iiinova — where your digital self keeps evolving

Daniel Fisch @danielhfisch.bsky.social · 02/10/2026

Yesterday was a special day in my career: I got to open the door to my own research group (and for the first time in my life to my own office 😂)! Thank you to @gimminstitute.bsky.social for giving me this opportunity, it's exciting to finally be here!

2162

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

In summary: TRAM is identified as the release factor enabling receptor-free and sustained myddosome signaling in the TLR pathway. I hope you enjoy the read: rdcu.be/jGdKxErXyAsa

Graphical summary of the proposed model for TLR signaling.

110

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

If long-lived myddosomes sustain inflammation, disrupting them should stop it. Conversely, MyD88 oligomerization inhibitors (MPP, TSI-13-57) dissolved already-formed myddosomes in cells. We detected for the first time native myddosomes in mice, which could be shut down by the same inhibitors.

Top: immunoblots of myddosome immunoprecipitations from splenocytes of mice injected with PBS, LTA or LPS and subsequently injected with TSI-13-57 or vehicle. IRAK2, IRAK4, TRAF6, IKKα, IKKβ, p65 and TBK1 co-precipitate with MyD88 after PAMP injection and are strongly reduced by TSI-13-57, while total MyD88 and actin are unchanged. Bottom: ELISA of serum TNFα, IL-6 and IL-12p40 at 1.5 and 4.5 hours after PAMP injection, with five mice per group. At 4.5 hours TSI-13-57 significantly reduces IL-6 and IL-12p40, whereas TNFα is unaffected.

100

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

We then reconstituted the release mechanism in vitro with three TIR domains. We found that TIRAP restrains MyD88 self-assembly and TRAM relieves that restraint. The purified domains are sufficient to reproduce the behavior seen in cells.

Turbidity assay plotting OD350 over 5 hours for recombinant TIR domains, with lines showing mean and shaded bands showing error. MyD88 TIR alone at 100 µM self-assembles and reaches the highest turbidity. Adding TIRAP TIR at a 1 to 20 ratio (vs MyD88) strongly suppresses assembly. Adding increasing amounts of TRAM TIR, at TIRAP to TRAM ratios of 20:1:1, 20:1:10 and 20:1:20, progressively restores MyD88 assembly towards the level of MyD88 alone.

100

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

We could separate the role of TRAM in the TRIF-arm from its role in the MyD88-arm of TLR signaling based on its localization. Forcing TRAM to the plasma membrane restores myddosome release but abolishes IFN. Endosome-restricted TRAM does the opposite.

Three bar charts showing TNFα, IL-6 and IFNβ secretion measured by ELISA from MyD88-AGF wild-type iBMDMs and ΔTicam2 iBMDMs reconstituted with empty vector, full-length TRAM, or TRAM chimeras carrying the N-terminal plasma membrane targeting sequences of Fyn, Lyn or the TIRAP phosphoinositide-binding motif, stimulated with LTA or LPS. TNFα is restored by all chimeras. IL-6 is restored only by the TIRAP-PBM chimera, which exceeds wild-type levels. IFNβ is restored only by full-length TRAM; all three plasma-membrane-targeted chimeras fail to induce it.

100

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

For releasing myddosomes TRAM acts alongside TIRAP in PI(4,5)P2-rich plasma-membrane microdomains. TRAM mutants, chimeras and CRISPR-Activation showed that the local TIRAP:TRAM ratio, rather than the abundance of either protein, determines whether myddosomes are released.

Fluorescence images of MyD88 localization, MyD88 in grey and nuclei in blue, in MyD88-AGF wild-type iBMDMs and in ΔTicam2 iBMDMs reconstituted with empty vector (EV) or with TRAM WT, G2A, 3X, E180A, C114H or DE88/89AA mutants. The top row is untreated and the bottom row LPS-treated for 1 hour. After stimulation, wild-type cells and cells reconstituted with TRAM WT or E180A show MyD88 assemblies distributed through the cytosol, whereas empty vector, G2A, 3X, C114H and DE88/89AA cells retain MyD88 at the plasma membrane. Scale bars 5 µm.

100

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

Across bacterial infections, TRAM was required for IL-6 but not TNFα production. In infected cells MyD88 again remained at the plasma membrane when TRAM was absent.

Left: three bar charts showing TNFα, IL-6 and IFNβ secretion measured by ELISA from wild-type (white) and Ticam2-/- (orange) BMDMs left uninfected or infected with E. coli (EC) or S. aureus (SA) at MOI 10 for 24 hours. TNFα is comparable between genotypes, IL-6 is strongly reduced in Ticam2-/- cells after both infections, and IFNβ is lost after E. coli but similar after S. aureus infection. Right: fluorescence imaging of MyD88-AGF wild-type and ΔTicam2 iBMDMs infected with E. coli for 3 hours, MyD88 in grey and nuclei in blue. Wild-type cells contain cytosolic myddosomes, whereas MyD88 is retained at the plasma membrane in ΔTicam2 cells, marked by orange arrowheads. Scale bars 5 µm.

100

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

This reflects in transcription: Primary response genes are induced without TRAM, but secondary response genes (e.g. IL-6, IL-12) are not. By RNA-seq, TRAM-deficient macrophages return to baseline by 6 h while WT cells remain activated.

