LynA regulates an inflammation-sensitive signaling checkpoint in macrophages

Abstract

Clustering of receptors associated with immunoreceptor tyrosine-based activation motifs (ITAMs) initiates the macrophage antimicrobial response. ITAM receptors engage Src-family tyrosine kinases (SFKs) to initiate phagocytosis and macrophage activation. Macrophages also encounter nonpathogenic molecules that cluster receptors weakly and must tune their sensitivity to avoid inappropriate responses. To investigate this response threshold, we compared signaling in the presence and absence of receptor clustering using a small-molecule inhibitor of Csk, which increased SFK activation and produced robust membrane-proximal signaling. Surprisingly, receptor-independent SFK activation led to a downstream signaling blockade associated with rapid degradation of the SFK LynA. Inflammatory priming of macrophages upregulated LynA and promoted receptor-independent signaling. In contrast, clustering the hemi-ITAM receptor Dectin-1 induced signaling that did not require LynA or inflammatory priming. Together, the basal-state signaling checkpoint regulated by LynA expression and degradation and the signaling reorganization initiated by receptor clustering allow cells to discriminate optimally between pathogens and nonpathogens.

Keywords: ITAMs; Src-family kinases; biochemistry; cell signaling; immunology; innate immunity; macrophages; mouse; receptors.

Figure 1.. Surface-marker expression of Csk^AS BMDMs.

Expression of the surface markers F4/80, CD11b, and CD11c in bone marrow-derived macrophages (BMDMs) from Csk^AS mice was assessed by flow cytometry. Data in this figure and those that follow are representative of three or more independent experiments. DOI: http://dx.doi.org/10.7554/eLife.09183.003

Figure 2.. Csk inhibition leads to rapid activation of the SFKs.

Adherent BMDMs generated from Csk^AS or WT mice were treated with 10 µM 3-IB-PP1. The resulting lysates were separated by SDS-PAGE and subjected to immunoblotting with antibodies specific to the inactive and active forms of the Src-family tyrosine kinases (SFKs) (pLyn^Y507 and pSFK^Y416, respectively). An immunoblot of total Syk protein shows the total protein content in each lane. DOI: http://dx.doi.org/10.7554/eLife.09183.004

Figure 3.. Depleted zymosan signals through the Src-family and Syk kinases.

(A) BMDMs were pulse-spun with intact zymosan or zymosan^dep (10 particles per cell) in the presence and absence of the Syk inhibitor BAY 61-3606 (1 µM). Signal transduction was assessed by immunoblotting with antibodies specific to activating phosphorylation sites of Syk, Erk, and Akt. Vinculin immunoblots are shown as loading controls. (B) The effect of the SFK inhibitor PP2 (20 µM) on zymosan^dep stimulation was also assessed. See Figure 3—figure supplement 1 for a model of signaling induced by intact and depleted zymosan. DOI: http://dx.doi.org/10.7554/eLife.09183.005

Figure 3—figure supplement 1.. Signaling through intact and depleted zymosan has different requirements for Syk activation.

After full depletion of TLR2 agonists by repeated boiling, sonication, and hot alkali treatment, zymosan^dep should activate BMDMs exclusively through the Src and Syk kinases in a Dectin-1-dependent manner. DOI: http://dx.doi.org/10.7554/eLife.09183.006

Figure 4.. SFK activation in the absence of receptor clustering fails to stimulate downstream signaling in BMDMs.

(A) Csk^AS BMDMs were pulse-spun with 3-IB-PP1 (10 μM) or with zymosan^dep (10 particles per cell). Signal transduction was assessed by immunoblotting with antibodies specific to activating phosphorylation sites of the SFKs, Syk, PLCγ2, and Erk (arrow). Total Erk1 is shown as a loading control. (B) Csk^AS BMDMs were pulse-spun in the presence and absence of 3-IB-PP1, and Erk phosphorylation was assessed by immunoblot. Total Erk1/2 is shown as a loading control. (C) WT BMDMs were treated with zymosan^dep for 5 min with varying concentrations of 3-IB-PP1 and analyzed by immunoblot. See Figure 4—figure supplement 1 for additional tests of 3-IB-PP1 specificity for Csk^AS, Figure 4—figure supplement 2 for evidence of that 3-IB-PP1 cotreatment suppresses signaling through the CSF-1 receptor, and Figure 4—figure supplement 3 for evidence that actin-remodeling agents cannot restore 3-IB-PP1-induced downstream signaling. DOI: http://dx.doi.org/10.7554/eLife.09183.007

Figure 4—figure supplement 1.. Further tests of 3-IB-PP1 specificity for Csk^AS.

