chapter 12 - ALS Flashcards
12.1. [als] describe the symptoms of ALS and explain how they arise
> adult-onset neurodegenerative disease
characterized by selective and progressive degeneration of motor
neurons
lower motor neuron degeneration
* muscle weakness
* muscle atrophy
* fasciculations
upper motor neuron degeneration
* spasticity
* slowness of movement
* incoordination
muscle atrophy= shrinkage of muscle, muscle wasting.
fasciculations= twitching = a brief spontaneous contraction affecting a small number
of muscle fibers, often causing a flicker of movement under the skin. spasticity= involuntary tightness of a muscle, some muscles are continuously contracted.
12.2. [als] summarize the genetics of ALS (A), and understand how ALS animal models can be generated based on familial ALS genes (B)
A
ALS can be sporadic (90%) or familial (10%)
Enzyme: SOD- superoxide dismutase: converts superoxide oxygen radicals into hydrogen peroxide and dioxygen
RNA binding and/or processing: FUS, TDP-43, TAF15, EWSR1, ANG, SETX, HNRNPA1, MATR3
Repeat expansions: C90RF72, ATX2
Proteostatic proteins:
UBQLN2, OPTN, SQSTM1, VCP, CHMP2B, FIG4
Excitotoxicity: DAO
Cytoskeleton/cellular transport:
VAPB, Peripherin, DCTN1, NFH, PFN1
Uncertain: Spastacin, Alsin
B
Mutant SOD1 transgenic mice recapitulae features of human ALS
- ubiquitous expression pattern
- show neurogenic atrophy
SOD1
SOD1 is ubiquitously expressed in all cells in the body. Yet, there is specific motor neuron (MN) degeneration. We don’t know yet why other things are not affected.
In this experiment, they generated a mouse model of ALS. They used a transgene to generate transgenic mice that either had the human WT SOD1 gene or with an ALS-causing mutation. Thus, these mice had, in addition to their own genome, a number of copies of the mutant or WT SOD1 gene. The lines in the picture on the right show age of symptom onset (squares) and life span (circles).
C and D: Section through the ventral horn of the spinal cord, with immunostaining for ChAT. This stains specifically for choline acetyl transferase, which synthesises ACh, and is thus specifically expressed in ACh neurons (i.e., MNs in the spinal cord ventral horn). C is a healthy mouse while D is an SOD1 mouse. There is loss in the SOD1 mouse of ChAT-positive ventral horn MNs.
H and J: Peripheral motor axons. Every circle is a myelinated axon, and many of those will be motor axons. In the SOD1 ALS mouse model (J), many axons are lost, while others are swollen.
The picture on the right shows muscle atrophy. The arrows indicate fibres that have atrophied and wasted away. All these findings are very reminiscent of human ALS patients.
Cell-autonomous and non-cell-autonomous mechanisms
Cell-autonomous mechanisms of SOD1 mean that MN degeneration is only due to toxicity in the MN itself, not other cell types. Non-cell-autonomous mechanisms means that the MN degeneration is not due to toxicity in the MN itself, but that other cell types are toxic to MNs. The fact that other cell types are not affected doesn’t mean that they couldn’t e.g., secrete molecules that are toxic to MNs and contribute to the MN degeneration.
To address this question of whether just the MNs themselves or other non-neuronal cells surrounding the MNs contribute to disease pathogenesis, they generated chimeric mice that have a population of WT cells and a population that carries the mutation. This can be done in two ways (that come down to the same thing):
Blastocyst cells (green) carry a mutant SOD1 gene (they tested a few different mutations). The inner cell mass is taken out of the blastocyst and WT cells that are YFP-labelled (so that you can recognise them later on) are added (yellow). Then let the chimeric blastocyst grow. You get chimeric animals with different percentages of chimerism.
Use morula (early-stage embryo consisting of 16 cells). Add WT cells (stained with Lac z, a blue staining) and SOD1 mutated cells (stained red). This results in chimeric mice.
