pancreas_lab_meeting

1. Cell intrinsic cause for type one diabetes

Group meeting by Roy Novoselsky
24.11.2022

2. Collaboration with Yuval Dors lab from HUJI

Prof. Yuval Dor
Dr. Agnes Klochender
Shani Peleg

3. The project is about…

Alu
elements
RNA
editing
Interferon
response
Type 1
diabetes

4. Introduction – autoimmune type 1 diabetes

• Caused by depletion of beta cells in islets of Langerhans (Insulitis)
• High abundance (>1:1000 in population)
Alu
elements
RNA
editing
Interferon
response
Type 1
diabetes
• Mechanism differs from T1D that follows T2D
• Treatment is by Insulin injection and monitoring of blood glucose level, but no cure
• Definitive cause is still unknown!
• Leading theory – viral infection causes diabetes:
o Autoimmune surveillance of hypersecreting mutants
o Molecular mimicry
o Inflammation
Marroque et al. 2021

5. Introduction – autoimmune type 1 diabetes

Alu
elements
RNA
editing
Interferon
response
Type 1
diabetes
Marroque et al. 2021

6. Autoimmune surveillance of hypersecreting mutants

Alu elements
• Short interspersed nuclear elements (SINEs)
• Retrotransposons (“hitchhikers” of LINEs)
• Found only in primates
• Length: ~ 300 bp
• > 1,000,000 in human genome
• > 10% of the human genome
• Play a huge role in evolution (contain 2 splice sites)
universität zu köln, Research highlights, Alu elements
Retro transposition (LINE)
Alu
elements
RNA
editing
Interferon
response
Type 1
diabetes

7. Alu elements

Adar proteins perform (A to I) RNA editing
• Editing performed by Adar enzymes on dsRNA
• >12,000 editing sites in >1,600 human genes
• Functions:
o Defense against dsRNA viruses
o Editing of splicing sites
o Regulation of miRNA
o Editing of protein-coding regions
o Creates intracellular and extracellular heterogeneity
o Editing of Alu elements (vast majority of editing)
Alu
elements
RNA
editing
Interferon
response
Type 1
diabetes

8. Adar proteins perform (A to I) RNA editing

dsRNA causes interferon response and expression of ISGs
ALU
IFNR
IFN
(Ifih1)
MAVS
Cytoplasm
Nucleus
NFkB
IRFs
IFN + ISGs
(Interferon Stimulated Genes)
ISGs
Alu
elements
RNA
editing
Interferon
response
Type 1
diabetes

9. dsRNA causes interferon response and expression of ISGs

Hypothesis
Viral infection
Lower RNA editing
dsRNA
Activation of IFIH1 (MDA5)
IFN signaling
T1D
Lower levels of RNA editing
might be a cause for type 1 diabetes
Alu
elements
RNA
editing
Interferon
response
Type 1
diabetes

10. Hypothesis

Experiment model and design
Tamoxifen
Ins
Lox
Rosa26
Cre
RNA sequencing
Adar1
Lox
Lox
Stop
Lox
YFP
Imaging

11. Experiment model and design

Knockout of Adar1 in beta cells causes mice to become diabetic

12. Knockout of Adar1 in beta cells causes mice to become diabetic

Massive inflammation in beta-AdarKO
*12 days post Tamoxifen injection

13. Massive inflammation in beta-AdarKO

Innate cells & Lymphocytes are present at beta-Adar -/- islet infiltrates
Beta-Adar +/-
Beta-Adar -/-

14. Innate cells & Lymphocytes are present at beta-Adar -/- islet infiltrates

Plan of action – check interferon response genes in conditional Adar
knockout via smFISH
Why smFISH:
• Could detect very subtle changes.
• Easy to analyze.
• Lack of antibodies for ISGs.
WT
smFISH for ISGs
B - KO
*7 days post Tamoxifen injection
Analysis

15. Plan of action – check interferon response genes in conditional Adar knockout via smFISH

Log10(mean expression in beta-ko)
Interferon target genes are upregulated in Adar knockout
*7 days post Tamoxifen injection
Log10(mean expression in wt)

16. Interferon target genes are upregulated in Adar knockout

Selected Interferon target genes for smFISH validation
Pdl1
WT
B-KO
Ifih1
WT
B-KO
B2m
WT
B-KO

17. Selected Interferon target genes for smFISH validation

IFN target genes are expressed also in non-endocrine pancreatic cell
types
Pdl1
B2m
Ifih1

