1. A SAGE Approach to Discovery of Genes Involved in Autophagic Cell Death

2. Genome Sciences Centre Suganthi Chittaranjan University of Maryland
Acknowledgements
Genome Sciences Centre
Victor Ling
Marco Marra
Functional Genomics
Suganthi Chittaranjan
Doug Freeman
Carrie Anderson
Shaun Coughlin
Sequencing
Sequencing Team
Bioinformatics
Steven Jones
Erin Garland
Richard Varhol
Scott Zuyderduyn
SAGE team
University of Maryland
Biotech Institute
Eric Baehrecke

3. Programmed Cell Death (PCD)
Function
Dysfunction
Cancer
Deleting damaged cells
Autoimmune diseases
Culling cell number
Neurodegenerative
diseases
Deleting structures
Developmental
abnormalities
Sculpting tissues

4. Programmed cell death Type I = Apoptotic cytoskeletal collapse
condensation and fragmentation
of chromatin and cytoplasm
preservation of organelles
phagocytosis by macrophages or neighbouring cells (heterophagy)
occurs in isolated cells
Type II = Autophagic
preservation of cytoskeleton
formation of vacuoles that
engulf cytoplasm and organelles
fusion of vacuoles with lysosomes for self-degradation
late chromatin condensation and nuclear degeneration
occurs in groups of cells

5. Autophagic cell death in normal physiology
Dictyostelium sorocarp formation
insect metamorphosis
intersegmental muscle, gut, salivary glands
mammalian embryogenesis
regression of interdigital webs, sexual anlagen
mammalian adulthood
intestine, mammary gland post-weaning, ovarian
atretic follicles

6. Autophagic cell death in disease
human neurodegenerative diseases (Alzheimer and Parkinson)
cardiomyocyte degeneration
spontaneous regression of human neuroblastoma
tamoxifen-treated mammary carcinoma cells (MCF-7)
bcl-2 antisense treatment of human leukemic HL60 cells
beclin-1 (apg6) promotes autophagy and inhibits tumorigenesis;
expressed at decreased levels in human breast carcinoma

7. Questions
What is the relationship between Autophagic Cell Death and Cancer?
common mechanism in breast and other cancers?
gene mutation
What is the therapeutic potential of autophagic cell death in cancer?
solid tumours
apoptotic-resistant tumours

8. Aims Identify the genes involved in autophagic cell death in vivo.
Determine which genes are necessary and sufficient for
autophagic cell death.
Determine if genes are conserved in mammalian autophagic
cell death.
Identify the autophagic cell death genes associated with human
disease and investigate potential as molecular markers and/or
therapeutic targets.

9. Experimental Approach
Gene expression profiling (SAGE):
Comprehensive
Gene Discovery
Drosophila model system:
Known cell death genes/pathways are conserved
Genetic and molecular tools
Sequence resources
FlyBase and GadFly databases
Multiple tissues undergo PCD; well-characterized

10. Drosophila salivary gland PCD
autophagic
stage-specific
synchronous
(Jiang et al., 1997)
hr (APF, 18°C)
known cell death genes are regulated transcriptionally
Reverse transcription
diap2
rpr
hid
RT-PCR analysis

11. An ecdysone induced transcriptional cascade regulates salivary gland cell death
BFTZ-F1
EcR/USP
BR-C
E74
E93
rpr
hid
ark
dronc
crq
diap2
Cell
Death
E75

12. Overview of SAGE tag abundance
Library
Total tags analyzed
Tag species
% of tag species seen at frequency: 1
2-10
11-100
>100
16 hr
34,989
3,126
32.7
55.7
10.4
1.2
20 hr
31,215
3,034
38.0
50.9
9.7
1.4
23 hr
30,823
2,963
33.3
54.0
11.2
Total number of different tag species in all three libraries is 4,628.

