a b c d e - mitweb.mit.edu/biophysics/papers/nature2009c_si.pdf · only the c-myc transgene and...

Supplementary Figure 1 | Strategy for long-term culturing and follow-up of monoclonal populations.

Schematic with representative images and immuno-staining demonstrating the long term-culturing strategy. a, In the

minority of cases, after single cell plating of Pre-B cells or monocytes in DOX, cells with ES-like morphology emerged

(black arrow) and resulted in Nanog-GFP detection by FACS within 2 weeks. b, In the rest of the wells, non/semi adherent

round cells grew that could be propagated in the presence of DOX. At the end of each week FACS was performed to test

for the appearance of Nanog-GFP + cells, and when negative (as shown), approximately 2-2.5x105 cells were re-plated on

gelatin and in the presence of DOX for continued follow-up analysis. c, This panel demonstrates the appearance of cells

with ES-like morphology among the small round “intermediate cells”. Nanog-GFP+ signal could be readily detected by

microscope and FACS. d, Following detection of Nanog-GFP+ the cells were plated in the absence of DOX, and iPSC

colonies were readily observed (black arrows) while partially reprogrammed non-adherent cells ceased to grow by DOX

withdrawal. Stable iPSC Nanog-GFP lines were established and expanded by 1-2 passages. e, Results of FACS analysis for

GFP detection performed on randomly selected monoclonal populations from different experimental groups. Middle

column shows results on non-adherent fraction of cells only obtained from the supernatant of growing wells. The analysis

shows that Nanog-GFP iPS cells can be detected only in the samples the contain the adherent fraction, consistent with

iPSCs being fully adherent on gelatin coated tissue culture dishes. % of Nanog-GFP+ cells is indicated.

a b c d e

SUPPLEMENTARY INFORMATIONdoi: 10.1038/nature08592

www.nature.com/nature 1

0

20

40

60

80

100

0.1-0.25% 0.25-0.5% 0.5%-1% > 1%

(7/12)

(10/12)

(12/12) (4/4)

Pe

rce

nt

of

su

cce

ssfu

l

de

riva

tio

n o

f N

an

og

-GF

P+

DO

X in

de

pe

nd

en

t iP

SC

lin

es

a b

0

20

40

60

80

100

Pe

rce

nt

of

su

cce

ssfu

l

de

riva

tio

n o

f N

an

og

-GF

P+

DO

X in

de

pe

nd

en

t iP

SC

lin

es

Fro

m m

onoclo

nal popula

tion w

ith

>0

.5%

Na

no

g-G

FP

fra

ctio

n

(12/12) (12/12) (12/12) (23/24)

B cells CD11b+ cells

Supplementary Figure 2 | Threshold for isolation of DOX-independent Nanog-GFP+ iPSC lines. a,

Experiments to determine the FACS detection threshold for Nanog-GFP in order to classify a given well as “being

reprogrammed”. As described in Fig. 2a, B cells were isolated from NGFP1 chimeras. Next, single cells were cloned and

grown in the presence of DOX. After five weeks of DOX induction, clonal populations were analyzed for the percentage of

Nanog-GFP+ cells in each population and sub-grouped into four categories based on Nanog-GFP fraction: 0.1%-0.25%,

0.25%-0.5%, 0.5%-1%, and >1%. Cells were subsequently grown in the absence of DOX and were passaged at least twice.

Each cell population was scored for the presence of DOX-independent Nanog-GFP+ clones. The presence of 0.5% GFP+ cells

in a given well reproducibly facilitated the isolation of DOX-independent Nanog-GFP+ iPSC lines. b, Experiments described

in (a) were performed with various different cell types: wild-type and p53KD B cells and CD11b+ myeloid cells. Data only

from wells in which initial detection of Nanog-GFP+ was >0.5% are shown. This analysis demonstrates that a Nanog-GFP+

detection threshold of 0.5% is a reliable marker for isolation of iPSCs across different donor cell types. c, Representative

images of iPSC lines isolated and detection of endogenous Nanog-GFP.

