. Amyloid fibrils in FTLD-TDP are composed of TMEM106B and not TDP-43. Nature. 2022 May;605(7909):304-309. Epub 2022 Mar 28 PubMed.


Please login to recommend the paper.


  1. What a surprise! Nowadays, one would assume that rather simple pathological abnormalities, such as intracellular amyloid filaments, should have been seen by neuropathologists for a long time. Not so. Suddenly, three independent publications pop up and report the widespread existence of intracellular filaments composed of TMEM106b in many neurodegenerative diseases, as well as normal and pathological aging.

    It is unlikely that these filaments are artifactually generated by the extraction protocols, as they are formed in a strictly age-dependent manner. But why were they not seen before?  One obvious possibility is that detection of TMEM106b fibers requires antibodies to the luminal region, which selectively accumulates within the deposits, and which were not used in previous research.

    These striking findings will completely reset research on TMEM106b and its involvement in FTLD as a risk/protective factor. TMEM106b is well-known in the field, since it has been genetically associated with FTLD-TDP caused by PGRN haploinsufficiency (Van Deerlin et al., 2010). 

    We found some time ago that TMEM106b is a type 2 intramembrane protein, which is predominantly located within late endosomes and lysosomes (Lang et al., 2012). The protein undergoes regulated intramembrane proteolysis similar to the amyloid precursor protein (Brady et al., 2014). Shedding of the ectodomain occurs by an unknown protease, most likely a lysosomal protease, whereas the membrane-retained N-terminal stub is removed by intramembrane cleavage executed by Signal Peptide Peptidase-like 2 (SPPL2a) (Brady et al., 2014), an intramembrane cleaving protease of the GxGD type (Steiner et al., 2000). 

    As for Aβ-peptide generation, proteolytic processing appears to be a prerequisite for TMEM106b fiber formation, since all three manuscripts consistently report a fibrillar core encompassing roughly amino acids 120 to 254. TEMEM106b fibers are found intracellularly. 

    Strikingly, we and many others reported that PGRN deficiency is associated with lysosomal dysfunction (Götzl et al., 2016; Götzl et al., 2018). One could therefore speculate that endosomal/lysosomal TMEM106b is deposited in these organelles during aging and disturbs their physiological function. Alternatively, dysfunctional lysosomes could induce deposition and thus accelerate secondary malfunction. Increased expression and shedding should then drive the disease, a hypothesis that can now easily be checked. A first hint may come from the finding that TMEM106b increased in an age-dependent manner in PGRN knockout animals and exacerbated lysosomal dysfunction (Zhou et al., 2017Götzl  et al., 2014). 

    However, currently we only know that TMEM106b forms intracellular inclusions. It remains unclear if they occur in late endosomes/lysosomes, or are distributed throughout the entire cell body, maybe even due to rupture of these organelles.  Importantly, we do not even know yet if these filaments are associated with disease, or are directly causative. They may still be rather innocent bystanders, like lipofuscin.

    So as always, an important discovery raises a lot of new questions. Answering them may allow us to finally understand the functional TMEM106b/PGRN interconnection, and the mechanism of how lysosomal function is disturbed in FTLD and probably other neurodegenerative diseases as well. 


    . Common variants at 7p21 are associated with frontotemporal lobar degeneration with TDP-43 inclusions. Nat Genet. 2010 Mar;42(3):234-9. PubMed.

    . Membrane Orientation and Subcellular Localization of Transmembrane Protein 106B (TMEM106B), a Major Risk Factor for Frontotemporal Lobar Degeneration. J Biol Chem. 2012 Jun 1;287(23):19355-65. PubMed.

    . Regulated intramembrane proteolysis of the frontotemporal lobar degeneration risk factor, TMEM106B, by signal peptide peptidase-like 2a (SPPL2a). J Biol Chem. 2014 Jul 11;289(28):19670-80. Epub 2014 May 28 PubMed.

    . Glycine 384 is required for presenilin-1 function and is conserved in bacterial polytopic aspartyl proteases. Nat Cell Biol. 2000 Nov;2(11):848-51. PubMed.

    . Impaired protein degradation in FTLD and related disorders. Ageing Res Rev. 2016 May 7; PubMed.

    . Early lysosomal maturation deficits in microglia triggers enhanced lysosomal activity in other brain cells of progranulin knockout mice. Mol Neurodegener. 2018 Sep 4;13(1):48. PubMed.