They found that some WT MNs surrounded by mutant non-MN SOD1 cells still degenerated. And that some of the mutant SOD1 MNs that were surrounded by WT non-MN cells did degenerate, but delayed. Thus, the toxicity from mutant SOD1 acting in non-MN cells contributes to MN degeneration. ALS thus has both cell-autonomous and non-cell-autonomous aspects.
Another approach to study this is to create mice with a specific transgene: LoxP, mutated SOD1 gene, stop codon, LoxP, GFP. Without Cre, the gene is expressed, but then there is a stop codon, so the GFP is not expressed. When the mouse is crossed with a Cre mouse, the cells will start to express Cre, and Cre combines the two LoxP sites. Because the two sites are in the same orientation on the same DNA strand, the sequence in between them (in this case mutant SOD1) is removed. What’s left over is GFP, which will now be expressed.
They crossed the mice with this transgene with mice expressing Cre only in one cell type (motor neurons or astrocytes or oligodendrocytes). Then they evaluated what happens when SOD1 is expressed in all cell types except that one where Cre is expressed. (Note that the Cre transgenic mice will not remove the part between the LoxP stuff in all neurons where it should. It’s not 100% efficient). Results:
F-H: Mice with (red) or without (blue) Cre in the MNs. The age of onset of the disease (F) did not change much. However, the duration of the early disease (G) is longer in the Cre mice and the survival (H) is also longer. Thus, the toxicity of mutant SOD1 in MNs in particular contributes to the early disease stages in this mouse model.
J-L: Cre in the microglia. There is a substantial extension of the lifespan (J) without SOD1 in the microglia. There is no difference in the early disease stages (K), but the late disease stages (L) are substantially delayed. Thus, the activity of mutant SOD1 in microglia contributes to the late disease stages (i.e., it results in toxicity of MNs in the later disease stages).
f-g: Cre in astrocytes. The age of onset and early disease stage (f) are not changed. The late disease stages (g) are changed. Thus, the activity mutant SOD1 in astrocytes results in toxicity of MNs in the later stage of the disease.
a: Cre in oligodendrocytes. Already early in the disease, there’s a delay. When Cre is activated (?), the early stage and survival are delayed. Thus, toxicity of SOD1 in oligodendrocytes may contribute to the earlier stages of disease.
Thus, ALS is a disease of MNs but also of non-neuronal cells like astrocytes and microglia. From these experiments, it can be concluded that non-cell-autonomous toxicity to MNs induced by SOD1 mutant microglia, astrocytes, and oligodendrocytes contributes significantly to ALS pathogenesis. These experiments in transgenic mice were also confirmed in co-culture experiments where MNs were co-cultured either with astrocytes or microglia (where one is the mutant and the other the WT). Then you can really see what the contribution of toxicity within the MNs or within the microglia/astrocytes is to MN death (typically the phenotype that is investigated in these co-culture experiments).
However, questions still remain. One question is whether this is also true for other genetic forms of ALS (non SOD1-familial ALS) and whether it’s true for sporadic ALS. Another question is whether non-cell-autonomous contributions of skeletal muscles could have a role in ALS. This is interesting since skeletal muscles are ultimately the tissue type that’s affected secondary to MN degeneration and that ultimately give rise to the phenotype. So far, the potential contribution of skeletal muscle to ALS has been controversial. It has been investigated mainly in SOD1 mouse models since it’s clear that mutant SOD1 is intrinsically toxic to skeletal muscle and induces skeletal muscle pathology. However, whether this pathology contributes to the progressive loss of MNs is a bit more controversial and there have been quite some indications that this may not be the case.
12.3. [als] Paraphrase and design experimental approaches to evaluate whether non-cell-autonomous toxicity contributes to motor neuron degeneration in ALS
non-cell autonomous mechanisms of ALS
approach 1
- chimeric mice containing of mixture of wt and mutant sod1-expressing cells (mice contain genes of 2 genotypes)
- some wild type motor neurons surrounded by mutant sod1 nonneuronal cells degenerate
- wt nonneuronal cells surrounding mutant SOD1 motor neurons delay motor neuron degeneration
– leads to toxicity from mutant sod1 acting in nonneural cells that contributes to motor neuron degeneration
approach 2
- generation of transgenic mice carrying mutant SOD1 transgene flanked by loxP SITES (here lies the difference btw the two approaches)
- combined cre (in motor neurons, astrocytes) and loxP genes used to inactivate motor neurons
- microglia: resident immune cells in the cns which act as macrophages when activated. Microglia activation is a hallmark of ALS.