%GFP+ Cell fraction at each well at

initial GFP detection time point

c

NG

FP

1

#7

NG

FP

1-

p53K

D #

22

Phase contrast Nanog-GFP

doi: 10.1038/nature08592 SUPPLEMENTARY INFORMATION


Supplementary Figure 3 | Characterization of B cell populations following DOX induction. a, Isolated B

cells were cultured in the presence of DOX and tested for surface expression of the indicated markers at time 0 (-DOX) and

3 and 5 days on DOX. Consistent with previous reports (Mikkelsen et al. Nature 2008 and Stadtfeld et al. Cell Stem Cell

2008) somatic cell markers (CD43, CD19, CD45 and IL-7R) were efficiently silenced by 5 days of transgene induction.

Percentage of positive cells is indicated in comparison to isotype match antibody control. b, FACS analysis of surface

markers after OSKM transgene induction in reprogramming Pro B derived samples grown in the presence of DOX. Percent

of positive cells is indicated in comparison to isotype match antibody control. Information regarding each clone is presented

as: clone #, week on DOX induction was tested, Nanog-GFP+ >0.5% status, and week # when Nanog-GFP+ became

significantly detected (>0.5%). Note that all reprogramming populations had lost expression of B/hematopoietic cell

markers consistent with suppression of somatic cell identity (shown in a). SSEA1 detection was heterogeneous between

different populations and was not a predictive parameter for whether a monoclonal populations would give rise to iPSC at

relatively early or late time points (compare clone #66 and #53). These results are consistent with fluctuating expression

pattern of SSEA1 surface marker during reprogramming (note clone #1 at w5 and w8 and Mikkelsen et al. Nature 2008).

B cell

(- DOX)

B cell

(3d on

DOX)

B cell

(5d on

DOX)

>95% 36% 3%

>95% 41% 2%

>95% 31% 0%

>95% 29% 2%

0% 0% 8%

CD43

CD19

CD45

SSEA1

IL-7R

CD43

CD19

CD45

SSEA1

IL-7R

0% 0% 0% 0% 0% 0% 0%

0% 0% 0% 0% 0% 0% 0%

0% 0% 0% 0% 0% 0% 0%

0% 0% 0% 0% 0% 0% 0%

25% 19% 0% 0% 15% 0% 8%

a b



Cu

mu

lati

ve %

Na

no

g-G

FP

+ W

ell

s

0

1

2

3

4

5

6

0 2 4 6 8 10 12

CD11b+ NGFP1 CD11b+ NGFP1-p53KD

Nu

mb

er

of

we

lls

gen

era

tin

g G

FP

+

cells p

er

week

%G

FP

+ C

ell

s a

t each

we

ll a

t

init

ial G

FP

dete

cti

on

tim

e p

oin

t

b

c

Latency (Weeks on DOX) Number of Events

Clone # week >0.5%GFP td (hr)

1 5 - 19.5

2 5 - 22.6

7 3 - 19.4

7 5 + 20.5

11 5 + 21.4

21 5 + 19.0

33 5 - 21.8

40 5 + 18.4

Mean 20.33

Stdev 1.49


5 3 - 9.59

5 5 + 9.50

18 5 + 9.65

21 5 - 10.0

Mean 9.69

Stdev 0.22

0510

15

0.51.01.52.02.53.03.54.04.55.05.56.0

CD11b+ NGFP1CD11b+ NGFP1-p53KD

0 5 10 15 200

20

40

60

80

100

CD11b+ NGFP1 cells n=24

0 5 10 15 200

20

40

60

80

100

CD11b+ NGFP1 n=48Cu

mu

lati

ve %

Na

no

g-G

FP

+ W

ell

s

50%

88%

Latency [Weeks on DOX]