    . Elevated TMEM106B levels exaggerate lipofuscin accumulation and lysosomal dysfunction in aged mice with progranulin deficiency. Acta Neuropathol Commun. 2017 Jan 26;5(1):9. PubMed.

    . Common pathobiochemical hallmarks of progranulin-associated frontotemporal lobar degeneration and neuronal ceroid lipofuscinosis. Acta Neuropathol. 2014 Mar 12; PubMed.

    View all comments by Christian Haass
  2. This series of three papers clearly shows that amyloid fibrils composed of the C-terminal region of TMEM106b are abundant in FTLD-TDP-43 as well as during aging and some other neurodegenerative disorders. Combined with prior data showing that a polymorphism in TMEM106b alters risk in FTLD-TDP-43 due to progranulin mutations, the new findings make it likely that the aggregation of TMEM106b into an amyloid conformation is directly linked to pathogenesis of not only FTLD-TDP-43 but also other diseases where it forms amyloid.

    This important set of findings should now prompt new studies to understand what leads TMEM106b to aggregate, what the consequences are in animal and cellular models, and how to prevent TMEM106b aggregation.

    View all comments by David Holtzman
  3. These are really exciting findings from structural biology, and they open up many questions about TMEM106B. It's fantastic that multiple groups are seeing similar things, since that helps the field, especially those of us who do not know that much about cryo-EM.

    The first thing I considered is why we had not seen TMEM106B in inclusions from neuropathological specimens, since our group had stained many brain sections in 2013 (Busch et al., 2013). I think the answer here is that we used an N-terminus antibody. At that time, there were no commercial antibodies against TMEM106B that worked well for much, so we had raised antibodies to both C-terminus and N-terminus epitopes, but the N-terminus antibody much more clearly worked. It looks like these TMEM106B species are C-terminus. 

    The second thing I considered is whether the data from these papers agrees with "clues" we had from other lines of investigation. The exciting thing is that they do. First, many groups, including ours, have seen evidence for a dimer form of TMEM106B, as well as for glycosylation of TMEM106B, both of which are in line with the Fitzpatrick group paper's findings.

    It will be important to understand what the role of missense variants in TMEM106B, glycosylation of TMEM106B, dimerization of TMEM106B, is in driving the formation of fibrils. It seem very possible that increased expression of TMEM106B, which we have linked to specific FTD-associated genetic polymorphisms through a CTCF-mediated mechanism (Gallagher et al., 2017), may make fibrillization events more likely.

    Moreover, the fact that TMEM106B fibrils might be found in synucleinopathies and in FTD, but not in AD, accords with our clinical finding that TMEM106B genotypes associate with rate of decline in synucleinopathy and in FTD, but not in AD (Tropea et al., 2019).

    All in all: an exciting time to be a TMEM106B researcher!


    . Expression of TMEM106B, the frontotemporal lobar degeneration-associated protein, in normal and diseased human brain. Acta Neuropathol Commun. 2013 Jul 11;1(1):36. PubMed.

    . A Dementia-Associated Risk Variant near TMEM106B Alters Chromatin Architecture and Gene Expression. Am J Hum Genet. 2017 Oct 16; PubMed.

    . TMEM106B Effect on cognition in Parkinson disease and frontotemporal dementia. Ann Neurol. 2019 Jun;85(6):801-811. PubMed.

    View all comments by Alice Chen-Plotkin
  4. It’s surprising to see three very similar papers on TMEM106B appear at the same time.

    First impressions on these findings from the sarkosyl pellets from a diverse series of postmortem brain specimens raise questions about whether this is a new type of intraneuronal amyloid. The classical definition of an amyloid, i.e., Congo red negative birefringence, does not appear to have been met. Rather, we are presented with a neuronal cytoplasmic fibrillar aggregate that is (not always) associated with autofluorescent lipofuscin (Schweighauser et al.). These fibrillar aggregates are not specific for a particular disease, but are clearly age-related.

    All the fibrils were extracted using sarkosyl extraction protocols, albeit with some variation in the stage of sarkosyl addition: (I) in homogenate before the first centrifugation step (Jiang et al. and Schweighauser et al.) or (II) later, after low-speed centrifugation (Chang et al.). It was argued that the protocol change from II to I was essential for detecting abundant TMEM106B filaments.

