A Proteomics and Transcriptomics Approach to Identify Leukemia Stem Cell (LSC) Markers

A Proteomics and Transcriptomics

Approach to Identify Leukemia Stem

Cell (LSC) Markers

Presented by:

Somayeh Haji Kazem Nili, Kalyani Rajalingham,

Nataliia Samus

Bonardi et al., (2013) MCP 12(3): 626-637

Summary

PM proteins mediate hematopoietic stem cells interaction

with their niche

Changes in these interactions may cause Acute Myeloid

Leukemia (AML)

Aims:

• Characterization of Plasma Membrane (PM) composition in AML

• Identifying markers within PM to recognize and target AML

Creator/Presenter: Somayeh Haji Kazem Nili

Background

What is AML?

How AML maintains?

What is LSC?

LSC’s PM proteome

characterization and AML

development

Understanding PM

proteome improves LSCs• Identification

• Isolation

• Targeting

Creator/Presenter: Somayeh Haji Kazem NiliRoboze et al., (2009) Expert Rev

Hematol. 2(6): 663-672

Leukemia development

Research Components1. Proteomics

• Sample: two groups of untreated AML cells (CD34+ and CD34- )

• Method: nano-LC/MS/MS

• Identification of CD34+-specific PM protein profile

2. Transcriptomics

• Sample: AML CD34+/CD34- and normal CD34+/CD34-

• Method: Illumina bead microarray

• Classification of eight AML subgroups associated specifically to

PM expression profile

3. Characterization

• Sample: PM marker genes of AML

• Method: Gene Set Enrichment Analysis (GSEA)

• Characterization of identified subgroups based on specific cellular

processes and prognosis


Proteomics Experimental procedure

1. Select cells from two patients :

• AML1: a poor risk AML patient

• AML2: myeloid blast crisis patient

2. Sample sorting by MoFLo-XDP

3. Membrane purification by simplify purification procedure

4. Complexity reduction by MuDPIT combined with nano-

LC/MS/MS

5. Protein digestion by trypsin

6. Peptides separation by a SCX chromatography and RP

chromatography column coupled with an LTQ-OrbiTrap MS

7. The MS/MS result has been searched against ipi-Human

database using Mascot, Sequest, and X!Tandem

Creator/Presenter: Somayeh Haji Kazem NiliBonardi et al., 2013

Proteomics workflow

ProteomicsPM proteins identified in CD34+

Bonardi et al., 2013Creator/Presenter: Somayeh Haji Kazem Nili

Novel

proteins

Described proteins

Proteomics

B, Number of proteins per sampleC, Composition of samples

Results:

Creator/Presenter: Somayeh Haji Kazem NiliBonardi et al., 2013

D, Number of PM

in CD34+ fractions

E, GO anotation for

biological processes

DiscussionConclusion

• 619 unique PM proteins in CD34+ from AML1

• 386 unique PM proteins in CD34+ from AML2

• novel markers;

CD82, CD97,CD99, PTH2R, ESAM, MET, ITGA6

• Previously described markers;

CD44, CD47, CD135, CD96, ITGA5

Objections

• Low amount of material limits quantification of less

abundant PM proteins

• Only two patients

• Proteome approach did not allow quantitative evaluation

• Combining proteomics with transcriptomics approaches


Transcriptomics

1- Control:

NBM – Normal Bone Marrow

Treatment:

AML CD34+, AML CD34-

2- Illumina Bead Array:

Microarray

3- Select significant genes using

statistical test

4- 238 AML CD34+ up-

regulated genes

Creator/Presenter: Kalyani RajalinghamBonardi et al., (2013) MCP 12(3): 626-637

Transcriptomics versus Proteomics

1- Compare Transcriptomics

and Proteomics

2- 59 proteins found in both the

proteins, and transcriptomics

sections (plasma membrane

proteins)

3- Function of the 59 proteins -

leukemic stem cell markers


Proteins found in both procedures depicted

Transcriptomics


Over-expressed,

and found in both

the

transcriptomics,

and proteomics

(in blue)

Putative

Marker

(Should have

been found –in

blue)

Transcriptomics


On the

protein, and

transcript

list, some

known

markers

were not

present (eg:

CD135).

