and Ter119 (TER-119; all BioLegend); granzyme B (NGZB), Rorγt (AFKJS-9), and EOMES (DAN11mag, all eBioscience); pS6S235/236 (D57.2.2E; Cell Signaling Technology, Danvers, MA); and pAktT308 (545007; R&D Systems, Minneapolis, MN). Human flow cytometry antibodies included CD45 (2D1), CD3 (UCHT1), CD8 (HIT8a), CD4 (A161A1), CD44 (IM7), IFN-γ (4S.B3), CD11b (CBRM1/5), CD14 (M5E2), CD15 (HI98), and HLA-DR (human leukocyte antigen-DR isotype) (L243; all BioLegend). Dead cells were excluded from analysis using Zombie Yellow (BioLegend). Intracellular staining for cytokines, transcription factors, and granzyme B was performed using the Fixation/Permeabilization Solution Kit (eBioscience). Flow cytometry was performed on the Attune NxT Acoustic Focusing Cytometer (Thermo Fisher Scientific, Waltham, MA). FACS was performed on a SY3200 (Sony, Tokyo, Japan). Data were analyzed using FlowJo v.10.1 (Treestar, Ashland, OR). Bone marrow–derived cultures were harvested by aspiration of mouse femurs, stimulated with GM-CSF (20 ng/ml; BioLegend), and harvested on day 5 as we previously described (57). Splenocytes were prepared by manual disruption. PBMCs were prepared for flow cytometry using a Ficoll-Paque PLUS gradient (GE Healthcare, Uppsala, Sweden).

Cell size and volume determination

For cell size determination in adherent myeloid cells, CD11b+ cells were isolated, and cells were plated at 1 × 106 cells/ml in Opti-MEM (minimum essential medium) media with or without phorbol 12-myristate 13-acetate (40 ng/ml) (Sigma-Aldrich). After overnight incubation, cells were fixed in 4% paraformaldehyde (PFA), permeabilized with 0.25% Triton X-100, blocked, and then stained with Phalloidin-488 (Invitrogen, A12379) and 4′,6-diamidino-2-phenylindole (DAPI). Images were acquired at 20× on the CellInsight CX7 (Thermo Fisher Scientific). The Cell Spreading algorithm was used to quantify cell perimeter and area.

Patient-derived organotypic tumor spheroids

PDOTS were prepared as previously described with slight modifications (20, 21). Briefly, human surgically resected tumor specimens were received fresh in Dulbecco’s minimum essential medium (DMEM) on ice and minced to submillimeter pieces in 10-cm petri dishes. Minced tumors were resuspended in DMEM + 10% FBS with collagenase type IV (100 U/ml) to obtain spheroids. Partially digested samples were pelleted, resuspended in fresh DMEM + 10% FBS, and then strained over both 100- and 40-μm filters to generate S1 (>100 μm), S2 (40 to 100 μm), and S3 (<40 μm) spheroid fractions, which were subsequently maintained in ultralow attachment tissue culture plates. An aliquot of the S2 fraction was pelleted and resuspended in type I rat tail collagen, and the spheroid-collagen mixture was then injected into the center gel region of the DAX-1 3D microfluidic cell culture chip (AIM Biotech, Singapore). After 30 min at 37°C, collagen hydrogels containing PDOTS were hydrated with media with indicated treatments. Spheroids were harvested on day 3 for analysis by flow cytometry. Images were captured on a Nikon Eclipse 80i fluorescence microscope equipped with Z-stack (Prior) and a CoolSNAP charge-coupled device camera (Roper Scientific, Trenton, NJ). Image capture and analyses were performed using NIS-Elements AR software package (Nikon, Tokyo, Japan). Human biological samples were sourced ethically, and their research use was in accord with the terms of the informed consents under an Institutional Review Board–approved protocol.

Mechanical stimulation of adherent cells

For adherent myeloid cells, cells were seeded at 1.5 × 105 cells per well on top of 96-well glass-bottom Easy Coat plates of varied substrate stiffness (50 and 0.2 kPa) populated with quinone groups, which allow strong nucleophile bonds with cellular surface proteins (Matrigen, Brea, CA). Control standard endotoxin (200 ng/ml; Cape Cod Incorporated, East Falmouth, MA) was added to select wells at equal concentrations. After 18 hours, secretion of TNF-α in the supernatant was measured by Quantikine enzyme-linked immunosorbent assay according to the manufacturer’s protocol (R&D Systems).