– leads to non-cell-autonomous toxicity to motor neurons induced by sod1 mutant microglia, astrocytes and oligodendrocytes
- this was confirmed by co-culture experiments
Now some studies that looked at the non-cell autonomous contribution of other cell types, in particular skeletal muscle, in a mouse model of ALS FUS.
FUS is a DNA/RNA-binding protein involved in several steps of gene expression regulation, including regulation of transcription, splicing, mRNA subcellular localisation, and even microRNA biogenesis.
FUS is one of the protagonists in a group of RNA-binding proteins that have all been linked to ALS, indicating that defects in RNA-biogenesis are likely involved in ALS pathogenesis. Other RNA-binding proteins linked to ALS are TDP-43, TAF15, EWSR1, ataxin 2, hnRNPA2B1, hnRNPA1, and matrin-3. Beyond ALS, cytoplasmic aggregates of proteins including FUS and TDP-43 are also a pathological hallmark found in frontotemporal dementia (FTD), the second most common form of dementia after Alzheimer’s disease. They are found in 10% (FUS) and 45% (TDP-43) of FTD patients.
The picture shows a schematic overview of the functional domains of the FUS protein:
Low-complexity (LC) domain: Located at the N-terminus and thought be involved in the formation of RNA granules by interacting with other low-complexity domains (either in other FUS proteins or other RNA-binding proteins that contain this domain, like TDP-43).
RGG motif rich domain: Rich in RGG. There are three in FUS.
RNA recognition motif (RRM): Classic RNA-binding domain.
Zinc finger motif (ZnF): Involved in DNA- and likely also RNA-binding. It might be important for the binding of FUS to DNA when it’s involved in gene expression regulation.
Nuclear export signal (E) and nuclear localisation signal (L): L is located at the C-terminus. There’s a nuclear localisation signal because FUS is a protein that’s predominantly localised to the nucleus under physiological conditions. However, it can shuttle between the nucleus and cytoplasm.
Pathogenetic ALS-causing mutations that have been reported in the FUS genes are also indicated. They are colour coded: yellow = found in a single patient; green = found in more than one patient; red = segregates with disease in a family (i.e., in a family with multiple cases of ALS, every person with ALS has the mutation while no-one without ALS has the mutation). The segregation within a family is the strongest evidence for disease causality.
There’s a clustering of mutations within the nuclear localisation signal or right before it. Within the nuclear localisation signal there are often missense mutations (one AA is changed for another one), meaning that the nuclear localisation is mutated. Before the nuclear localisations signal, there are often frame shift mutations, meaning that the nuclear localisation is completely missing. The consequence of either is that FUS is not properly functioning and mis-localises to the cytoplasm.
Under physiological conditions, FUS is mostly located in the nucleus. However, ALS-associated mutations in FUS result in FUS aggregation in the cytoplasm. This means that at the same time, there’s a nuclear clearance of FUS (there are reduced levels of FUS in the nucleus, and sometimes even a complete lack of FUS).
12.4. [als] explain how loss-of-function changes can be distinguished from gain-of-toxic-function changes in ALS animal models
LOS VS. GOTF
Originally, it was unclear whether ALS-FUS was caused by nuclear loss-of-function (LOF) or cytoplasmic gain-of-toxic-function (GOTF)
* LOF: reduction (or even complete loss) of FUS in the nucleus -> endogenous functions of FUS are no longer executed -> LOF.
GOTF: cytoplasmic mis-localisation and potential aggregation of mutant FUS protein -> GOTF Because of the disease-causing mutation, the protein has gained a new interaction with other proteins that can be toxic.