    None of these studies observed fibrils of TDP-43, which have been identified before in ALS/FTD (Arseni et al., 2021) or even in FTD-TDP (Nonaka et al., 2013) using the preferred sarkosyl extraction protocol (I) (sarkosyl in homogenate). However, abundant non-filamentous aggregates of TDP-43 were identified by Jiang et al. in FTD-TDP extracts.

    Clearly, there’s still a lot of work to do to determine the biologic and pathologic significance of these TMEM106B fibrillar aggregates. If they are simply the end-stage detritus of age-related lysosomal processing, akin to lipofuscin, then it’s going to be difficult for them to get traction in the field. TDP-43 (whether or not in an aggregated fibrillar state) still remains of great interest, because of its FTD/ALS disease-association status.


    . Structure of pathological TDP-43 filaments from ALS with FTLD. Nature. 2022 Jan;601(7891):139-143. Epub 2021 Dec 8 PubMed.

    . Prion-like properties of assembled TDP-43. Curr Opin Neurobiol. 2020 Apr;61:23-28. Epub 2019 Dec 18 PubMed.

    View all comments by Colin Masters
  5. These represent breakthrough findings in the field.

    During the past decade, TMEM106B has been associated with brain aging, myelination disorder, and many neurodegenerative diseases. The demonstration by three independent studies of a TMEM106B C-terminal fragment as the amyloid fibril component raises many interesting questions to be answered.

    1) How much TMEM106B protein is cleaved during the disease state, and what is the effect on TMEM106B function?

    2) What triggers the cleavage of TMEM106B in the lysosomal lumen and aggregation of the C-terminal fragment?

    3) Are TMEM106B fibrils extracellular or inside of the lysosome?

    4) What are the effects of TMEM106B fibril formation on lysosomal function and aggregation of other proteins?

    Ultimately, these answers will help us understand the association of TMEM106B with brain aging and neurodegeneration.

    View all comments by Fenghua Hu
  6. These are fantastic papers illustrating the immense potential, but also potential caveats to be considered in the interpretation, of cryo-EM in neurodegenerative disease research.

    All three papers identified similar fibrils of C-terminal fragments of TMEM106b in the brains of TDP-43 proteinopathy cases. While Jiang et al. did not detect TMEM106 fibrils in their control samples, suggesting specificity for TDP-43 proteinopathies, Schweighauser et al. and Chang et al. identified TMEM106b fibrils also in a wide variety of other neurodegenerative conditions, as well in older neurologically healthy controls, thereby providing sufficient evidence that TMEM106b aggregation seems to be rather a common event in neurodegenerative conditions and upon aging.

    These findings raise crucial questions that will keep the field busy for a while, e.g.:

    (i) What are the neuropathological correlates of the TMEM106b aggregates in brain sections? Which cells develop aggregates? Association with lipofuscin? Association with inclusions composed of TDP-43, tau, synuclein in the distinct conditions?

    (ii) Mechanism of TMEM106b cleavage and aggregation?  

    (iii) Functional consequences of TMEM106b fibrils (loss of function/gain of toxic function/neither?).

    (iv) Link and role of TMEM106b fibril formation in the pathogenesis of the various neurodegenerative diseases (relevant/irrelevant?).

    Specifically in the context of TDP-43 proteinopathies, I think these studies clearly emphasize the future need to modulate/compare extraction methods and to carefully pay attention to the used sample preparation for cryo-EM analysis.

    While these three papers could not identify TDP-43 fibrils in their preparations, a previous study by Arseniy et al. did report the presence of TDP-43 filaments by using a slightly different extraction method, raising the possibility that TDP-43 filaments might have been lost in the preparation process in the three TMEM106b papers. In that context, it might be worth recalling that TDP-43 aggregates in the human TDP-43 proteinopathies are not stained with amyloid dyes such as Thioflavin-S (in contrast to inclusions composed of tau, synuclein, Aβ). Therefore it seems plausible that extraction protocols for detection of TDP-43 filaments might have to be adjusted.   

    I am very much looking further not only to the answers to the questions raised above on TMEM106β, but also to additional cyro-EM studies dissecting TDP-43 aggregates. 


    . Structure of pathological TDP-43 filaments from ALS with FTLD. Nature. 2022 Jan;601(7891):139-143. Epub 2021 Dec 8 PubMed.

    View all comments by Manuela Neumann

Make a Comment

To make a comment you must login or register.

This paper appears in the following:


  1. Surprise! TMEM106b Fibrils Found in Neurodegenerative Diseases