Quantitative polymerase chain reaction

For qPCR, total RNA was extracted using an RNeasy mini kit (QIAGEN, Valencia, CA), and complementary DNA (cDNA) was synthesized using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA). Real-time qPCR was performed in duplicate for each sample using the Bio-Rad Real-Time PCR System (Bio-Rad). Each reaction mixture contained 10 μl of SYBR Green Master Mix (Applied Biosystems), 0.5 μl each of forward and reverse primers (Invitrogen) and 3 μl of cDNA (corresponding to 50 ng of RNA). The qPCR conditions were as follows: 50°C for 2 min, 95°C for 10 min, followed by 40 cycles at 95°C for 15 s, and 60°C for 1 min. Amplification of specific transcripts was confirmed by melting curve profiles generated at the end of the PCR program. Expression levels of target genes were normalized to the expression of glyceraldehyde phosphate dehydrogenase (internal control) and calculated on the basis of the comparative cycle threshold (Ct) method (2−ΔΔCt). The Rb1 primer sequences used in the study were CAGGGCTGTGTTGACATCGGAGTA (forward) and TCCACGGGAAGGACAAATCTGTTC (reverse).

Immunohistochemistry and microscopy

For histological analysis, specimens were fixed with 10% buffered formalin, dehydrated in ethanol, embedded with paraffin, and stained with hematoxylin and eosin. For immunofluorescence staining of enriched mouse leukocyte populations, cells were fixed in 4% PFA on ice. Cells were then probed with antibodies directed against CD45 (30-F11, BD Biosciences), CD3 (17A2), CD11b (M1/70), Piezo1 (Polyclonal; Novus Biologicals, Littleton, CO), and DAPI (Vector Labs, Burlingame, CA). Images were acquired using a Zeiss LSM700 confocal microscope with ZEN 2010 software (Carl Zeiss, Thornwood, NY) and analyzed using ImageJ.

Single-cell RNA sequencing

After confirming the integrity of the cDNA, quality of the libraries, number of cells sequenced, and mean number of reads per cell, as a quality control, we used the cellranger package to map the reads and generate gene-cell matrices. A quality control was then performed on the cells to calculate the number of genes, unique molecular identifiers, and the proportion of mitochondrial genes for each cell using scSeqR R package (v0.99.0) (https://github.com/rezakj/scSeqR), and the cells with low number of covered genes (gene count, <500) and high mitochondrial counts (mt genes, >0.1) were filtered out. Then, the matrix was normalized on the basis of ranked geometric library size factor (ranked glsf) using the scSeqR. Geometric library size factor normalization is a common normalization method used by popular tools such as DESeq2 (58); however, here, we use only the top-ranked genes (top 500 genes sorted by base mean) to correct for dropout size factors. General gene statistics were then performed to calculate gene dispersion, base mean, and cell coverage to use to build a gene model for performing principal components analysis (PCA). Genes with high coverage (top 500 ranked by base mean) and high dispersion (dispersion, >1.5) were chosen for PCA. Top highly expressed genes and highly dispersed/variable genes are a good list of genes to define the identity of the cells. Once PCA analysis was completed on our model, gene clustering was performed on principal components with high SDs (top 10 PCs) (scSeqR options; clust.method = “kmeans,” dist.method = “euclidean,” index.method = “silhouette”), and t-distributed stochastic neighbor embedding (t-SNE) was generated. Marker genes for each cluster were then determined on the basis of fold change and adjusted P value (t test), and average gene expression for each cluster was calculated using the scSeqR. Marker genes were visualized on heatmaps, bar plots, and box plots for each cluster and were used to determine the cell types using the ImmGen database (www.immgen.org/). Cell type identities of known population markers were assigned as follows: GMPs, ElanehiPrss34hiMcpt8hiCtsghiPrtn3hi; macrophages, Cldn13hiGlypahiSlc4a1hiTspo2hiTrim10hi; monocytes, DstnhiPrtn3hiFcnbhiLtb4r1hiAldh2hi; neutrophils, Cd77hiMgst1hiCybbhiTkyhiIfitm6hi; MDPs, Igst6hiEmbhiAp3s1hiNclhiMpn1hiCcr2hi; cytidine diphosphate, S100a6hiLgals3hiFcer1ghiTyrobphiCrip1hi.

