TRIM11 protects against tauopathies and is down-regulated in Alzheimer's disease

Research tools used in this peer-reviewed study.

AUMsilence™ sdASO™ (self-delivering antisense oligonucleotide)

Experimental methodology

Primary cortical neurons were prepared from wild-type or PS19 mouse pups, PS19 expressing human tau P301S, and seeded on plates on day 1. AUMsilence sdASO targeting TRIM11, or a scrambled control oligonucleotide, was added straight into the medium at 10 μM on day 4, with a second dose on day 14 for long-term knockdown. Knockdown itself was read by Western blot three days after a dose. Preformed tau fibrils assembled from myc-K18/P301L, a truncated tau carrying the P301L mutation, were added to the PS19 neurons at 1.5 μg per well on day 7, seeding aggregation of the neurons' own tau. Neurons were analyzed on day 21 using immunofluorescence with AT8 antibody (recognizing tau phosphorylated at Ser202 and Thr205) and MC1 antibody (recognizing a disease-specific conformation of tau). AT8 and MC1 signals were normalized on the basis of cell numbers.

Results achieved

AUMsilence sdASO entered the neurons and silenced TRIM11 to different extents across the five sequences tested, read on a Western blot against an HSP90 loading control three days after dosing (supplementary figure S12B). That blot is not quantified, so the study reports the silencing as graded rather than as a percentage, and no knockdown percentage appears anywhere in the paper or its supplement. In PS19 cortical neurons challenged with preformed tau fibrils, silencing TRIM11 exacerbated tau aggregation, increasing AT8- and MC1-reactive tau by approximately 50 to 90%. That is one range covering both markers together, and the paper does not divide it between them. Figure 5C and supplementary figure S12D each mark the fibril-treated comparison at P < 0.05, mean ± SEM, n = 3. A scrambled control oligonucleotide was the comparator throughout, and the specificity argument the authors draw is a graded one: silencing TRIM11 with the various oligonucleotides reduced the viability of cortical neurons in a manner that was correlated with the effect of each one on TRIM11 expression. The deeper the knockdown, the worse the viability. This is the loss-of-function arm of the study, and its result is that lowering TRIM11 makes seeded tau pathology worse in primary neurons.

Page reference

Page 7-8, Figures 5B-C, Supplementary Figure S12

• View product details

What the study's figures show, and which arm of it used the AUM product.

Figure 2: TRIM11 protein is reduced in sporadic Alzheimer's disease brains

Immunoblot of postmortem frontal cortex gray matter from 14 control and 23 sporadic Alzheimer's disease individuals, age- and sex-matched (Figure 2A-D). Every Alzheimer's subject in table S2 is Braak stage V or VI and every control is Braak I or II. TRIM11 protein was present at a substantially lower level in Alzheimer's than in control tissue, an approximately 55% reduction on average, while TRIM10 and TRIM55 protein were comparable between the two groups, both marked not significant on Figure 2C. All three transcripts were present at similar levels in the two groups, so the authors conclude that the change in TRIM11 expression is probably caused by a post-transcriptional mechanism, and they read the unmatched messenger RNA and protein levels in individual Alzheimer's samples as suggesting that the TRIM11 reduction preceded neuronal loss. The protein reduction was corroborated by immunohistochemistry, which put TRIM11 lower and AT8 higher in Alzheimer's tissue at P < 0.05 (Figure 2E-F), and by immunofluorescence controlled for the number of neurons (Figure 2G-H), both at mean ± SD, n = 4. Across control and Alzheimer's tissues there was a strong inverse correlation between TRIM11 expression and the level of each of four abnormally phosphorylated tau species. Figure 2I prints a Pearson coefficient on each panel: AT8 r = -0.3504, p = 0.0335; pSer262 r = -0.3961, p = 0.0152; AT180 r = -0.3809, p = 0.0200; PHF-1 r = -0.4201, p = 0.0096. The correlation held among the Alzheimer's tissues alone, where Figure 2J prints pSer262 r = -0.4277, p = 0.0418 and PHF-1 r = -0.4369, p = 0.0371, with AT8 and AT180 in supplementary figure S4I.

