of 27
Identification of a distinct ductal subpopulation with self
-
renewal and differentiation potential from the
adult murine pancreas
Tremblay and Ortiz
et al.
Supplementary Figure 1:
Gating strategy for fractionation of the
parent
CD133
high
CD71
low
population.
Supplementary Figure 2:
Immunofluorescent control staining for pancreas cell markers.
Supplementary Figure 3:
Cell
-
adhesion marker staining of pancreas colonies and controls.
Supplementary Figure 4:
Quantification of nuclei per individual FSC
mid
-
high
cluster
.
Supplementary Figure
5
:
Hprt
-
dsRed
mouse tool validation.
Supplementary Figure
6
:
Drop
-
Sequencing QC analysis and cluster identification.
Supplementary Figure
7
:
FSC
low
and
FSC
mid
-
high
sub
-
populations
qRT
-
PCR.
Supplementary Figure
8
:
Top upregulated GO and KEGG pathways among all clusters.
Supplementary Figure
9
:
Pancreas progenitor gene expression pattern among all clusters.
Supplementary Figure
10
:
StemID
analysis of Drop
-
sequencing results and identification of
StemID
cluster
03 units among all clusters.
Supplementary Figure 1
1
:
Differentially expressed (DE) genes between cluster 1, TP units,
and
StemID
cluster 03.
Supplementary Figure 1
2
:
Control samples for
acinar injury model and TP cell analysis among
normal
intercalated, small interlobular, and main ducts.
Supplementary Figure 1
3
:
Gating strategy for control and injured pancreas samples.
Supplementary Figure 1
4
:
Male and female cohorts data separated fo
r 3
-
days after acinar injury.
Supplementary Figure 1
5
:
Epithelial rosettes are still present 14 days after acinar injury and express
ductal
makers.
Supplementary Figure 1
6
:
Graphical abstract.
Supplementary
Table
1
:
Taqman
probes used for conventional and microfluidic qRT
-
PCR analyses.
Supplementary
Table 2
:
List of
a
ntibodies.
Supplementa
ry
Figure 1. Gating strategy for fractionation of the parent CD133
high
CD71
low
population.
Pancreas c
ells
stained for CD133 and CD71
were sequentially gated using the following regions (R): R1
eliminated cell debris; R2 and R3 eliminated cell doublets; R4 eli
minated dead cells
based on DAPI
; R5 gated
CD133
high
CD71
low
cells. R5 cells were
further
analyzed by size (FSC
-
A) and granularity (SSC
-
A), which revealed
4 sub
-
populations. Percentages of cells in gates are shown from a representative experiment. Abbreviat
ions:
FSC
-
A, forward scatter
-
area; SSC
-
A, side scatter
-
area; FSC
-
W, forward scatter
-
width; SSC
-
W, side scatter
-
width; APC, Allophycocyanin; PE
-
Cy7, Phycoerythrin
-
cyanine dye
7
.
Supplementary Figure
2
.
Authentication of antibodies used in
i
mmunofluorescen
ce
(IF)
staining
with
positive control pancreas tissues
.
Double
and triple
immunofluorescence (IF) analysis
of
mouse pancreas
at
e
mbryonic day
12.5
(
E12.5;
left
and
center) and
p
ost
-
natal day 0
(
P0;
right)
were used
to
validate
the
expected
staining
patterns of
the
primary
antibod
ies
that
we employed
.
Scale bar
s
=50
m.
Supplementary
Figure 3. Cell
-
adhesion marker staining
of pancreas colonies and
negative
control
staining
.
(a)
S
amples treated with s
econdary
antibodies
only
were
used to
rule out
non
-
specific staining on FSC
mid
-
high
clusters
for E
-
cadherin (Goat anti
-
488), JAM
-
A and ZO
-
1 (Rabbit anti
-
Cy3)
. See also Fig. 4.
Scale bar
s
=
2
0
m.
(
b
-
c
)
IF analysis
of FSC
mid
-
high
fraction
-
derived
3
-
week
-
old
c
olonies grown in
Matrigel/RSPO1
colony
assay
.
DAPI
(blue
)
identif
ies i
ndividual cells
and
E
-
cadherin (green)
demarcates
cell
-
to
-
cell borders. DAPI and E
-
cadherin
were co
-
stained with either
JAM
-
A (red,
b
) or ZO
-
1 (red,
c
).
Negative controls using s
econdary antibod
ies
only
were used to rule out
non
-
specific staining
on pancre
as
colonies
for E
-
cadherin (Goat anti
-
488), JAM
-
A and ZO
-
1 (Rabbit anti
-
cy3).
Scale bar
s
=
5
0
m.
Supplementary Figure
4
.
Quantification of nuclei per individual FSC
mid
-
high
cluster.
Three independent sorting experiments were performed to
collect FSC
mid
-
high
clusters followed by
Wright
-
Giemsa
staining to
manually
quantify
the number of nuclei in each cluster
.