To study GOTF vs. LOF, they generated
Fus knock in mice for Eus. The whole Fus gene is shown at the top. It has 15
exons (dark blue are coding regions, light blue are non-coding untranslated regions). In intron 12, a cassette is introduced that consists of exon 13 and 14 and a stop codon, which are flanked by Loxl sites. Thus, (without Cre) a truncated version of the protein is created that doesn’t contain the last 20 amino acids from exon 15. Those amino acids contain the nuclear localisation signal. Thus, the truncated protein is expected to localise to the cytoplasm and have reduced
levels in the nucleus.
This is more or less what they found. The pictures show spinal cord ventral horn immunostained for ChAt, FUS protein, and a nuclear stain.
* In control animals (top row): FUS is expressed in the nuclei of motor neurons and non-motor neurons.
Mouse with two KI alleles and no WT FUS: FUS is mostly localised to the cytoplasm and nuclear levels are strongly reduced. It doesn’t form cytoplasmic aggregates however.
They also generated a KO mouse that has a LOF of FUS protein. Then, by comparing phenotypes of the KI and KO mice, you can say which are caused by LOF (the phenotypes that are shared between KI and KO) and which by TGOF (phenotypes only found in KI).
Most phenotypes are shared in KI and KO (and thus due to LOF): they both dig soon after birth because of respiratory defects, the mice are blueish, the lungs are deflated (see picture), and the lungs don’t float. There are less MN cell bodies in the ventral horn of KI mice than controls, but not in the KO mice. So this is most likely due to a gain of toxic function mechanism.
Then they crossed the KI mice with mice expressing Cre only in the motor neurons. Thus, the WT protein is expressed in the MNs in those mice. This way, they could check whether the reduction in MNs is due to toxicity in the MNs purely, since the MN loss should be rescued. If it’s due to other cell types, it will not be completely rescued. They found that the number of MNs goes back to WT. Thus, the loss of MNs in KI mice is due to intrinsic toxicity of the mutant FUS within the MNs.
12.5. [als] design experimental approaches to investigate the contribution of skeletal muscle pathology in ALS mouse models
Heterozygous Fus mice as model for ALS-FUS
Since ALS patients are heterozygous for FUS mutations, they wanted to find out whether heterozygous mice could serve as a good mouse model for ALS-FUS, or at least capture some of the features.
The mice could recapitulate some features:
Ubiquitin pathology: A typical finding in ALS is aggregations of ubiquitinated proteins in MN cell bodies. The stains against ubiquitinated proteins show aggregates in the MN cell bodies in the heterozygous mice, but not the controls.
Progressive MN loss: At 10mo, there’s no difference in the number of motor neurons in the ventral horn between the heterozygous and control mice. At 22mo, there’s a reduction of 25-30%. Thus, between 10mo and 22mo, there’s progressive degeneration. Both the Nissl staining (histochemical staining to stain all cells in the ventral horn: the cells that are larger than a certain size are classified as MNs) and the ChAT staining (stains specifically ACh neurons, i.e., MNs) both show this result. They used both stainings here since FUS is a protein that’s involved in gene expression regulation and could regulate ChAT, which would mean that it is regulated differently in those mice. Then you’d think there are less MNs when in reality the marker is just expressed less. Thus Nissl excludes that there’s an effect of FUS on ChAT. And the other way around, ChAT is also used because Nissl is a bit arbitrary (have to classify them based on size).
Motor and EMG deficits: They are found from 10mo onwards, but only rather mild. Nothing leading to complete paralysis.
They also tested heterozygous FUS KO mice, which don’t develop any of these phenotypes. Thus, gain of function toxicity is required to get the phenotypes that are akin to ALS symptoms. This doesn’t necessarily mean that LOF can’t contribute, but it’s not sufficient.
Then the mice were crossed with ChAT-CRE mice (i.e., mice that expressed Cre in the MNs), meaning that WT FUS is expressed in the MNs again.
The MN loss at 22mo is fully rescued, indicating that MN loss itself is due to intrinsic toxicity of mutant FUS in the MNs.
The onset of the motor deficits was later, but tthey still occurred, despite the fact that the MN loss doesn’t occur. Thus, cell types other than MNs contribute to those motor deficits.