85%

8.2

a

CD11b+ NGFP1 cells n=24

CD11b+ NGFP1 n=48

*

Median Latency (Weeks on DOX) Median Cd, Population-averaged # of Cell

Divisions before Nanog-GFP detection

0 2 4 6 8 10 12

CD11b+ NGFP1, n=48

CD11b+ NGFP1-p53KD, n=48

0 20 40 60 80 100

CD11b+ NGFP1, n=48

CD11b+ NGFP1-p53KD, n=48

d

Supplementary Figure 4 | Reprogramming of CD11b+ derived cell populations. a, CD11b+ derived cells were

seeded as single cells and cultured in DOX. Reprogramming of NGFP1 clonal CD11b+ populations was measured as the

cumulative number of wells that became Nanog-GFP+ over time. After 13 weeks, greater than 85% of wells generated iPSCs,

and after 16 weeks greater than 88% of wells generated iPSCs. Asterisk indicates that measurements were taken every 4

weeks. b, Exponential growth described the growth well for each clone (R2=0.97-1.0), and doubling time, td, was calculated

from these fits. No difference in doubling times was statistically significant (p<0.05) except between NGFP1 (blue) versus

NGFP1-p53KD (red) groups. c, Left: For reprogramming of CD11b+ wells, the number of wells generating Nanog-GFP+ cells

each week is plotted against the time in DOX containing medium. Right panel shows the %Nanog-GFP+ cells for each

individual well (open markers) at the week when GFP was initially detected (>0.5% GFP+). Horizontal solid lines represent

the average value of GFP+ fraction in the populations upon initial detection of a positive signal (defined >0.5% in the study;

see Supplementary Fig. 2) and indicate no significant difference between NGFP1 and NGFP1-p53KD populations (1.54% and

1.85%, respectively [p>0.05]). d, Median times and cell divisions during latency determined through parametric statistical

analysis as described in Supplementary Fig. 12. Notably, transgene induction levels where highly similar between monocytes

and B cell populations (Supplementary Fig. 8), possibly explaining the similar reprogramming kinetics observed for the two

distinct cell types.



Endoderm Mesoderm Ectoderm

NGFP1 #1 (W12)

NGFP1 #66 (W13)

40 XY

NGFP1 #1

(W12)

NGFP1 #72

(W14)

NGFP1 #16

(W5)

NGFP1 #20

(W10)

NGFP1 #9

(W3)

NGFP1 #47

(W10)

a

c

40 XY

NGFP1 #72 (W14)

40 XY

b

NGFP1 #66

(W13)

NGFP1 #7

(W5)

NGFP1 #53

(W5)

Supplementary Figure 5 | See figure legend on next page.



Supplementary Figure 5 | Characterization of NGFP1 Pre-B derived iPSC clones at different times following

DOX induction. a-c, Randomly selected iPSC lines were selected for analysis. Clone # is indicated for each panel and time on

DOX required to derive each line is indicated in parenthesis (W; weeks on DOX). a, Normal karyotype was observed by analyzing

different iPSC clones derived after 12-14 weeks on DOX. b, Characterization of B cell derived and DOX-independent iPSCs

demonstrating ES-like morphology, specific Nanog-GFP reporter detection by flow cytometry, and ability to generate in vivo

differentiated teratomas. c, Well differentiated teratomas were derived from all lines tested with evident formation of endoderm,

mesoderm and ectoderm structures. Table on the right summarizes the results of immunostaining / fluorescent detection of the

indicated markers on the DOX-independent iPSC clones. d, Chimera generated following injecting NGFP1 #72 iPSC clone

derived after 14 weeks on DOX induction in vitro. High contribution is evident by the agouti coat color.

NGFP1 #72

(W14)

d



Supplementary Figure 6 | Background apoptosis levels in reprogramming populations from NGFP1 B cells.