Verification,

and

validation of

putative

markers

Transcriptomics

Validation of expression of CD135

(FLT3), CD47, CD96, PTH2R,

and CD49f (ITGA6)

1- Patients grouped as 2002-120,

2005-289, etc…

2- Using Illumina BeadArray,

confirmed expression of CD135

3- Repeat using either FACS or

Array for the remaining


CD

135

Transcriptomics

Validation of expression of CD135

(FLT3), CD47, CD96, PTH2R,

and CD49f (ITGA6)

They found that:

CD135 (FLT3)

CD47

ITGA6

CD96

PTH2R

Showed “enhanced expression”


CD

135

Discussion


Conclusion

• Transcriptomics show that 238 genes were found to be upregulated, of

which 200 were associated with the PM

• Of the 200 genes, 59 were also detected in the proteomics section

• Validation of CD135, CD47, ITGA6, CD96, and PTH2R shows that they

are over-expressed in AML CD34+

• CD135(FLT3) – strongest marker, overexpressed, found in both

proteomics/transcriptomics

• New markers: CD82, PTH2R, ESAM, MET, and ITGA6

Discussion


Objections

• 141 transcripts were not detected via proteomics

• Putative markers NOT detected in initial study

Characterization

Creator/Presenter: Nataliia Samus

8 types of acute myeloid leukemia:

The Leukemia & Lymphoma Society

http://www.lls.org/#/somedayistoday

Characterization

Creator/Presenter: Nataliia SamusBonardi et al., (2013) MCP 12(3): 626-637

Aim: to evaluate differences in

plasma membrane composition

of leukemia subtypes

Method: Selection of uncorrelated

membrane markers of CD34+ cells

Algorithm:

1. set of significantly upregulated genes

in AML CD34+ with relevant GO

annotation;

2. calculate information gain for selected

genes (base on gene expression level)

– allows to rank genes depending of

their predictive value;

3. find gene with max information gain;

4. remove all genes that are correlated to

selected genes;

5. repeat step 3 &4.

Characterization


Results: 8 membrane markers were identified.

Fig.4 A Supervised cluster analysis of expression of the 8 proteins

Conclusion: AML CD34+ samples can be separated from NBM CD34+

samples on the basis of the expression of the 8 markers.

Characterization


Aim: to evaluate whether these

8 subgroups would be

characterized by specific cell

processes.

Method: Selection of genes correlated

with the

8 membrane markers.

Algorithm:

1. expression of all genes was ranked

according to their correlation

coefficient in relation to the 8

membrane markers;

2. 8 lists of genes was formed;

3. gene set enrichment analysis (GSEA)

– to characterize genes functions in

each of the 8 phenotypes using

statistic approach

4. evaluation of good or poor prognosis

gene set

Results: 8 identified subgroups associate

with specific gene signatures.

Characterization


Fig.4 B Enrichment of membrane marker subgroups with genes

associated with specific biological processes

Characterization


Conclusion

1. Eight plasma membrane markers were identified that were

uncorrelated within cohort of 60 AML samples.

2. GSEA analyses indicates that these subgroups are characterized by

specific cellular processes (mostly associated with cancer

development).

3. Strong positive correlation with good or poor prognosis signature

was found (not yet supported by clinical data).

4. Further functionally validation studies are required.


Thank you!

Any questions?

A Proteomics and Transcriptomics Approach to Identify Leukemia Stem Cell (LSC) Markers

Science

aml cd34 cd34

normal cd34 cd34method

cd34 method

cd34 bonardi

cd34 fractionse

summary pm proteins

number of pm

somayeh haji kazem nilibonardi