Department of Pathology, Joint Program in Transfusion Medicine, Harvard Medical School, and Department of Laboratory Medicine, Children's Hospital Boston, Karp Family Research Building, Room 10214, Boston, MA 02115
Edited by Solomon H. Snyder, Johns Hopkins University School of Medicine, Baltimore, MD, and approved August 9, 2006 (received for review July 7, 2006)
Neutrophil spontaneous death plays essential roles in neutrophilhomeostasis and resolution of inflammation, whereas the underlyingmolecular mechanisms are still ill-defined. Neutrophils diebecause of programmed cell death or apoptosis. However, treatmentwith inhibitor of caspases, which are responsible for the majorityof apoptotic cell deaths, does not prevent the spontaneous deathof neutrophils. PKB/Akt possesses prosurvival and antiapoptoticactivities in a variety of cells. In this study, we show thatAkt activity decreases dramatically during the course of neutrophildeath. Both phosphatidylinositol 3-kinase and Akt inhibitorsenhance neutrophil death. Conditions delaying neutrophil death,such as treatment with granulocyte–macrophage colony-stimulatingfactor, granulocyte colony-stimulating factor, or IFN-, restoreAkt activity. Finally, we demonstrate that neutrophils depletedof PTEN, a phosphatidylinositol 3'-phosphatase that negativelyregulates Akt activity, live much longer than WT neutrophils.Thus, we establish Akt deactivation as a causal mediator ofneutrophil spontaneous death.
apoptosis | PTEN
PNAS | October 3, 2006 | vol. 103 | no. 40 | 14836-14841
Neutrophils are the most abundant cell type among circulatingwhite blood cells and constitute the first line of host defenseagainst invading pathogens (bacteria, fungi, viruses, etc.)(1–3). These cells are terminally differentiated and usuallyhave a very short lifespan (1–4 days in tissues). Theydie via spontaneous programmed cell death (apoptosis). The dailyturnover of human neutrophils is 0.8–1.6 x 109 cells perkg of body weight. Properly regulated death program is essentialfor neutrophil homeostasis. Augmented neutrophil death leadsto a decrease of neutrophil counts in the blood (neutropenia),which will increase the chance of contracting a bacterial orfungal infection. On the other hand, delayed neutrophil deathelevates neutrophil counts in the blood (neutrophilia), whichoften is associated with pathological conditions such as bacterialinfection, myeloid leukemia, and acute myocardial infarction.Programmed neutrophil death is also an essential cellular eventfor maintaining neutrophil numbers in the sites of infectionand inflammation. Neutrophils are recruited to the infectedtissues to engulf, kill, and digest invading microorganisms.However, the enzymes and reactive oxygen species released byneutrophils also can damage the surrounding tissues. Thus, thedeath program in neutrophils need to be well controlled to providea nice balance between their immune functions and their safeclearance. Delayed clearance of neutrophils in inflamed tissuescauses unwanted and exaggerated tissue inflammation (4–7).
PtdIns(3,4,5)P3/Akt signaling pathway possesses prosurvivaland antiapoptotic activities in a variety of cell types. PtdIns(3,4,5)P3(phosphatidylinositol 3,4,5-trisphosphate) contains two hydrophobicfatty acids and, therefore, are mainly localized on the plasmamembrane (8). PtdIns(3,4,5)P3 exerts its function by mediatingprotein translocation via binding to their pleckstrin homolog-domains(9, 10). PKB/Akt, a serine/threonine protein kinase with oncogenicand antiapoptotic activities, is one of the major downstreamfactors of PtdIns(3,4,5)P3 (11, 12). Akt contains apleckstrinhomolog domain, which specifically binds PtdIns(3,4,5)P3. ThePtdIns(3,4,5)P3-mediated membrane translocation of Akt is essentialfor its phosphorylation and activation. Activated Akt, in turn,phosphorylates a variety of proteins, including several associatedwith cell survival/death pathways such as BAD, Forkhead, ASK1,and NF-B, leading to diminished apoptotic cell death (12, 13).PtdIns(3,4,5)P3 level on the plasma membrane is regulated byphosphatidylinositol 3-kinases (PI3K) (12, 14–16) andthe tumor suppressor PTEN (Phosphatase and tensin homologuedeleted on chromosome 10), a phosphatidylinositol 3'-phosphatasethat converts PtdIns(3,4,5)P3 to PtdIns(4,5)P2 (17, 18).