(A) WT BMDMs were treated for 2 min with recombinant M-CSF (10 or 50 ng/ml) in the presence or absence of 3-IB-PP1 (10 µM). Treatment with 3-IB-PP1 does not affect M-CSF-induced phosphorylation of Erk or Akt. (B) WT and Csk^AS BMDMs were treated to a time course of 3-IB-PP1. Only Csk^AS cells showed any measurable change in tyrosine phosphorylation (4G10 + pY20 anti-pTyr antibodies). DOI: http://dx.doi.org/10.7554/eLife.09183.008

Figure 4—figure supplement 2.. Inhibitory effect of 3-IB-PP1 on M-CSF-induced signaling in Csk^AS cells.

Csk^AS BMDMs were treated for 5 min with recombinant M-CSF (2, 5, 20, or 50 ng/ml) in the presence or absence of 3-IB-PP1 (10 µM). Addition of 3-IB-PP1 impairs M-CSF induction of Akt and Erk phosphorylation (arrows), suggesting an active, inhibitory effect of 3-IB-PP1 treatment that blocks downstream signaling. DOI: http://dx.doi.org/10.7554/eLife.09183.009

Figure 4—figure supplement 3.. Actin-remodeling agents do not restore signaling downstream of 3-IB-PP1 treatment.

BMDMs were stimulated with 3-IB-PP1 or zymosan^dep in combination with the actin-remodeling agents Cytochalasin D (CytD, 10 µM), Latrunculin A (LtrA, 0.5 µM), and Jasplakinolide (Jas, 1 µM). Although each compound affected zymosan^dep signaling, none restored Erk phosphorylation after 3-IB-PP1 treatment (CytD n = 2). Note that these images were developed with high signal, so the basal Erk phosphorylation looks high even though the cells have not been activated. The important point is that there is no increase in Erk phosphorylation with 3-IB-PP1 treatment. DOI: http://dx.doi.org/10.7554/eLife.09183.010

Figure 5.. Inflammatory priming is required for SFK activation to produce downstream signaling in the absence of receptor clustering.

(A) Csk^AS BMDMs were incubated 12–16 hr in non-priming medium or in medium containing 25 U/ml IFN-γ. Signal transduction after 3-IB-PP1 or zymosan^dep treatment was assessed by immunoblotting with antibodies specific to activating phosphorylation sites of DAP12, Syk, LAT, PLCγ2, PLCγ1, and Erk (arrow). Total Erk1 is shown as a loading control. (B) Signal transduction in the PI3K/Akt and JNK pathways was assessed by immunoblotting with antibodies specific for phosphorylated c-Cbl, PI3K, Akt (arrow), and JNK (arrow). Total JNK is shown as a loading control. NFAT1 activation was assessed by the faster migration of cellular NFAT1 upon dephosphorylation. See Figure 5—figure supplement 1 for further characterization of signaling in primed and unprimed cells. DOI: http://dx.doi.org/10.7554/eLife.09183.011

Figure 5—figure supplement 1.. Further characterization of signaling in primed and unprimed BMDMs.

(A) Treatment of IFN-γ-primed and unprimed Csk^AS BMDMs with 3-IB-PP1 led to activating phosphorylation of the cytoskeleton-associated SFK substrates HS1, Vav, and FAK. (B) Comparison of background Erk phosphorylation in unprimed and IFN-γ-primed Csk^AS BMDMs with pulse-spin only. DOI: http://dx.doi.org/10.7554/eLife.09183.012

Figure 6.. Activated SFKs initiate negative-regulatory ITIM pathways.