13. Tag-to-gene Mapping in Drosophila
(E. Pleasance, M. Marra and S. Jones, in preparation)
Resources: Drosophila genomic sequence, full-length cDNAs,
ESTs, salivary gland ESTs
4,628 tags:
2866 (61.9%) – known or predicted genes
289 ( 6.2 %) – genomic DNA and EST (but no predicted gene)
(25.3%) – genomic DNA and/or reverse strand of gene
303 ( 6.5%) – no match

14. Comparison of SAGE and real-time quantitative RT-PCR
I. Direction of Change:
91/96 samples = 95% concordance
II. Correlation coefficient between fold-difference values (64 samples):
Fold-difference by SAGE
Fold-difference by QRT-PCR
Correlation coefficient = 0.5

15. SAGE Identifies Genes Associated Previously With Salivary Gland Death
BFTZ-F1
EcR/USP
BR-C
E74
E93
rpr
hid
ark
dronc
crq
iap2
Cell
Death
E75

16. 1244 genes are expressed differentially prior to salivary gland PCD
512 genes have associated
biological annotations (Flybase Gene Ontology)
732 genes have
unknown functions
377 of these
were unpredicted

17. Secondary screening of differentially expressed genes
I. Data Mining
II. Gene expression in salivary gland cell death-defective
mutant (E93)
III. Gene expression in other dying tissues
IV. Loss-of-function and gain-of-function mutant analyses

18. I. Data mining by sequence similarity searches and keyword queries
SAGE tag maps to Drosophila gene
SAGE tag maps to novel EST
Extract gene sequence from GadFly Database
GadFly – Swissprot Homology (tBLASTX)
tBLASTX search EST vs Swissprot
Keyword query of Swissprot comments,
keywords and identification fields
Keyword query of FlyBase and PCD database

19. Keyword Gene Sim (%) (aa) Swiss Prot id Database DescriptionSG 16 23
Gene
Sim
(%)
Score
Length
(aa)
Swiss
Prot id
Database Description
death
apoptosis
survival
autophagy
hormone
cancer
tumor
TNF
BH3 SP FB
PCD 4 1 2 45 10 65
EST only 15 6 5 7 61 73 12
102 Nc
debcl
cact
Traf1
CG4719
Ptpmeg
CG4859
CG10777
Rpn2
stck
chrw
ciboulot
CG8706
Atet
CG13907
botv
sp6
CG11335
CG10990
CG4091
CG2023
27.2
25.5
37.1
37.0
39.3
31.3
41.6
18.7
23.7
65.7
35.3
68.6
29.6
39.9
33.2
48.6
28.3
42.7
59.9
59.6
53.9
251
127
322
300
232
101
772 87
160
1276
257
114
1204
469
311
2044
427
423
592
330
247
253
133
203
179
116 80
407
444
267
321
170 34
984
240
838
361
182
188
228
ICE6_HUMAN
BCL2_HUMAN
IKBA_HUMAN
TRA1_HUMAN
BAR1_HUMAN
PTND_HUMAN
MM11_MOUSE
WRN_HUMAN
PSD2_HUMAN
PINC_HUMAN
RB24_MOUSE
TYB4_HUMAN
LRP2_HUMAN
ABG2_HUMAN
MOT1_HUMAN
EXL3_HUMAN
MASP_MOUSE
LYOX_HUMAN
CASPASE-6 PRECURSOR
APOPTOSIS REGULATOR BCL-2
NF-KAPPAB INHIBITOR ALPHA
TRAF1
BRCA1-ASSOCIATED RING DOMAIN
FAS-ASSOCIATED PTP-1
STROMELYSIN-3 PRECURSOR
WERNER SYNDROME HELICASE
26S PROTEASOME S2
PINCH PROTEIN
RAS-RELATED PRTEIN RAB-24
THYMOSIN BETA-4
LDL RECEPTOR-RELATED PROT
BREAST CANCER RESISTANCE
MONOCARBOXYLATE TRANSPORT
TUMOUR SUPPRESSOR EXL3-LIKE
PROTEASE INHIBITOR 5
PROT-LYSINE 6-OXIDASE PREC.
MM 'APOPTOSIS PROTEIN MA-3'
HS 'TNF-INDUCED PROTEIN GG2-1'