Recently, we demonstrated that deactivation of PtdIns(3,4,5)P3/Aktsignal characterizes both caspase-dependent and -independentcell death (19). In the present study, we investigated the contributionof PI3K-Akt pathway in neutrophil spontaneous death. We demonstratethat PtdIns(3,4,5)P3/Akt signal is deactivated significantlyduring neutrophil death. Inhibition of PtdIns(3,4,5)P3/Akt signalfurther promotes neutrophil death. Moreover, augmentation ofPtdIns(3,4,5)P3/Akt signal by depleting PTEN prevents neutrophilspontaneous death. Thus, we establish Akt deactivation as acausal mediator in neutrophil spontaneous death.
Using this system, we investigated the role of caspases in neutrophilspontaneous death. As reported in ref. 20, the broad-spectrumcaspase inhibitor zVAD-fmk blocked caspase-dependent apoptosis,such as staurosporine-induced death of HEK293 cells and H2O2-eliciteddeath of HeLa cells, whereas the same drug did not suppressspontaneous neutrophil death at all (Fig. 1 C–E).
Akt Signaling Pathway Is Down-Regulated During Neutrophil Spontaneous Death. We recently showed that deactivation of PtdIns(3,4,5)P3/Aktsignal characterizes both caspase-dependent and -independentcell death. To investigate whether the PtdIns(3,4,5)P3/Akt survivalpathway is involved in neutrophil spontaneous death, we examinedthe activation of this signal in neutrophil spontaneous death.Akt is recruited onto the plasma membrane through its specificbinding to PtdIns(3,4,5)P3. Only the Akt molecules on the plasmamembrane can be phosphorylated and activated by two phosphatidylinositol-dependentprotein kinases and get activated, thus Akt phosphorylationhas been used widely as an indicator of Akt activation (12,14, 15, 19, 21–23). Our results show that during the courseof neutrophil death, levels of phospho-Akt decrease dramatically,whereas levels of total Akt do not change (Fig. 2A). This resultwas detected in all of the blood donors we examined (>10donors). The level of phospho-Akt declines by more than one-halfin only 15 h in culture (Fig. 2B). We next measured the Aktactivity directly. We used glycogen synthase kinase-3 (GSK-3),a substrate of Akt (24, 25), as a marker for in vivo Akt activation.Levels of phospho-GSK-3 decline during neutrophil death, witha time course similar to the decline in levels of phospho-Akt.By contrast, total GSK-3 levels do not change (Fig. 2C). Noticeably,the Akt activity started to decrease before any apoptotic morphologychanges were detected, indicating that Akt deactivation mightbe a causal mediator of neutrophil death. The decreases in phospho-Aktare selective. We measured activity of MAPK p44/42 (Erk1/2)pathway (Fig. 2D). During neutrophil spontaneous death, we detectno alteration in any of these phosphokinases 3–48 h afterblood was drawn from the donors.
Inhibition of PtdIns(3,4,5)P3/Akt Signaling Promotes Neutrophil Death. To test the hypothesis that deactivation of PtdIns(3,4,5)P3/Aktsignaling is a causal mediator of neutrophil spontaneous death,we examined whether inhibition of this pathway can promote furtherneutrophil death. We used a newly developed Akt inhibitor, Aktinhibitor I, to suppress the PtdIns(3,4,5)P3/Akt signaling inneutrophils. Moreover, because PI3K is upstream of Akt and inhibitionof PI3K is associated with deactivation of Akt, we also treatedneutrophil cultures with the PI3K inhibitor wortmannin. Treatmentwith these drugs markedly deactivates Akt without altering totalAkt levels (Fig. 3A and B). Both drugs promoted neutrophil deathas monitored by FACS analysis (Fig. 3C).