Apoptosis rates were quantified in reprogramming populations obtained from NGFP1, NGFP1-p53KD and NGFP1-p21 KD B

cells. a, Annexin V- FITC and propidium iodide staining on randomly selected clones are shown, and demonstrate background

(<1%) levels of apoptotic fraction (AnnexinV+ PI- in lower right quadrants). Information regarding each clone is presented as:

clone #, week on DOX when then induction test was performed, Nanog-GFP+ >0.5% status. We verified a Nanog-GFP negative

signal before analysis to avoid false positive detection of Annexin V FITC. Pro/Pre B cells were isolated from a mouse carrying

only the c-Myc transgene and Rosa26-M2rTta and grown for 3 days on OP9 feeders in the presence of IL-7 (Markoulaki et. al.

Nature Biotechnology 2009). Early apoptosis was quantified following 8 hr of DOX mediated c-Myc transgene induction. b, Flow

cytometry based terminal deoxynucleotidyl transferase dUTP nick and labeling (TUNEL) assay was also performed on randomly

selected cells throughout the process (weeks 3-10 on DOX). This independent analysis confirmed insignificant levels of apoptotic

cells in the DOX supported populations. These results are consistent with the notion that Oct4, Sox2, c-Myc and Klf4 could act in

concert as context dependent oncogenic and anti-apoptotic factors enabling hematopoietic cell propagation in the context of

optimized OSKM transgenes expression without additional transformation.

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

0 1 2 3 4

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H100

101

102

103

104

FL1-H

100

101

102

103

104

FL1-H

PI

11% <1%

<1%

<1% <1% <1%

<1% <1% <1%

<1% <1% <1%

NGFP1

NGFP1-

p53KD

NGFP1-

p21KD

-DOX +DOX (8h)

M2rTta +/-

C-Myc +

#1,w5,- #85,w6,- #20,w8,- #59,w10,-

#9,w3,- #42,w4,- #63,w5,-

#3,w3,- #11,w5,- #14,w3,-

a

NGFP1

(n=10)

NGFP1-

p53KD

(n=10)

NGFP1-

p21KD

(n=10)

% T

UN

EL

+ c

ell

s

b

p-value >0.12

p-value > 0.4



Supplementary Figure 7 | Characterization of NGFP1 primary iPSC lines used. a-b, NGFP1 iPSCs were infected

with lentivirus PsicoR plasmid encoding hairpins against mouse p53, p21 and CD8 (used as control against an irrelevant gene).

Infected cells were subcloned and viral integration was verified by PCR specific detection. Knockdown was specifically verified

by western blotting on the selected subcloned lines (NGFP1-p53KD, NGFP1-p21KD and NGFP1-Ctrl KD). c, NGFP1 line was

infected with a TetO-Nanog virus, and viral transgene integration was verified by southern and PCR (data not shown). Nanog

expression was verified by RT-PCR in NGFP1-NOE (Nanog over-expresser) iPSC line and derived B cell clones (1-3) grown in

the presence of DOX. d, FACS analysis for NGFP1 iPSC lines with and without additional genetic perturbations. Cells were

pre-plated on gelatin before the analysis to deplete feeders and differentiated fractions. The previously described biphasic

pattern of Nanog expression (Chambers et al. Nature 2007) was not altered by modulating the p53 pathway or overexpressing

Nanog. Moreover, the median fluorescence intensity (MFI) of Nanog-GFP+ positive cells was not significantly altered.

100 101 102 103 104

FL1-H

M1M2

100 101 102 103 104

FL1-H

M1M2

100 101 102 103 104

FL1-H

M1M2

100 101 102 103 104

FL1-H

M1M2

100 101 102 103 104

FL1-H

M1M2

68% 32%

69% 31% 68% 32%

70% 30% >99%

CO

UN

TS

NGFP1 NGFP1-p53KD

NGFP1-p21KD NGFP1-NOE

V6.5 ES

(negative control)

