WP04: Biomarkers and modifers of CLN3

WP Leader

Thomas Braulke, PhD    
University Medical Center Hamburg-Eppendorf (UKE)

WP Summary

 

 

Objectives


To provide novel biomarkers of CLN3 disease by analysing using genetic high-throughput screening and genome-wide microarray approaches

 

  • a simple well-studied and genetically tractable fission yeast model for CLN3 disease  
  • homozygous mouse neuronal precursor Cln3Δex7/8 cell lines
  • CLN3-depleted cells

To identify and exploit

  • conserved pathways that ameliorate the effect of mutations in btn1 (the homologous yeast gene to CLN3)
  • dysregulated potential biomarker genes in Cln3Δex7/8 neurons
  • modifier genes differently expressed during acute down-regulation of CLN3 modelling rapid course alterations


 

Yeast model for CLN3 disease


Partner 3 has shown that a simple fission yeast model for CLN3 disease can be used to predict the severity of disease caused by different mutations in CLN3, consistent with identified metabolic changes and diverse phenotypic effects.  The basis for these effects may be alterations that physically distort the Golgi apparatus and affect many downstream pathways. The ease of genetic manipulation in this organism makes this model unique in terms of the potential to identify novel pathways that ameliorate the effects of mutations in CLN3.

Integrated yeast strains carrying mutations equivalent to the common juvenile CLN3 disease 1 kb (∆ex7/8) deletion or p.Glu295Lys, a mutation associated with a particularly slow disease progression have been successfully created.

The strains have been characterized in terms of marker phenotypes exhibited by a strain completely deleted for btn1.

In order to use these strains in any screens to identify genes that rescue btn1 mutant phenotype(s) or act in the same pathway as btn1 mutant proteins, conditions needed to be identified that robustly differentiate between the growth of wild-type yeast, btn1del, and the mutant strains.  This had been successfully achieved.

 

Identification of genes rescuing btn1 mutant phenotypes in yeast

 

Towards a high-thoughput screen, conditions were identified that robustly differentiate between the growth of wild-type yeast and btn1del strains. A screen of over 1000 FDA-approved small molecules was performed. Three molecules showed efficacy in the yeast disease model, in mammalian cells in vitro, and in a zebrafish morpholino model of CLN3 disease.

A novel transposon mutagenesis approach has been established as a sophisticated method that also allows more comprehensively to identify the genetic interactions of btn1. The method allowed mutagenesis of a yeast culture by making insertions of a modified transposon at a random position in the genome, generally at a frequency of one insertion per cell. This tool is sufficiently well-developed that it is routine to create a library with an estimated 109 insertion-mutated cells. This method has been applied to investigate the genetic interactions of btn1.

Six libraries in wild type and cells lacking btn1, have been generated, sequenced, with >23 million good quality reads obtained at an average of 170 reads/kb.  Bioinformatic analysis, which is complex, is still being optimized, but it is clear that further optimisation of the method is possible.

Synthetic genetic arrays (SGAs) are a well established tool that can be used to systematically examine the consequences of loss of every gene in a query disease strain, highlighting genetic interactors of the disease gene. Those that enhance fitness are considered positive interactors (genetic suppressors) and those that decrease fitness are negative interactors (synthetically lethal or sick). Positive interactors can provide novel therapeutic candidates, and overall SGAs reveal where the query gene fits within the network of cellular pathways. We performed SGAs in four independent btn1D isolates from two different backgounds that highlighted 288 positive and 119 negative interactors (Figure 3).

To conclude, genes and small molecules that ameliorate CLN3 disease in the yeast model that can be followed up translationally have been successfully identified. A manuscript is in preparation:

Bond M et al. Genome-wide analysis reveals a central role for Tor signalling in a yeast model for neurodegeneration.

 


Fig. 3. The genetic interactions of btn1. SGAs were performed in two genetic backgrounds lacking btn1. This yielded 119 negative interactions (indicated in red), enriched for stress response, and 288 positive interactions (indicated in green).


Comparative gene expression analyses of 1-kb deletion models


Gene expression analysis in CbCln3Δex7/8 cerebellar precursor cells


The first milestone of this project part was a genome-wide microarray analysis of mouse cerebellar wild-type and Cln3Δex7/8 cell lines derived from a knock-in mouse model containing the most common CLN3 mutation (1-kb deletion of exons 7 and 8) which is found in 85% of all CLN3 patients. In the time between application and start of the Dem-CHILD project, Dr. Cotman (Boston, MA), who had provided the project with these cell lines, already published a gene expression analysis of Cln3Δex7/8 cells (Cao et al. PloSone 2011), and found 981 dysregulated genes belonging to gene sets related to metabolic processes (fatty acid and amino acid metabolism and iron transport), protein transport or mitochondrial membranes. Therefore, it was decided to extent these studies by analyzing the expression of microRNAs in the same mouse cerebellar wild-type and Cln3Δex7/8 cell lines.