Factors Preventing Neutrophil Spontaneous Death Restore Akt Kinase Activity. If Akt deactivation mediates neutrophil spontaneous cell death,then treatments that prevent neutrophil death should inhibitthe deactivation. Many extracellular factors, such as granulocyte–macrophagecolony-stimulating factor (GM-CSF), granulocyte colony-stimulatingfactor (G-CSF), and IFN- (LPS), have been shown to block neutrophilspontaneous death (4–7, 26, 27). We have confirmed theantineutrophil death effect of these reagents in our experimentalsystem (Fig. 4A). In addition, we showed that these factorsalso can prevent deactivation of Akt during neutrophil deathwith no alterations in total Akt levels (Fig. 4 B and C). Theseresults are consistent with our hypothesis that Akt deactivationis an essential causal mediator in neutrophil spontaneous death.
Augmentation of PtdIns(3,4,5)P3/Akt Signal by PTEN Depletion Prevents Neutrophil Spontaneous Death. If Akt deactivation mediates neutrophil spontaneous death, thenelevating this pathway should diminish neutrophil death. Totest this idea, we tried to augment PtdIns(3,4,5)P3/Akt signalingby disrupting PTEN gene. The tumor suppressor PTEN is a phosphatidylinositol3'-phosphatase that converts PtdIns(3,4,5)P3 to PtdIns(4,5)P2.Depletion of this lipid phosphatase leads to accumulation ofPtdIns(3,4,5)P3 on the plasma membrane and, thus, elevationof PtdIns(3,4,5)P3/Akt signaling. To examine the effect of elevatedPtdIns(3,4,5)P3/Akt signaling on neutrophil death, we isolatedneutrophils from PTEN knockout mice. Because of the early embryoniclethality of conventional Pten–/– mice (28), weused a conditional Pten knockout mouse, in which two loxP sequenceswere inserted on either side of the exon 5 of PTEN encodingthe phosphatase domain (Fig. 8, which is published as supporting informationon the PNAS web site). We then crossed this mouse with a myeloid-specificCre line, in which the Cre recombinase gene was inserted intothe endogenous M lysozyme locus (Fig. 8). Mice that are homozygousfor this allele are viable, fertile, normal in size, and donot display any gross physical or behavioral abnormalities (datanot shown). The expression of WT PTEN protein is completelyabolished in neutrophils isolated from either Cre+/–;PTENloxP/loxP or Cre+/+;PTEN loxP/loxP mice (Fig. 5A and data notshown). Thus, Cre-mediated deletion of the loxP-flanked PTENgene in myeloid cells is highly efficient. The amount of Crerecombinase expressed from only one copy of the Cre gene isenough to initiate PTEN deletion (Fig. 5A).
In neutrophils depleted of the PTEN gene, PtdIns(3,4,5)P3/Aktsignaling, monitored by Akt phosphorylation, is dramaticallyenhanced (Fig. 5A). This result is consistent with the roleof PTEN as a PtdIns(3,4,5)P3 phosphatase. PTEN-null neutrophilsdifferentiate normally, and neutrophil count in peripheral bloodis the same between PTEN knockout and WT mice (Fig. 5B and datanot shown). We measured the cell death rate of neutrophils directlyisolated from the bone marrow. We found that Pten-null neutrophilslive much longer than WT neutrophils, and this effect was observedat all time points examined. Only 5% of WT neutrophils couldlive>72 h in the culture, whereas 40% PTEN-null neutrophilswere detected healthy under the same condition (Fig. 5C). Delayeddeath of PTEN-null neutrophils can be reversed by treatmentwith the PI3K inhibitor wortmannin, suggesting this delayedneutrophil death is directly mediated by PtdIns(3,4,5)P3 signaling(Fig. 5D). The same result was obtained when in vitro-generatedneutrophils were used (Fig. 6). More than 85% of cells in WTgranulocyte colonies died at day 22, whereas only <10% ofPTEN-null neutrophils went to apoptosis.