CO

UN

TS

C

OU

NT

S

CO

UN

TS

CO

UN

TS

MFI=115

MFI=109

MFI=105

MFI=108

ba

Actin

p53

NG

FP

1 NGFP1

-p53KD

p5 p8

Actin

p21

NG

FP

1

NG

FP

1

-p21K

D

NG

FP

1

-Co

ntr

ol

KD

c

Fo

ld m

RN

A t

o G

AP

DH

Viral Nanog

NGFP1-NOE pre-B

d



0

0.5

1

1.5

2

2.5

3

Fo

ld m

RN

A t

o G

AP

DH

Klf4

Sox2

Oct4

c-Myc

NGFP1 Pre-B

0

0.5

1

1.5

2

2.5

3 NGFP1-p53KD Pre-B

Fo

ld m

RN

A t

o G

AP

DH

0

0.5

1

1.5

2

2.5

3 NGFP1-p21KD Pre-B

Fo

ld m

RN

A t

o G

AP

DH

0

0.5

1

1.5

2

2.5

3

Fo

ld m

RN

A t

o G

AP

DH

NGFP1-NanogOE Pre-B

a

NGFP1

(n=10)

NGFP1-p53KD

(n=6)

NGFP1 (n=4)

Fo

ld m

RN

A t

o G

AP

DH

NGFP1-p21KD

(n=6)

NGFP1-NanogOE

(n=9)

Average per group of monoclonal populations on DOX

Supplementary Figure 8 | See figure legend on next page.



Supplementary Figure 8 | Summary for transgene induction levels. RT-PCR analysis of OSKM transgene

induction levels in reprogramming B (a) and CD11b+ (b) derived samples grown in the presence of DOX. Average relative

expression levels and standard deviation from 2-3 RT-PCR reactions are shown for all four factors. Polyclonal freshly

isolated cells grown in the absence of DOX were used as negative controls. Information regarding each clone is presented as:

clone #, week on DOX induction was tested, Nanog-GFP+ >0.5% status. Average induction levels for OSKM in different

genetic backgrounds following genetic perturbation are also summarized in the lower part of each subpanel. Overall the

results indicate that transgene expression was not altered by the different perturbation and that difference in transgene

expression was not an underlying cause for the difference in reprogramming latencies between monoclonal population

within each experimental group.

0

0.5

1

1.5

2

2.5

3

0

0.5

1

1.5

2

2.5

3

0

0.5

1

1.5

2

2.5

3

NGFP1 (n=4)

NGFP1-p53KD (n=3)

Fold

mR

NA

to G

AP

DH

NGFP1 CD11b+ NGFP1-p53KD CD11b+

Average per group of monoclonal populations on DOX

-Dox

#5,w3,- (+Dox)

#5,w5,+ (+Dox)

#21,w5,- (+Dox)

-Dox

#2,w5,- (+Dox)

#7,w3,- (+Dox)

#7,w6,+ (+Dox)

#23,w6,+ (+Dox)

b



Supplementary Figure 9 | Proliferation data for perturbed NGFP1 monoclonal populations. Tables

show the doubling time (td) for each Pre-B derived clonal population, listed as “clone #, weeks on DOX, Nanog-

GFP >0.5% status (+ or -).” Boxed rows delineate cases where the same clonal population was measured at

different times during DOX induction or after iPSC isolation. Lower lines labeled in green indicate data obtained on

subcloned or parental iPSC lines. No difference in doubling times was statistically significant (p<0.05) except

between the different perturbation groups.