Principal component analysis of RNA-Seq data from wild-type and Ticam2-/- primary BMDMs, untreated or stimulated with LTA or LPS for 2 or 6 hours. PC1 accounts for 81.1% and PC2 for 9.8% of the variance. At 2 hours both genotypes move away from baseline along a shared TLR2/4 early response trajectory. By 6 hours wild-type samples travel further along arrows labelled sustained signaling and TRAM plus TRIF interferon (IFN) response, whereas Ticam2-/- samples curve back towards baseline along an arrow labelled TRAM deficiency.

100

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

The release is what makes TLR signaling last. While WT macrophages sustain NF-κB for hours, TRAM-deficient cells produce a single peak and then terminate signaling early.

Phase plots of p65 nuclear to cytoplasmic ratio, a measure of NF-κB activity, in single cells over 12 hours of stimulation, with 200 cells per condition shown as horizontal traces and activity color-coded from blue (none) to orange (high). Columns show wild-type, ΔMyD88, ΔTirap, ΔTicam1 (TRIF) and ΔTicam2 (TRAM) iBMDMs; the top row is LTA- and the bottom row LPS-stimulation. Wild-type and ΔTicam1 cells sustain NF-κB activity for hours. ΔMyD88 and ΔTirap cells are almost uniformly inactive. ΔTicam2 cells show a bright band of activity only at the start that rapidly returns to baseline.

100

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

Using minute-by-minute co-IPs we found that MyD88 first binds TLR:TIRAP (proto-myddosome), TRAM then releases MyD88 so it can recruit IRAK2/4 and TRAF6 (mature myddosome). Without TRAM, MyD88 stays bound to TLR:TIRAP for hours.

Immunoblots of myddosome immunoprecipitations from MyD88-AGF wild-type and ΔTicam2 iBMDMs stimulated with LPS for 0 to 12 hours, probed for MyD88, TLR4, TIRAP, TRAM, IRAK4, IRAK2, TRAF6 and actin. In wild-type cells, TLR4 and TIRAP associate with MyD88 only within the first minutes, TRAM appears transiently around 30 minutes, and IRAK4, IRAK2 and TRAF6 remain associated for up to 12 hours. In ΔTicam2 cells, TLR4 and TIRAP stay bound for hours, and IRAK and TRAF6 recruitment is delayed, unstable and quickly lost.

100

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

Without TRAM, MyD88 does not move. Proto-myddosomes still nucleate at the plasma membrane but are never released, and no cytosolic myddosomes form. Other adaptor knockouts behaved as expected; ΔTRAM was unique in that myddosome assembly starts but maturation fails.

100

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

To find what builds a myddosome, we performed a genome-wide CRISPR screen, sorting for cells with no myddosomes and no NF-κB activity. We obtained one very unexpected hit: TRAM (Ticam2), the adaptor long exclusively assigned to the MyD88-independent TRIF pathway.

Two rank plots showing gene-level enrichment from MAGeCK robust rank aggregation of the genome-wide CRISPR screen, for LTA- and LPS-stimulated cells sorted for absence of cytosolic myddosomes and NF-κB activation. −log10(MAGeCK score) is plotted against gene rank, with canonical TLR proteins, the oligosaccharyltransferase complex, signaling proteins and controls highlighted in different colors. Rela scores highest in both screens. TRAM is labelled in orange near the top of both rankings, alongside canonical TLR signaling factors MyD88, Traf6, Tirap, Irak4, Tlr2 and Tlr4, whereas Ticam1 (TRIF) ranks far lower.

110

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

We began the study by exploring which cells actually respond to TLR-ligands? Using a side-scatter-based sort we separated macrophages into myddosome-containing responders and myddosome-free bystanders. Only responders switch on the TLR gene program.

Left: Schematic of MyD88-AGF iBMDMs stimulated with PAMPs. Centre left: flow cytometry plot of FSC-W against SSC-A showing the gating strategy, with untreated cells in black and LPS-treated cells (6 h) in blue, separated into SSC-low bystanders and SSC-high responders. Centre right: images of the sorted populations, MyD88 in grey; responders show myddosomes in the cytosol, bystanders do not. Scale bars 10 µm. Right: principal component analysis of RNA-Seq from the sorted populations, in which bystanders separate from responders, and responders cluster with bulk LPS-treated cells.

110

Daniel Fisch @danielhfisch.bsky.social · 23/09/2026

Out now @nature.com! We report a new means of cell communication: receptor-free signal transduction. TLRs engage MyD88 for minutes, yet myddosomes signal for hours. TRAM releases myddosomes from TLRs, letting them mature and sustain signaling in the cytosol. rdcu.be/jGdKxErXyAsa Details below:

43214

Daniel Fisch @danielhfisch.bsky.social · 11/07/2025

Fantastic three days of science at #HFSPmtg25! Thank you @hfspo.bsky.social for hosting us in Melbourne!

030

Daniel Fisch @danielhfisch.bsky.social · 07/07/2025

Thank you @inflammasomelab.bsky.social and @s-burgener.bsky.social for the invitation to visit the IMB@UQ in Brisbane! I had a fantastic time presenting my postdoc work in the Kagan lab (@jkagan1.bsky.social) and talking science!

Sign of the University of Queensland, Brisbane

0114

Daniel Fisch @danielhfisch.bsky.social · 30/06/2025

Thank you @embo.org for a fantastic Fellows’ Meeting 2025. So much amazing science and brilliant young scientists gathered in one place really gives back hope for the future of biomedical research! 🔬🧪

030