20. II. Analysis of differential gene expression in E93 mutant salivary glands
E93 is a DNA binding protein required for salivary gland cell death
(Lee et al. 2000)
65 different E93 chromosomal binding positions are known
Expression of known cell death genes is reduced in E93 mutant
salivary glands
we tested 20 differentially expressed genes with map positions
corresponding to E93 binding sites

21. Gene expression is reduced in E93 mutant salivary glands
Cytological location
OreR SAGE (16:23)
OreR RT-PCR (16:23)
E93 RT-PCR (16:23)
E93 RT-PCR (16:30)
CecB
CG4091
CecA1
Doa
CG4859
CG14995
BACR19J1.2
CG3845
Ptpmeg
CG9321
Cp1
CG12789
CG1216
larp
CG7860
Cyp1
Phm
Sox14
CG13448
CG8149
ark (control)
3R 99E4-99E4
2R 59F5-59F5
3R 98F1-98F2
2R 60D10-60D10
3L 64A7-64A8
X 1B10-1B10
2R 49E1-49E1
3L 61C1-61C1
2L 29E4-29E4
2R 50C20-50C20
2L 28A1-28A1
3L 61A6-61A6
3R 98C3-98C3
X 13E3-13E3
X 14B15-14B15
2R 60A14-60A14
3L 71E1-71E1
3R 85D25-85D25
N/A 
> 30
105
> 106
> 38 12
> 22 34 25
> 4 23 7
> 11
10.5 11
> 7
7.2 6 15
2765.7
206.7
177.5
71.9
47.2
40.8
34.6
29.6
14.2
13.7
12.8
10.7
9.1
7.8
4.9
3.5
3.4
2.9
2.5
2.1
2.0
0.1
0.8
0.4
1.1
0.6
1.4
0.5
0.7
0.9
0.2
3.2
3.0
34.2
2.7
2.6
1.2
1.5
1.9
1.8
2.2
19/20 upregulated genes showed a reduction in relative levels of
transcription.

22. III. Expression of salivary gland genes in the embryo coincides with patterns of PCD
CG3132
Cp1
akap200
akap200
3/15 genes examined have embryonic expression patterns
that at least partially coincide with patterns of PCD

23. IV. Loss-of-function mutant analyses.
salivary glands
embryos
retinas
e.g. akap200 mutants indicate a possible defect in PCD:
wild-type (41 hr APF)
akap200EP2254 (41 hr APF) * * * *

24. Differentially expressed genes reveal molecular features associated with autophagic cell death
Autophagic cell death involves the induction of genes required
for protein synthesis
e.g. 6 different translation initiation factors
Novel transcription factors appear upregulated coordinately with
known transcriptional regulators
e.g. maf-S, CG3350
Components of multiple signal transduction pathways are involved
e.g. TNF-a like pathway, akap200, Doa

25. Differentially expressed genes and molecular
features of autophagic cell death, continued.
apg-like genes can be regulated transcriptionally and this regulation
is associated with autophagic cell death
e.g. genes involved in two ubiquitin-like pathways
CG6194 (apg4;novel cysteine protease)
CG5429 (apg6/beclin-1)
Other autophagy-associated genes are likely involved
e.g. lysosomal enzymes, rab-7
Autophagic and apoptotic cell death appear to utilize at least some
common pathways or pathway components

26. Future Directions Autophagic cell death in cancer
1. Extent of role of autophagic cell death in different types of cancer
Gene expression comparisons btn wt, cancers and Dm SAGE (Erin)
Need to develop autophagic PCD specific markers
Expt: expression profiling in tamoxifen-treated MCF-7 cells
2. Is it possible to induce autophagic PCD in tumours?
candidate genes to test: CG6194 (a cysteine protease)
(Is it possible to induce autophagic PCD in apoptotic-resistant tumours?)
3. What molecules are necessary and/or sufficient for autophagic PCD?
Mutant analyses in Drosophila (existing mutants; RNAi)
GAL4/UAS system to induce tissue-specific ectopic expression