Immunoblots show phosphorylation of negative regulatory proteins Sirpα, SHIP1, and SHP-1 in unprimed and IFN-γ-primed Csk^AS BMDMs treated with 3-IB-PP1 or zymosan^dep. See Figure 6—figure supplement 1 for evidence that the negative-regulatory adaptor Dok-3 is also activated in response to 3-IB-PP1 treatment. DOI: http://dx.doi.org/10.7554/eLife.09183.013

Figure 6—figure supplement 1.. Dok-3 activation by 3-IB-PP1.

Co-immunoprecipitation and immunoblot experiments show that 3-IB-PP1 treatment of Csk^AS BMDMs induces Dok-3 association with SHIP1 and Dok-3 tyrosine phosphorylation. DOI: http://dx.doi.org/10.7554/eLife.09183.014

Figure 7.. Lyn upregulation after priming coincides with Erk signaling after SFK activation.

(A) Csk^AS BMDMs were incubated in non-priming medium or in priming medium containing 25 U/ml IFN-γ and then analyzed directly or treated with 3-IB-PP1 for 5 min. Erk phosphorylation and Lyn expression were assessed by immunoblot. See Figure 7—figure supplement 1 for uncropped LynA blots. (B) Csk^AS BMDMs were primed overnight in 25 U/ml IFN-γ or 10 ng/ml GM-CSF. Erk phosphorylation after 3-IB-PP1 treatment and basal Lyn expression are shown. (C) Immunoblots show basal expression of Lyn, Hck, Fgr, Syk, and PLCγ2 with and without priming. DOI: http://dx.doi.org/10.7554/eLife.09183.015

Figure 7—figure supplement 1.. Uncropped Lyn blots.

(A) Blots displayed in cropped format in Figure 7A and (B) Figure 7B are shown here in their entirety. Note that total Lyn protein disappears with 3-IB-PP1 treatment. This observation is introduced in the main text during the discussion of Figure 8. DOI: http://dx.doi.org/10.7554/eLife.09183.016

Figure 8.. IFN-γ-potentiated signaling through Erk after Csk inhibition depends on Lyn expression.

(A) Signaling in IFN-γ-primed Csk^AS and Lyn^−/−Csk^AS BMDMs was assessed by immunoblotting as described previously. Erk phosphorylation is highlighted by a single arrow, total LynA level is highlighted by a double arrow. (B) Immunoblots show SFK expression in unprimed and IFN-γ-primed WT and Lyn^−/− BMDMs. See Figure 8—figure supplement 1 for evidence that Lyn-deficient BMDMs display normal surface markers. DOI: http://dx.doi.org/10.7554/eLife.09183.017

Figure 8—figure supplement 1.. Lyn deficiency does not impair BMDM generation.

Comparison of surface markers in Lyn^−/−Csk^AS and Csk^AS BMDMs. DOI: http://dx.doi.org/10.7554/eLife.09183.018

Figure 9.. LynA is selectively degraded after Csk inhibition.

(A) Csk^AS and WT BMDMs were treated with 3-IB-PP1. Total kinase levels were detected by immunoblotting with antibodies to LynA, LynA+LynB, Hck, and Fgr. Total Syk is shown as a loading control. (B) SFK levels for the first 5 min of 3-IB-PP1 treatment were quantified by densitometry. Error bars reflect the standard deviation from three separate experiments. (C) Antibodies were used to immunoprecipitate Hck or LynA from BMDM lysates with or without treatment with 3-IB-PP1 for 30 s. The ubiquitination of each immunoprecipitate and the specificity and efficiency of the immunoprecipitation procedures were assessed by immunoblotting. See Figure 9—figure supplement 1 for assignment of immunoblot bands, Figure 9—figure supplement 2 for direct detection of LynA and activated SFK laddering in whole-cell lysate, and Figure 9—figure supplement 3 for a sequence alignment of LynA and LynB that shows putative ubiquitination and phosphorylation sites unique to LynA. DOI: http://dx.doi.org/10.7554/eLife.09183.019

Figure 9—figure supplement 1.. Identification of individual SFKs in immunoblots of whole-macrophage lysate.