MicroRNAs are proposed as master regulators of gene expression in important cellular pathways. By analyzing their expression profile a deeper insight into the gene-regulating network in Cln3-defective brain supporting our studies on biomarker and modifier genes could be gained. To analyse microRNA data sets and to compare them with the already existing gene expression data sets from mouse models and patients, the advanced bioinformatic Ingenuity Software has been used. A concurrent license for this software had been purchased by the DEM-CHILD coordinator allowing its usage in a collaborative way by partners 1, 2 3, and 6 for WP04 as well as for WP 05.

A total of 21 up- and 8 down-regulated microRNAs have been identified in Cln3Δex7/8 cell lines. Up- and down-regulated microRNAs are supposed to act as suppressors and activators of the targeted mRNAs, respectively. From 19 of the 29 candidates, 663 mRNAs represent high potential transcripts estimated by filtering for metabolic processes, signaling, transport, and endocytosis. Among them, 117 mRNAs are experimentally demonstrated. For the remaining microRNAs no targets are known. Additionally, out of the 19 microRNAs we selected three microRNAs for further studies. In two of them, we could confirm the dysregulation by RT-PCR.  

Also, the microRNA data were combined with the gene expression analysis data of the same CbCln3Δex7/8 cell line published by Cao et al. Approximately 500 dysregulated genes were found to be clustered and these included genes that may regulate potential biomarkers of CLN3 disease.

Secondly, an analysis of the total proteome of the cln3 ki cells was performed. For this purpose an in vivo protein labeling approach (SILAC; stable isotope labeling by amino acids in cell culture) followed by quantitative mass spectrometry (MS) was used. Approximately 200 dysregulated proteins were identified. Among these, candidates which correspond to potential targets of the previously identified microRNAs were found.

Taken together, this is the first comprehensive analysis of the lysosomal CLN3 proteome which identifies subsets of lysosomal enzymes impairing lysosomal homeostasis and lysosomal targeting routes that might represent novel targets for therapeutic approaches.

These data have been presented by Dr. G. Makrypidi at the 19th European Study Group on Lysosomal Disorders (ESGLD), Sept. 2013, in Graz, Austria, and at the Working Group for Paediatric Metabolic Society (APS) March 2014, in Fulda, Germany and a manuscript is being prepared:

Schmidtke C, Markrypidi G, Pohl S, Thelen M, Schweizer M, Jabs S, Storch S, Cotman SL, Gieselmann G, Braulke T, Schulz A: Cln3Dex7/8 impairs endocytic receptor trafficking and efflux of lysosomal amino acids.

 

 

Evaluation of possible modifier genes on protein/mRNA level and in patient tissue

 

In parallel to the above described microRNA and pr0teome analyses, the gene expression data from mouse cerebellar Cln3Δex7/8 cells were compared with genome-wide expression analyses in lymphocytes (PBMCs) of eight CLN3 patients homozygous for the 1 kb deletion who manifest different phenotypes characterized by fast, average and slow progression of the disease, and of six age- and gender-matched control individuals (Lebrun et al. Mol. Med 2011). In the latter study five genes were identified which were dysregulated in all CLN3 patients and present candidate biomarkers of the disease (Table 4).

RT-PCR analysis revealed that only DUSP2 (dual specificity phosphatase 2) is upregulated in all PBMCs collected form CLN3 patients of all phenotypes as suggested by the microarray data.

Further studies on DUSP2 have shown that the tissue and age-specific expression of DUSP2 appears to be a promising target to modulate both the onset and severity of CLN3 disease. A manuscript with detailed description of these studies is being prepared:

Makrypidi G, Schulz A, Ballabio A, Braulke T: TFEB  control  through ERK signaling by Dual Specificity Phosphatase 2.

 

Table 4: Biomarker candidates dysregulated in CLN3 patients (Lebrun et al. 2011)

 

 
   

 


Comparative gene expression analyses in a model of acute down-regulation of CLN3


A genome-wide microarray analysis of acutely CLN3 down-regulated cells was performed. Acute down-regulation of CLN3 by siRNA treatment in Hela cells represents a cell model for molecular alterations in CLN3 patients with severe loss of CLN3 function or rapid progression of juvenile CLN3 disease (Lebrun et al. Mol Med 2011). Thus, this cell model was used for the analysis of modifier genes in CLN3 disease. Treatment of HeLa cells for 96 hours with three different siRNAs resulted in a down-regulation of CLN3 mRNA by 92 % in comparison with scrambled siRNA treated cells. The microarray analysis was performed in collaboration with SME partner 12. We identified in total 49 dys-regulated genes from which 7 were up-regulated and 42 were down-regulated. None of these genes were found within the gene data set derived from genome-wide microarray analysis derived from lymphocytes of CLN3 patients with rapid progression of the disease.


Conclusion


The work in WP4 has contributed to the identification of biomarker and modifier of CLN3 disease by

(1)    identifying a novel signalling pathway and potential therapeutic targets, together with three small molecules in the yeast model

(2)    comparative mRNA, microRNA arrays and SILAC protein analyses of cerebellar Cln3Δex7/8 percursor cells providing new insights in the dysregulated gene network in CLN3 disease.

(3)    evaluating the dual specificity phosphatase 2 (DUSP2) as  a highly potential biomarker of CLN3 disease