27. SAGE (Velculescu et al. 1995)
Stage 2
Stage 1
Expression level
Bioinformatics:
Quantitate tags
Correlate tags to genes
genes
genes

28. Autophagic cell death shares morphological features with autophagy

29. Transcription Profiling of Autophagic Cell Death
Gorski S, Anderson C, Chittaranjan S, Freeman D, Garland E, Jones S, Varhol R, Zuyderduyn S, Marra M
Verification of Differential Expression
1. Abstract
4. Tag-to-Gene Mapping in Drosophila
7. Verification of Differential Expression
During Drosophila metamorphosis, the larval salivary glands undergo steroid hormone-regulated programmed cell death (PCD). While dying salivary gland cells are detectable by markers of apoptosis, morphological analyses indicate that salivary gland destruction occurs by autophagy. Distinguished by the formation of acidic autophagic vacuoles that facilitate cellular self-destruction, autophagic cell death is a type of PCD that is not well characterized. To identify the genes involved in autophagic cell death, we used Serial Analysis of Gene Expression (SAGE) to examine salivary gland transcripts from three successive stages of wild-type salivary gland development leading up to PCD. We detected over 800 differentially expressed genes, including known cell death genes demonstrated previously to be upregulated prior to salivary gland PCD. Bioinformatic analyses revealed additional differentially expressed genes with sequence similarities to cell death related genes from other organisms. For a subset of the differentially expressed genes, we are currently verifying differential expression by quantitative real-time RT-PCR, examining gene expression in mutant backgrounds, and analyzing gene expression in cell death stage Drosophila embryos.
To identify genes corresponding to SAGE tags, we created a Drosophila tag-to-gene mapping program that utilizes genomic DNA sequence and cDNA sequences from the Drosophila Genome Project. In addition, to facilitate specifically the mapping of our salivary gland SAGE tags, we constructed a salivary gland specific cDNA library and generated 3’ ESTs quality 3’ ESTs were clustered and found to represent 1696 different transcripts. Of these, 1355 matched predicted genes and 341 (20%) are potentialy novel genes. These novel genes were used to map 210 tag species in the SAGE libraries.
To verify independently the differential expression determined by SAGE, we conducted real-time quantitative RT-PCR analyses for a subset of 50 genes. The graph below indicates the fold-difference in expression (between 16 hr and 23 hr) as determined by SAGE and real-time RT-PCR.
4628 tag species
Correlation coefficient = 0.64
2561 map to predicted genes
(96% unambiguous)
Align me!
1764 do not map to predicted genes
&/or genomic sequence
1461 map to ESTs
(6.5% of total tag species)
303 have no match
Fold-difference by real-time RT-PCR
Align me!
210 correspond to salivary gland ESTs
that represent potentially novel genes
2. Salivary Gland Cell Death
5. Known salivary gland death-related genes identified
Multiple Pictures
Fold-difference by SAGE
Salivary gland cell death is regulated by a transcriptional hierarchy induced by the steroid hormone 20-hydroxyecdysone (ecdysone). Following a pulse of ecdysone at the prepupal-pupal stage transition, the transcription of several known and highly conserved cell death genes is regulated. Shown below are the results of RT-PCR analyses using the known anti-death gene, diap2, and the pro-death genes, reaper (rpr) and hid. Developmental stages are indicated in hours (hr) after puparium formation (APF) at 18°C.
Several ecdysone-induced and cell death genes associated previously with salivary gland death (e.g. Jiang et al., 1997; 2000; Lee et al. 2000) were identified in the SAGE libraries. BR-C, E74, E75 and E93 are ecdysone-induced primary response genes that act upstream of the cell death genes reaper (rpr), hid, (not detected), ark, dronc, iap2 and croquemort (crq). The gene expression profiles generated by SAGE indicate that upstream transcriptional regulators can be distinguished temporally from downstream death effector molecules.
8. Data Mining
To annotate further the set of genes expressed prior to Drosophila salivary gland death, we are developing automated data mining tools that incorporate publicly available databases, sequence similarity, and user-defined keywords. RT hr
diap2
rpr
hid
                                                                           