Lysates from WT, Hck-deficient, and Lyn-deficient BMDMs immunoblotted with various antibodies were compared to assign bands to individual SFKs LynA, LynB, Hck, and Fgr. DOI: http://dx.doi.org/10.7554/eLife.09183.020

Figure 9—figure supplement 2.. Laddering of activated Lyn in whole-cell lysate.

We treated Csk^AS BMDMs with 3-IB-PP1 (10 μM) or medium alone and observed laddering in activated SFK and total LynA, but not total Hck or inactive SFK immunoblots. DOI: http://dx.doi.org/10.7554/eLife.09183.021

Figure 9—figure supplement 3.. Predicted ubiquitination site in the LynA insert.

Alignment of the amino acid sequence of the two isoforms of mouse Lyn reveals lysine and tyrosine residues unique to LynA. Lysine K40 in LynA only is predicted to be a site of ubiquitination. DOI: http://dx.doi.org/10.7554/eLife.09183.022

Figure 10.. Csk inhibition by 3-IB-PP1 produces transient LynA activation.

(A) Immunoblots of inactive and active SFKs in Csk^AS BMDMs treated with 3-IB-PP1. Inactive Lyn and Hck were detected with the antibodies pLyn^Y507 and pHck^Y527, respectively. Active Lyn, Hck, and Fgr were detected with the pSFK^Y416 antibody. (B) Levels of inactive SFKs were quantified by densitometry. Error bars reflect the standard deviation from three separate experiments. p-values reflect two-tailed t tests. (C) Levels of active SFKs. A 2-way Anova shows significant pairwise differences between LynA and each other SFK species. See Figure 10—figure supplement 1 for assignment of immunoblot bands and Figure 10—figure supplement 2 for data rendered as fold increase in SFK activation with 3-IB-PP1 treatment. DOI: http://dx.doi.org/10.7554/eLife.09183.023

Figure 10—figure supplement 1.. Identification of individual active and inactive SFKs in immunoblots of whole-macrophage lysate.

Lysates from WT, Hck-deficient, and Lyn-deficient BMDMs immunoblotted with various antibodies were compared to assign bands to individual SFKs. (A) Specific antibodies were used to identify immunoblot bands representing inhibitory-tail-phosphorylated Lyn and Hck. (B) Bands representing activation-loop-phosphorylated Hck, LynA, and LynB were similarly assigned. (C) The identity of the activated Fgr band was confirmed by treating Lyn-deficient Csk^AS cells with 3-IB-PP1 to activate the SFKs, immunoblotting for all active SFKs, partially inactivating the Horseradish peroxidase (HRP) enzyme, and then reprobing with a monoclonal antibody against Fgr. DOI: http://dx.doi.org/10.7554/eLife.09183.024

Figure 10—figure supplement 2.. Fold increase in SFK activation by 3-IB-PP1 treatment.

Data displayed as fold increase from basal. Fgr shows the highest fold increase because its basal activation is very low. LynA and LynB are both activated more rapidly than are Hck and Fgr, and activated LynA is uniquely short-lived. DOI: http://dx.doi.org/10.7554/eLife.09183.025

Figure 11.. LynA persists in primed cells treated with 3-IB-PP1.

(A) Immunoblots from unprimed and IFN-γ-primed Csk^AS BMDM lysates show levels of LynA and corresponding Erk phosphorylation during 3-IB-PP1 treatment. (B) Kinetics of LynA degradation relative to the maximum value within each experiment (primed cells, t = 0). (Quantified by densitometry. Error bars reflect the standard deviation over four independent experiments. The p-value is derived from a ratio paired, two-tailed t test.) See Figure 11—figure supplement 1 for independently normalized comparison of Lyn degradation in primed and unprimed cells. DOI: http://dx.doi.org/10.7554/eLife.09183.026

Figure 11—figure supplement 1.. Priming does not inhibit LynA degradation.

IFN-γ priming does not change the rate of Lyn degradation after Csk inhibition. DOI: http://dx.doi.org/10.7554/eLife.09183.027

Author response image 1.

DOI: http://dx.doi.org/10.7554/eLife.09183.030

Author response image 2.

DOI: http://dx.doi.org/10.7554/eLife.09183.031