Figure 5B-C and supplementary figure S12C-D: silencing TRIM11 with AUMsilence™ sdASO™ worsens seeded tau aggregation in PS19 neurons

Five antisense oligonucleotides against Trim11, and a scrambled control, all designed and synthesized by AUM BioTech, were put straight into the medium of primary cortical neurons at 10 μM. A far-red labeled scrambled control was used to watch uptake (supplementary figure S12A), and knockdown was read by Western blot three days later. Supplementary figure S12B is that blot: one lane for the scrambled control and one for each of the five sequences, TRIM11 against an HSP90 loading control. It is not quantified, and the study reports the result as graded, that the oligonucleotides entered the neurons and silenced TRIM11 to different extents, rather than as a percentage. PS19 cortical neurons were then challenged with preformed tau fibrils assembled from myc-K18/P301L, which produced neuritic thread-like inclusions reactive to AT8 and to MC1. Silencing TRIM11 exacerbated that aggregation, increasing AT8- and MC1-reactive tau by approximately 50 to 90%, one range covering both markers together. Relative AT8 intensity is Figure 5C and relative MC1 intensity is supplementary figure S12D, and in both the marked comparison is between the control and the TRIM11 oligonucleotide in fibril-treated neurons, at P < 0.05, mean ± SEM, n = 3.

Figure 5F-M: silencing TRIM11 also costs synapses, neurofilament and viability

In wild-type cortical neurons the same oligonucleotides reduced synaptophysin-positive puncta by approximately 40% (Figure 5F-G) and PSD95-positive puncta by approximately 30% (Figure 5H-I), and reduced their juxtaposition as well (supplementary figure S12J-K). TRIM11-depleted neurons contained lower levels of neurofilament light chain, an approximately 40% reduction, while microtubule-associated protein 2 was unchanged, its dendrite-length panel marked not significant (Figure 5J-L). Viability fell too, and it fell in a manner correlated with the effect each oligonucleotide had on TRIM11 expression: Figure 5M sets the control oligonucleotide against three of the sequences and marks two of the three significant and the third not. That panel is n = 6; every other panel in Figure 5 is n = 3, mean ± SEM. Synaptophysin and PSD95 signals were normalized to dendrite length, neurofilament light chain to cell number.

Figure 6: an adeno-associated virus carrying TRIM11, not an oligonucleotide, reduces tau pathology and improves cognitive and motor performance in PS19 mice

No animal in this study received an oligonucleotide. AAV9-TRIM11 or AAV9-GFP was delivered to the hippocampus of PS19 mice of either sex by bilateral stereotaxic injection at 2.5 months of age, 2×10^10 genome copies, and pathology and behavior were analyzed at approximately 10 months (Figure 6A). Hippocampi of the TRIM11 group contained approximately 55% less tau pathology by AT8 immunohistochemistry (Figure 6B, mean ± SD, n = 6 mice), and immunoblot showed reductions in both soluble and insoluble phosphorylated tau species (Figure 6C), quantified in supplementary figure S13, B and C, at n = 3 mice: against HSP90 the falls are approximately 81% and 52% for AT8 in the insoluble and soluble fractions and approximately 98% and 89% for PHF-1, and against total tau in each fraction approximately 59% and 56% for AT8 and approximately 83% and 92% for PHF-1. The effect was dose dependent, hippocampi expressing more TRIM11 containing fewer phosphorylated tau species. GFAP immunoreactivity fell by approximately 50% and microgliosis by approximately 40%, each, the authors write, to the level seen in age-matched wild-type littermates (Figure 6D-G). The two panels are marked differently and the difference is worth stating: Figure 6E puts the untreated PS19 group significantly above wild-type for GFAP and the treated group indistinguishable from it, while Figure 6G marks the untreated PS19 group as not significantly different from wild-type for Iba1, so the only significant comparison there is treated against untreated. Immunoreactivity rose for MAP2 by approximately 50%, for neurofilament light chain by approximately 70% and for NeuN by approximately 40% (Figure 6H-M). The paper puts the neurofilament light chain figure at a level comparable to wild-type, and Figure 6I still marks MAP2 in the treated group significantly below wild-type. Panels D to M are mean ± SD, n = 3 to 7 mice. In an object recognition test the GFP group showed no difference between the familiar and the novel object, approximately 50% preference to each, while the TRIM11 group showed an approximately 77% preference for the novel object, similar to wild-type mice (Figure 6N). In a wire hang test the TRIM11 group held on for approximately 75 seconds against approximately 50 seconds for the GFP group, still short of wild-type (Figure 6O). Both behavioral panels are mean ± SD, n = 4 to 7 mice.

1. 01Zi-Yang Zhang, Dilshan S. Harischandra, Ruifang Wang, Shivani Ghaisas, Janet Y. Zhao, Thomas P. McMonagle, Guixin Zhu, Kenzo D. Lacuarta, Jianing Song, John Q. Trojanowski, Hong Xu, Virginia M.-Y. Lee, Xiaolu Yang. (2023). TRIM11 protects against tauopathies and is down-regulated in Alzheimer's disease. Science. DOI: 10.1126/science.add6696