EcR/USP
BR-C
E93
E74
E75
rpr
dronc
ark
hid
crq
diap2
death
cell
                                         
SAGE tag maps to Drosophila
predicted gene
Extract corresponding protein sequence
From GadFly Database
GadFly – Swissprot Homology (BLAST)
Keyword search of Swissprot description
SAGE tag maps to novel EST
BLAST search EST vs Swissprot
Following the induction of cell death gene expression, the entire salivary gland undergoes cell death in a rapid, stage-specific, and virtually synchronous manner. Acridine orange staining, shown below, indicates the temporal specificity of salivary gland cell death. Tissue destruction is evident by 24 – 25 hrs APF at 18°C.
6. Differential gene expression
To identify additional genes expressed differentially prior to salivary gland cell death, we conducted pairwise comparisons between the SAGE libraries. Below is a view of the 16 hr vs 23 hr comparison that was generated using SAGEspace, an in-house graphical tool for visualizing results of SAGE experiments.
22 hrs: no staining detected
23 hrs: staining in one nucleus
24 hrs: staining in ALL nuclei
522 genes (12%) are upregulated (p <.05)
331 genes (8%) are downregulated
16 hr 23 hr
9. Pilot screen: embryonic gene expression
3. Salivary gland SAGE libraries
small vertical clearance check
The expression of several known cell death genes foreshadows death not only in the salivary gland but in other tissues as well. To identify other genes with a possible role in cell death beyond the salivary gland, we initiated a pilot screen to examine the expression pattern of upregulated salivary gland genes in cell death stage embryos. Preliminary results indicate some overlap with regions of embryonic apoptotic cell death.
Micro-SAGE libraries were constructed from three successive stages of salivary gland development: 16 hrs APF (SG16), 20 hrs APF (SG20), and 23 hrs APF (SG23). The total number of SAGE tags analyzed includes tags of >95% sequence quality seen at least twice in all libraries combined. The total number of different tag species in all three libraries is 4,628.
TUNEL
Library
SAGE
Total tags
analyzed
Tag species
% of tag species seen at frequency: 1
2-10
11-100
>100
16 hr
20 hr
23 hr
34,989
31,215
30,823
3,126
3,034
2,963
32.7
55.7
10.4
1.2
38.0
50.9
9.7
1.4
33.3
54.0
11.2
CG2444
Cp1
RNA
in situ
Cell death occurs mainly in the head region.
Lateral view of mid-stage embryos.
Cell death occurs predominantly in the nervous system.
Ventral view of late-stage embryos.
Dr. Michael Smith
10. Acknowledgements
S. Gorski is a Research Fellow of the National Cancer Institute of Canada supported with funds provided by the Terry Fox Run. E. Garland is a Michael Smith Foundation for Health Research Trainee. M. Marra is a Michael Smith Foundation for Health Research Scholar. We gratefully acknowledge the support of the BC Cancer Agency, BC Cancer Foundation, and NSERC.
Founding Director

30. 1244 genes are differentially expressed prior to salivary gland PCD
16 hour
23 hour
522 genes (12%) are upregulated (p < 0.05)
331 genes (8%) are downregulated
1244 genes
512 genes with biological annotations
(Gene Ontology)
732 genes with unknown function
377 of these were unpredicted

31. PCD is highly conserved
Dcp-1
Drice
Decay
Damm
Diap1
Diap2
Deterin
Hid
Grim
Reaper
Ras
Raf
Mapk
dBorg-1
dBorg-2
Ark
Cell Death
Egfr
Dm-P53
EcR/USP
Dredd ?
Strica
Dronc
cytC UV
dFadd

32. Autophagy
bulk degradation of cellular proteins through an autophagosomic-
lysosmal pathway
can be induced in yeast and in mammalian cells by starvation
conditions
genetic screens in yeast have identified the genes involved
at least 16 autophagy-defective (apg)
and autophagy (aut) genes identified
genes encode components of two ubiquitin-like systems
unknown whether the same molecular mechanisms are involved in
autophagic cell death