Chromatin immunoprecipitation sequencing

All of the reads from the ChIP-seq for each sample were mapped to the mouse reference genome (mm10/GRCm38.74) using Bowtie2 (v2.2.4), and duplicate reads were removed using Picard tools (v.1.126) (http://broadinstitute.github.io/picard/) (59). Low-quality mapped reads [MQ (mapping qualities), <20] were removed from the analysis. Read per million–normalized BigWig files were generated using BEDTools (v.2.17.0) and the bedGraphToBigWig tool (v.4) (60). Peak calling was performed using MACS (v1.4.2), and peak count tables were created using BEDTools (61). Differential binding analysis was performed using DESeq2 (62). ChIPseeker (v1.8.0) R package and HOMER (v4.8) were used for peak annotations, and motif discovery was performed using HOMER (63). ngs.plot (v2.47) and ChIPseeker were used for transcription start site binding site visualizations and quality controls (64). Kyoto Encyclopedia of Genes and Genomes pathway analysis and Gene Ontology analysis were performed using the clusterProfiler R package (v3.0.0) (64). To compare the level of similarity among the samples and their replicates, we used two methods: classical multidimensional scaling or PCA and Euclidean distance–based sample clustering. Downstream statistical analyses and generating plots were performed in R environment (v3.1.1) (www.r-project.org/).

Bacterial culture

To evaluate anaerobic bacteria in whole blood and peritoneal fluid, sixfold serial dilutions of the samples were performed in sterile PBS. Each dilution was plated onto blood agar plates (BD Difco) and incubated anaerobically at 37°C for 60 to 72 hours. Viable counts of bacteria were calculated and interpreted as colony-forming units per milliliter.

Electrophysiology

For patch clamping, myeloid cells were patched after magnetic isolation from murine spleens using the CD11b+ cell Isolation Kit (Miltenyi Biotec, Auburn, CA). Cells were plated on 15-mm round-glass 0.01% poly-l-lysine–coated coverslips for 30 min and then washed thoroughly with extracellular solution immediately before patching. Patch-clamp experiments were performed in whole-cell configuration using an Axon MultiClamp 700A amplifier and an Axon Digidata 1550A digitizer (Molecular Devices) at room temperature. Currents were low-pass filtered with an eight-pole Bessel filter (−3 dB at 1 Hz) and digitized at 3 kHz (Digidata 1550A, Molecular Devices) using pClamp v10.5 software (Molecular Devices). Patch electrodes were manufactured (Zeitz puller, Germany) using borosilicate glass (1.5-mm outer diameter; World Precision Instruments, Sarasota, FL) and had tip resistances of 2.5 to 3.5 megohms when filled with the following: 133 mM CsCl, 10 mM Hepes, 5 mM EGTA, 1 mM CaCl2, 1 mM MgCl2, 4 mM MgATP, and 0.4 mM Na2GTP (pH 7.3 with CsOH; osmolarity, 280 ± 10 mOsm) as previously reported (3). The pipette solution was supplemented with 30 μM Yoda1 to maximize pressure sensitivity during patch-clamp recordings. The extracellular solution consisted of 127 mM NaCl, 3 mM KCl, 1 mM MgCl2, 10 mM Hepes, 2.5 mM CaCl2, and 10 mM glucose (pH 7.3 with NaOH; osmolarity, 300 ± 10 mOsm) as previously reported (19). Mechanically activated whole-cell currents were elicited as previously described using a Clampex-controlled high-speed pressure clamp system (HSPC-2, ALA Scientific Instruments) (19). Data were not corrected for the liquid junction potential, which was calculated to be 5.0 mV. The whole-cell capacitance and series resistance were compensated to levels greater than 80%. Currents were expressed as picoamperes per picofarad after correcting for cell size by dividing membrane current by the cell capacitance. Leak subtraction was performed after acquisition. Representative patch-clamp recordings were compiled in Origin 8.1 (OriginLab, Northampton, MA).