In the myeloid-specific Cre line, the Cre recombinase gene wasinserted into the lysozyme locus and, therefore, the endogenouslysozyme gene is disrupted. To ascertain that the phenotypewe observed is caused by PTEN disruption instead of lysozymedeletion, we examined the neutrophils isolated from Cre+/+;PTEN wt/wt mice, which express WT PTEN but contain two mutatedlysozyme alleles. These cells die at a similar rate as WT neutrophils,suggesting that depletion of lysozyme does not affect neutrophilspontaneous death. Furthermore, the diminished cell death ofPTEN-null neutrophils could be reversed by treatment with PI3Kor Akt inhibitors, suggesting that this effect is directly causedby elevated Akt activity in PTEN-null neutrophils (data notshown). In fact, we routinely use only Cre heterozygous micein our assay (Fig. 5 and 6). Thus, effects caused by lysozymedepletion will be minimized.
The molecular mechanism governing this caspase-independent neutrophilspontaneous death is still largely unknown. The susceptibilityof neutrophils to apoptosis appears to depend on the balancebetween proapoptotic and survival (antiapoptotic) signals. Celldeath can be triggered by augmenting proapoptotic signals orattenuating survival signals. In recent years, the activationof proapoptotic pathways in neutrophil spontaneous death hasbeen studied extensively, and several important players suchas BAD, reactive oxygen species, and p38MAPK have been identified(4–7). Nevertheless, the contribution of the deactivationof survival signals to neutrophil death has not yet been investigated.In this study, we reported that the deactivation of a well knownsurvival signal, PI3K-Akt pathway, plays a causal role in neutrophilspontaneous death. Augmentation of PtdIns(3,4,5)P3/Akt signalby PTEN depletion significantly prevents neutrophil spontaneousdeath, thus opening a previously undescribed avenue for interveningin neutrophil death. This study also advances our knowledgeof the molecular mechanism of cell death (particularly the caspase-independentapoptosis) in general and helps us to understand death signalingin other types of cells, such as hematopoietic stem cells, monocytes/macrophages,lymphocytes, and leukemia cells.
The upstream deactivators of Akt remain to be elucidated. Thebest established activator of Akt is PtdIns(3,4,5)P3 generatedby PI3K (12, 15). Conceivably, deactivation of PI3K is responsiblefor Akt's deactivation. PtdIns(3,4,5)P3 level also can be regulatedby the tumor suppressor PTEN and SHIP (SH2-containing inositol5'-phosphatase), which converts PtdIns(3,4,5)P3 to PtdIns(4,5)P2and PtdIns(3,4)P2, respectively (18, 30). PTEN or/and SHIP mightget activated during neutrophil spontaneous death, leading todown-regulation of Akt. Similar with what is discovered in thePTEN-null neutrophils, Gardai et al. (31) reported that thehalf-life of neutrophils depleted of SHIP also was dramaticallyincreased. Akt activation relies on its membrane translocationmediated by its specific association with PtdIns(3,4,5)P3 onthe plasma membrane. Recently, we demonstrated that two inositolphosphates, InsP7 and Ins(1,3,4,5)P4, compete for Akt-pleckstrinhomolog domain binding with PtdIns(3,4,5)P3, providing anotherlevel of regulation for Akt membrane translocation and activation(32). Thus, Akt deactivation also could be a result of elevatedInsP7 or Ins(1,3,4,5)P4 level.
Multicellular organisms defend themselves against invading pathogens(bacteria, fungi, and viruses) by mounting both innate and adaptiveimmune responses. Neutrophil accumulation at sites of infectionis essential for host defense. However, exaggerated accumulationcould be responsible for pathogenesis of many acute and chronicinflammatory diseases, such as pneumonia, asthma, multiple sclerosis,and rheumatoid arthritis. There is a consensus that the rateof neutrophil spontaneous death can determine the number ofneutrophils at the sites of inflammation and, therefore, areimportant in the regulation of inflammation. Results presentedin this paper provide a previously undescribed strategy andtarget for the treatment of a variety of infectious and inflammatorydiseases associated with delayed or enhanced neutrophil death.
Neutrophil spontaneous death also plays essential roles in neutrophilhomeostasis. We examined whether disruption of PTEN can increaseperipheral blood neutrophil count due to delayed spontaneousdeath but did not detect any alteration. Peripheral blood neutrophilcount is decided by multiple cellular processes, such as cytokine-elicitedmobilization from bone marrow, spontaneous death, transmigrationfrom blood to tissues, and clearance by phagocytic cells. Conceivably,depletion of PTEN can affect processes other than spontaneousdeath, and these effects are able to overcome the effect elicitedby delayed neutrophil death, leading to unaltered peripheralblood neutrophil count.