They thought it could be due to mutant FUS expression in the skeletal muscle, so they looked at the NMJ to test this. Green indicates the nerve terminals, labelled with an antibody for neurofilaments. The presynaptic site is labelled with some SV protein thing. Red indicates the postsynaptic sites (the end plates). In controls, the postsynaptic site is completely covered by the presynaptic site (i.e., the endplate is fully innervated). In the heterozygous KI mice, even in very young mice (1mo), there are some end plates that are not (fully) innervated by a nerve terminal. They are partially or fully denervated. This is also found in ALS patients. They also quantified the surface area of the end plate. This is slightly reduced in the KI mice, suggesting that there’s a problem with the expression of AChRs.
The pictures on the right show NMJ phenotypes in homozygous KO (-/-) and KI (ΔNLS/ΔNLS) mice compared to controls. The endplates are again labelled in red and the axons and presynaptic nerve terminals in green. The end plate size is very variable, even within the same animal (also in WT controls).
Left two graphs: The surface area of the end plates was reduced in KI mice but not KO mice, suggesting that it may be mediated by toxic GOF.
Middle two graphs: The innervation status (endplate occupancy), i.e., whether the endplate overlaps with the presynaptic nerve terminal, was 100% in all three groups (WT, KI, and KO mice).
Right two graphs: The total number of endplates in a muscle (the tibialis anterior) is reduced in both KO and KI mice (right two graphs). This may thus be due to a LOF.
Then they wanted to see whether the reduced end plate size could be due to FUS toxicity in the muscle. FOr this, they used MyoD-CRE mice, which selectively express Cre in the skeletal muscles (thus, these mice have WT FUS in the skeletal muscles).
Subcellular localisation of FUS (with immunostaining for FUS (green), phalloidin (a staining for muscle fibres; red), and DAPI (staining for nuclei; blue)): The FUS protein is localised to the nuclei of the muscle fibres and non-muscle nuclei. In homozygous KI mice (without Cre) (second row), FUS is localised to the cytoplasm. There is no longer nuclear enrichment, but it’s more or less equally distributed throughout the cell. In homozygous KI mice with MyoD-Cre, FUS protein is distributed similarly to the control animals again: it’s back in the nucleus. Thus, the MyoD-Cre reversed the phenotype back to WT.
Endplate area size: The reduced endplate size is rescued with Cre. Thus, the reduction in endplate area is due to intrinsic toxicity of mutant FUS protein specifically in the skeletal muscles.
These results suggest that maybe the mutant FUS protein reduces the expression of AChR genes (since that means there are less AChRs and thus smaller endplate areas).
They tested this theory. Remember that the expression of AChR subunit genes is stimulated by Agrin-LRP4-MuSK signalling (see previous lecture).
FUS levels in skeletal muscle. In WT mice, the subsynaptic nuclei (i.e., those underneath an endplate) have higher levels of FUS (about 2-3fold higher). This enrichment is lost in heterozygous KI mice.
Expression levels of AChR subunit genes: Are reduced in the muscles of both homo- and heterozygous KI mice.
They showed (with ChIP experiments) that FUS binds to the promotor regions of the AChR subunit genes and stimulates the transcription of those genes.
FUS and ERM are interdependent: if you knock out FUS or ERM, you get a reduction in expression of the AChR subunit genes. If you overexpress FUS, you get increased expression of AChR subunit genes, but if you then knock out ERM, this increase is suppressed. And vice versa: overexpression of ERM increases AChR subunit gene expression, and knocking out FUS suppresses this increase.
FUS and ERM also physically interact.
They came to the following model: Compared to the model from the previous lecture, FUS is added. It’s enriched in subsynaptic nuclei and physically interacts with ERM and stimulates the transcription of the subunit genes.
The mutations that cause a loss of the last exon (which are the majority, but not all) cause a specific LOF effect. In a control animal, FUS is expressed in all nuclei of the myofibres, but it’s enriched in the subsynaptic nuclei. This results in the proper transcription and expression of AChR subunit genes, normal endplate size, and neuromuscular transmission. In the heterozygous KI mice, the enrichment in the subsynaptic nuclei is lost, which leads to reduced expression of the subunit genes, smaller endplate sizes, and defective neuromuscular transmission.