1 3 + 9.50

2 3 - 9.17

2 5 - 9.48

4 3 + 9.80

9 3 - 9.62

9 5 + 9.40

22 3 - 9.37

42 5 + 9.77

17 3 - 9.70

17-iPS N/A + 10.2

p53KD-iPS N/A + 10.2


1 3 + 9.66

3 3 - 9.88

3 5 + 9.86

10 3 + 9.54

11 3 - 9.92

11 5 - 10.0

12 3 - 9.76

12 5 + 9.59

22 5 - 9.78

25 5 + 9.86

30 5 - 9.90

p21KD iPS N/A + 10.0

NG

FP

1-p

53K

D

NG

FP

1-p

21K

D


1 3 - 13.2

1 4 - 12.9

1 7 + 13.8

2 3 - 13.2

2 4 + 13.2

8 3 + 13.2

44 4 - 13.4

44 7 - 13.4

50 7 - 13.7

Lin28OE-iPS N/A + 13.6

NG

FP

1-L

in28O

E


1 4 - 17.0

1 7 + 16.3

10 4 - 16.5

10 6 + 16.7

15 4 - 16.5

15 7 - 17.0

25 4 - 17.2

25 6 + 16.1

32 6 - 16.4

32 7 + 15.9

41 6 - 16.8

41 7 + 16.7

NOE-iPS N/A + 14.8

NG

FP

1-N

an

og

OE



Supplementary Figure 10 | Characterization of NGFP1-p53KD B cell derived iPSC clones. a, NGFP1-p53KD

primary iPSC line showed normal karyotype at early passage (P6) and was injected into blastocysts to generate chimeras

from which Pre-B were derived for long term analysis. b, The NGFP1-p53KD B cell donor cells were derived from 3-5

week old mice; all analyzed clones carried different heavy chain rearrangement patterns, arguing against the possibility that

cells in individual wells were derived from a B cell tumor which may have resulted following p53 knockdown (which would

be expected to be mono/oligo clonal). c, Well differentiated teratomas were derived from all lines tested with evident

formation of endoderm, mesoderm and ectoderm structures. Table on the right summarizes the results of immunostaining /

fluorescent detection of the indicated markers on the DOX independent iPSC clones derived. Clone # is indicated for each

panel and time on DOX required to derive each line is indicated in parenthesis (w; weeks on DOX). d, Chimera generated

following injecting NGFP1-p53KD #42 iPSC clones derived following 5 weeks on DOX induction in vitro. Prior to injection

the cell line was labeled with a constitutively expressed lentiviral GFP expressing vector, and chimerism is evident by

specific GFP detection by fluorescent lamp (highlighted by the white arrow).

NG

FP

1-

p53K

D #

17

(w5)

NG

FP

1-

p53K

D #

22

(w4)

NG

FP

1-

p53K

D #

42

(w5)

NGFP1-p53KD Clone #

DH-J

HV

HJ5

58

VH7183

VHJQ

52

1 5 9 17 22 35 49 NGFP1-p53KD #42 (w5)

a b

c

NGFP1-p53KD line

(Passage #6)

40XY

d

Nanog-GFP



Supplementary Figure 11 | Reprogramming of NGFP1-p53KD CD11b+ derived cell populations.

As in Fig. 3c-d, latencies for reprogramming various clonal CD11b+ donor cell derived populations.

0 20 40 60 80 100 120 140 160 1800

20

40

60

80

100

50

90

NGFP1 (n=48)NGFP1-p53KD (n=48)

0 2 4 6 8 10 12 14 16 180

20

40

60

80

100

50

90

NGFP1 (n=48)

NGFP1-p53KD (n=48)

Cu

mu

lati

ve %

Na

no

g-G

FP

+ W

ell

s

Cu

mu

lati

ve %

Nan

og

-GF

P+

We

lls

a

Cd , Population-averaged # of Cell Divisions

on DOX before Nanog-GFP detection

b

Adjust latency

by td

*** p<0.0001

Latency, weeks

on DOX before Nanog-GFP detection

p=0.209

CD11b+ Wells CD11b+ Wells



Supplementary Figure 12 | Median latencies with parametric and non-parametric numerical analyses. a, For

each week, the percent of wells generating iPSCs is plotted for B cell NGFP1 and NGFP1-p53KD wells. The solid line is the

fitted gamma distribution for the two types of wells. Table on right lists the fit of various two-parameter distributions and the

associated chi-squared goodness of fit statistic ( 2) for all the events seen in the cell-division rate dependent acceleration and