For
cell sorting #1
(n=31)
the mean
number
of nuclei
wa
s found to be 8.1
±
3.5
SD
. For cell sorting #2
(n=31)
the mean
number of nuclei
was found to be
7.2
±
3.9 SD. For cell sorting #3
(n=31)
the mean
number of nuclei
was found to be 8.3
±
4.4 SD.
The data from
the three independent sorting experiments w
ere
combined and shown Fig. 4e.
Abbreviation:
Standard Deviation
(SD).
Supplementary Figure
5
. Hprt
-
dsRed mouse
tool
validation.
(a
-
c)
DsRed
fluorescence
image
analysis
of
pancrea
ta
from (a)
wild
-
type
female
(
Hprt
+
/
+
)
(negative control
, (b)
heterozygous
female
(
Hprt
DsRed
/
+
)
, and (c)
homozygous
male
(
Hprt
DsRed/
y
)
mice
(positive control)
.
As expected,
a
mosaic pattern of
membrane
-
bound DsRed labelled
-
cells (
cells in
red color)
in
heterozygous
female
(
Hprt
DsRed/+
)
was found
. (d
-
e)
Spleens were dissected from two 12
-
week
-
old
female
mice
and mashed with the plunger end
of a syringe to release splenocytes. Splenocytes were
filtered,
washed
,
and analyzed using flow cytometry. (d)
In wil
d
-
type female mice, DsRed staining was present at background levels (
4
% DsRed
+
cells
vs.
96% DsRed
-
cells
). (e) In
DsRed heterozygous female
mice, approximately half of the splenocytes expressed DsRed
(45%
DsRed
+
cells vs. 55% DsRed
-
cells
)
consistent with
a random mosaic pattern.
Supplementary
Figure
6
.
Drop
let
-
based RNA
-
Seq
uencing quality control (QC)
analysis and cluster
identification.
Pancrea
ta
from a total of 5 adult (2
-
4 month old) C57Bl/6 mice were collected, dissociated into a single
-
cell
suspension, stained with antibodies against CD133 and CD71, sorted using a fluorescence
-
activated cell sorter,
and
subjected to
droplet
-
based
RNA
-
seq
uencin
g
using the 10x Genomics
platform
.
Quality control, aggregation,
clustering, and a
nalysis
was done
using
Seurat
package
.
(a
) The
number
of
feature
genes (biologically
informative genes) and
m
RNA count
s
per
unit
showed that the sorted
FSC
low
fraction
has a unimodal distribution,
suggesting that it is a single population. In contrast, the
FSC
mid
-
high
fraction
appears to have a bimodal distribution,
suggesting the presence of both true single cells and clusters. The
percentage of
mitochondria genes per unit
was low in both populations, suggesting a maintenance of plasma membrane integrity
after tissue dissociati
on
.
(
b
) Approximately 1
5
% of events displayed higher than 10% mitochondria genes per unit
, indicating
a
compromised plasma membrane. Events that exceeded these criteria were excluded from subsequent analysis.
(
c
) There was a strong positive correlation
(R
2
= 0.98)
between the number of feature genes and the number of
mRNAs per unit, suggesting that the mRNAs identified were biologically informative. (
d
)
The elbow plot
analysis
indicated that
10
principal component
s were sufficient to capture the majority of
the variation in the data
.
(e) DB
index of multiple dimensions indicate
that
12 dimensions
contain the least variation
. (f)
Clustering
at 12
dimensions
with
a 0.6x
resolution
results in
8 clusters
(clusters 0 to 7)
.
(g)
Bubble
plot
showing
the
expression
of
lineage
markers
typically found in
the
pancreas.
D
uctal
lineage
gene
s are found
in most of the clusters (
0
through
6
)
,
with cluster 6 expressing
additional
acinar
marker
genes
.
C
luster 7 express
es
immune
cell
genes.
(h) UMAP of
clusters 0 to 7 generated with Seurat after
combin
ing the datasets for
FSC
low
and
FSC
mid
-
high
units.
(
i
)
Gene expression
levels of
various
lineage
markers
in
the
UMAP
plot
shown
in h
.
Supplementary Figure
7
.
Quantitative
RT
-
PCR
analysis of
the
pre
-
sort
ed
pancreas cells
and after sorting
for
FSC
low
and FSC
mid
-
high
fractions
.
Pancrea
ta
from a total of 5 adult (2
-
4 month old) C57Bl/6 mice were collected, dissociated into a single
-
cell
suspension, stained with an
tibodies against CD133 and CD71, FACS sorted, and
analyzed for expression of
Insulin 2
(a),
Amylase2a
(b)
, and
Sox9
(c)
using
conventional
qRT
-
PCR
relative to beta
-
actin
.
*p<0.05, **p<0.01,
***p<0.001, ****p<0.0001,
n=
3
.
Statistics were performed using
one
-
way ANOVA multiple comparison
s
two
-
tailed Student’s
t
-
test.
Supplementary
Figure
8
.
Top upregulated GO and KEGG pathways
in each
cluster
.
(a)
Upregulated
Gene Ontology (GO)
pathways for clusters 0
-
7.
(b)
Upregulated
Kyoto Encyclopedia of Genes
and Genomes (KEGG)
pathways for clusters 2
-
7
.
See also Fig. 6
and Supplementary Datasets 2
-
3.