18. IFN target genes are expressed also in non-endocrine pancreatic cell types

No clear overexpression of interferon target genes
eGFP
B2m
Gene
Ins
Ifih1
Gcg
DNA
Pdl1

19. No clear overexpression of interferon target genes

Quantification of smFISH images using ImageM
eGFP
ISG
Ins2
DNA

20. Quantification of smFISH images using ImageM

eGFP
ISG
Ins2

21. Quantification of smFISH images using ImageM

Dots per volume (1/um^3)
smFISH analysis reveals no significant upregulation of interferon target
genes

22. smFISH analysis reveals no significant upregulation of interferon target genes

No correlation between target genes and eGFP

23. No correlation between target genes and eGFP

Do we see the same effect in alpha-KO?
Lox
Rosa26
Beta-KO
Ins
Cre
Adar1
Lox
Lox
Stop
Lox
Alpha-KO
Gcg
Cre
YFP
Double-KO
Ins
Cre
Gcg
Cre

24. Do we see the same effect in alpha-KO?

No inflammation in alpha-AdarKO
*12 days post Tamoxifen injection

25. No inflammation in alpha-AdarKO

smFISH analysis reveals no significant upregulation of interferon target
genes

26. smFISH analysis reveals no significant upregulation of interferon target genes

A subset of islets show potential immune infiltration and highly elevated
levels of B2m, but most transcripts are in non-parenchymal cells
eGFP
B2m
Ins2
DNA

27. A subset of islets show potential immune infiltration and highly elevated levels of B2m, but most transcripts are in

Extreme beta cells are present in islets of Langerhans
~8% of beta cells in
healthy islets are
extreme beta cells

28. Extreme beta cells are present in islets of Langerhans

A subset of islets show potential immune infiltration and highly elevated
levels of PDL1, including in beta cells
eGFP
Pdl1
Ins2
DNA

29. A subset of islets show potential immune infiltration and highly elevated levels of PDL1, including in beta cells

A subset of islets show potential immune infiltration and highly elevated
levels of Ifih1
eGFP
Ifih1
Ins2
DNA

30. A subset of islets show potential immune infiltration and highly elevated levels of Ifih1

Contradiction between RNA-seq and smFISH
smFISH
No significant overexpression of the 3 ISGs
RNAseq
Very significant overexpression of the 3 ISGs

31. Contradiction between RNA-seq and smFISH

The RNAseq show very high variance in betaKO samples
B-KO

32. The RNAseq show very high variance in betaKO samples

A portion of infiltrated islands might explain the high variance in the
RNAseq data
Ifih1
B-KO: 1
2
Pdl1
3
4
B-KO: 1
2
Ins
3
4
B-KO: 1
2
3
Normal islets
ISGs
Ins
Immune
4

33. A portion of infiltrated islands might explain the high variance in the RNAseq data

Conclusions
• Non-infiltrated islets show no expression of ISGs
• No correlation between Adar1 knockout and ISGs expression in beta cells
• Infiltrates:
• Infiltration at 7 days <5%
• Beta cells lose Ins2 expression in the mRNA level (means the effect is on transcription, not stress
translation)
• Beta cells that remain in the infiltrates – become extreme
• Adar-KO cells are depleted (low % in late stage of infiltration)
• B2m & Ifih1 are rarely expressed in beta cells (eliminated first?)
• Pdl1 is induced in part of beta cells, co-occurrence with loss of Ins2 (protects from quick death?)

34. Conclusions

The results suggest even stronger human response to RNA editing
• Mouse don’t have Alu elements
• Alu elements are the main target of Adar in humans (>90%)
• A-I editing is more prevalent in the human transcriptome than in mice (at least 100-fold)
• This suggests that even a small defect in RNA editing could cause dramatic effects in
humans

35. The results suggest even stronger human response to RNA editing

My interpretation of the results
Stochastic SINE expression (methylation & histone modification H3K9Me)
Stochastic RNA editing
Stochastic IFIH1 binding
Alu expression
Cell #1
Cell #2
IFIH1 detection / MAVS activation – threshold
INF response
No INF response
Immune recruitment
MHC1 upregulation
INF response propagation to the whole islet
Insulitis
T1D

36. My interpretation of the results

Future plans
• Try to image more infiltrates (so far only 2 per mice) – later time points
• Quantify the change in ISGs in Glucagon area
• scRNAseq

37. Future plans

Thank you for listening!