Mouse Neutrophils. The conditional Pten knockout mouse (PTEN loxP/loxP) and themyeloid-specific Cre mouse were purchased from The Jackson Laboratory(Bar Harbor, ME). Mouse genotyping was performed by using astandard protocol provided by The Jackson Laboratory. Mouseneutrophils were prepared from bone marrow, which was isolatedfrom the femur and the tibia of 10-week-old mice. Bone marrowneutrophils are separated by centrifugation over a three-layerPercoll gradient (78%/69%/52%). We routinely obtain 7–8million cells from one mouse, and >90% of them are morphologicallymature neutrophils (donut-shape, segmented nucleus). All animalmanipulations were conducted in accordance with the Animal WelfareGuidelines of Children's Hospital and were monitored by theChildren's Hospital Animal Care and Use Committee. To generatemature neutrophils in vitro, 4 million total bone marrow cellswere cultured in semisolid METHOCULT GF M3534 medium (StemCellTech, Vancouver, BC, Canada) in which only monocytes and granulocytecolonies can be formed. We can routinely obtain 40–60granulocyte colonies from one plate. Neutrophil death in thesecolonies was analyzed as described below.
FACS Analysis of Neutrophil Spontaneous Death. Neutrophils were cultured for an indicated time and stainedby using an Annexin V Detection Kit (Caltag Laboratories, Burlingame,CA) following a protocol provided by the manufacturer. FACSwas performed by using a FACScan flow cytometer (Becton Dickinson,San Jose, CA) equipped with a 488-nm argon laser. Ten thousandcells were collected and analyzed by using the CellQuest software(Becton Dickinson).
Cell Lines and Cell Death Assays. HeLa cells, a human cervical carcinoma-derived cell line, andHEK293 cells, a kidney tumor cell line, were maintained in DMEMwith 10% FBS, 2 mM L-glutamine, and 100 units of penicillin–streptomycinat 37°C with 5% CO2 atmosphere in a humidified incubator.Cell death was induced by the addition of 1 mM H2O2 or 1 µMstaurosporine. Toxicity was assayed 12–14 h after drugexposure by microscopic examination with computer-assisted cellcounting. For staining of dead cells by TUNEL assay, cells werefixed in 4% paraformaldehyde/PBS and then stained by using aTUNEL Assay Kit by following protocols provided by the manufacturer(Molecular Probes, Eugene, OR). The broad-spectrum caspase inhibitor,zVAD-fmk (10 µM), was added 1 h before drug treatments.Cell death was determined as the ratio of dead (TUNEL-positive)-to-totalcell number and quantified by counting 1,000 cells. Westernblot analysis was conduced as described in ref. 33.
Statistical Analysis. Values shown in each figure represent mean ± SD. Statisticalsignificances were calculated with the Student t test. Differenceswere considered significant for P values <0.005.
We thank James Campbell, John Manis, and Li Cai for helpfuldiscussions. D.Z. is supported by National Institutes of Health(NIH) Training Grant HL066987. H.L. was supported by NIH GrantsNS052200 and GM076084 and a research scholar grant from AmericanCancer Society.
Abbreviations: PI3K, phosphatidylinositol 3-kinase; PtdIns(3,4,5)P3, phosphatidylinositol 3,4,5-trisphosphate.
*To whom correspondence should be addressed. E-mail: [email protected]
Author contributions: H.R.L. designed research; D.Z., H.H.,H.J., Y.J., K.K.S., F.L., and J.Y. performed research; Y.L.,M.H., and L.S. contributed new reagents/analytic tools; D.Z.,H.H., H.J., Y.J., K.K.S., F.L., and H.R.L. analyzed data; andH.R.L. wrote the paper.
The authors declare no conflict of interest.
This paper was submitted directly (Track II) to the PNAS office.
© 2006 by The National Academy of Sciences of the USA