Statistical analysis

Data are presented as means ± SE. Statistical significance was determined by Student’s t test and the log-rank test using GraphPad Prism 7 (GraphPad Software, La Jolla, CA). P ≤ 0.05 was considered significant. Significance for Gene Set Enrichment Analysis and differential gene expression based on scRNA-seq was determined using the Wilcoxon rank sum test with Bonferroni multiple comparison correction. Comparisons for more than two groups were calculated using two-way analysis of variance (ANOVA), followed by Bonferroni multiple comparison correction. Data on gene expression in human tissues were derived from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/). Survival was measured according to the Kaplan-Meier method and analyzed by log rank. The publicly available dataset GeneAtlas U133A, gcrma (65) on the BioGPS site was used to obtain cell type–specific gene expression patterns.

SUPPLEMENTARY MATERIALS

immunology.sciencemag.org/cgi/content/full/5/50/eabb5168/DC1

Fig. S1. Piezo1 is expressed in myeloid cells.

Fig. S2. Piezo1 does not affect myeloid cell size.

Fig. S3. Inhibition of Piezo1 protects against polymicrobial sepsis.

Fig. S4. Acetyl-histone H4 ChIP-seq of bone marrow cells.

Table S1. Raw data file (Excel spreadsheet).

REFERENCES AND NOTES

Acknowledgments: We acknowledge N. Berger and K. Kennedy for technical support. Funding: We acknowledge the use of the Genome Technology, Experimental Pathology, Microscopy, and Flow Cytometry core facilities at the NYU School of Medicine. These shared resources are partially supported by the Cancer Center Support Grant, P30CA016087, at the Laura and Isaac Perlmutter Cancer Center. This work was supported by NIH grants CA168611 (to G.M.), CA203105 (to G.M.), CA215471 (to G.M.), CA19311 (to G.M.), DK106025 (to G.M.), UL1 TR001445 (to J.I.K.), Department of Defense grant CA170450 (to G.M.), and the Deutsche Forschungsgemeinschaft grant AY 126/1-1 (to B.A.). The research described herein was also funded in part by Pfizer. Author contributions: B.A. conceived and performed most of the experiments, analyzed and interpreted the data, performed statistical analysis, and wrote the manuscript. R.C. conceived and performed experiments, analyzed and interpreted the data, performed statistical analysis, and wrote the manuscript. J.I.K. conceived and performed all electrophysiological experiments, analyzed and interpreted the data, performed statistical analysis, and edited the manuscript. D.W. performed PDOTS experiments, performed statistical analysis, and edited the manuscript. S.A.A.S. assisted in the orthotopic tumor and polymicrobial sepsis experiments. R.A. provided technical assistance for the execution of orthotopic tumor experiments. P.P. helped with the design and analysis of flow cytometry experiments. A.S. conceived and performed the ChIP-seq and immunofluorescence experiments. S.P. performed all bacterial cultures. J.L. performed experiments for the generation of scRNA-seq data and edited the manuscript. B.D. provided assistance for the design and execution of qPCR experiments and edited the manuscript. W.W. helped with the execution of orthotopic tumor experiments and edited the manuscript. G.W. provided assistance for polymicrobial sepsis experiments and edited the manuscript. M.B. helped with genotyping and orthotopic tumor experiments. S.K.B.L. provided technical assistance for the execution of orthotopic tumor experiments. A.K.-J. assisted in the analysis of the scRNA-seq and ChIP-seq data. D.S. assisted in the analysis of bacterial culture experiments. W.A.C. conceived and assisted in the analysis of electrophysiological experiments and wrote the manuscript. G.M. conceived all experiments, wrote the manuscript, obtained funding, and supervised the project. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The scRNA-seq data and ChIP-seq data are deposited in the Gene Expression Omnibus database under accession number GSE155340. All other data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials.
View Abstract

Stay Connected to Science Immunology

Navigate This Article