NGFP1 B cell cases. Lower 2 indicate better fits. b, Statistics for comparing different reprogramming dynamics against both

time and cell divisions (Cd) between the well type listed on the left and NGFP1 B cell wells. The “Increase” column signifies a

p-value<0.05 and that the cumulative probability of wells generating iPSCs increased over the NGFP1 B cell case. c, Median

latency times and 95% confidence intervals were calculated from best parametric gamma fits. For p-values see part b. d,

Parametric analysis on latency data with cell division number yielded the median number of cell divisions during latency and

95% confidence intervals. For p-values see part b.

Median Latency (Weeks on DOX) Median Cd, Population-averaged # of Cell


0%

5%

10%

15%

20%

25%

30%

35%

0 5 10 15 Latency (Weeks on DOX)

% W

ell

s g

en

era

tin

g G

FP

+

cell

s a

t each

week

a

c

b

NGFP1 NGFP1-p53KD

Gamma Fits

d

Parametric

Distribution

Location

parameter

Shape

parameter

Scale

parameter 2

Gamma N/A 5.1±0.3 16±1 107

Weibull N/A 2.7±0.1 89±2 113

Lognormal 4.3±0 N/A 0.49±0.02 114

Normal 78±1 N/A 30±1 133

Extreme Value 95±2 N/A 30±1 242

0 2 4 6 8 10 12

NGFP1, n=85

NGFP1, n=86

NGFP1, n=93

NGFP1-ctrl hairpin, n=90

NGFP1-Lin28OE, n=84

NGFP1-Lin28OE, n=82

NGFP1-NanogOE, n=88

NGFP1-NanogOE, n=86

NGFP1-p53KD, n=88

NGFP1-p21KD, n=86

0 20 40 60 80 100

NGFP1, n=85

NGFP1, n=86

NGFP1, n=93

NGFP1-ctrl hairpin, n=90

NGFP1-Lin28OE, n=84

NGFP1-Lin28OE, n=82

NGFP1-NanogOE, n=88

NGFP1-NanogOE, n=86

NGFP1-p53KD, n=88

NGFP1-p21KD, n=86

Compared to NGFP1 B cell

parametric gamma function fit

Compared to NGFP1 B cells using

non-parametric log-rank test

Well Type Cd data Time data Cd data

2 2 p-value Increase 2 p-value Increase

NGFP1 CD11b+ cells 1.1 0.38 0.5401 N 1.2 0.272 N

NGFP1-p53KD B cells 89 145 < 0.0001 Y 0.42 0.518 N

NGFP1-p21KD B cells 90 124 < 0.0001 Y 0.011 0.915 N

NGFP1-Lin28OE B cells 52 35 < 0.0001 Y 0.96 0.327 N

NGFP1-NanogOE B cells 300 96 < 0.0001 Y 41 < 0.0001 Y



Supplementary Figure 13 | Parametric analysis of reprogramming events observed in NGFP1 and NGFP1-

p53KD monoclonal populations. a, Using data from Fig. 3c-d, a gamma distribution was used to fit the events observed

(see parametric analysis summary in Supplementary Fig. 12; dashed lines indicate 95% confidence interval). The cumulative

number of wells that became Nanog-GFP+ for both NGFP1 and NGFP1-p53KD B cell derived lines were similar when time is

rescaled by doubling time. The probability density of generating Nanog-GFP+ cells (derivative of the left) is shown in the right

panel. This fit gamma distribution describes the probability of rate-limiting events occurring as a function of cell division

(right) as the monoclonal population grows from a single cell. These events do not necessarily reflect changes at the single cell

level and rather describes changes in the bulk population. b, Identical analysis as described in part a for NGFP1 and NGFP1-

p53KD CD11b+ cell derived lines (using data in Supplementary Fig. 11).

0

20

40

60

80

100

0 50 100 150

Pro

bab

ilit

y o

f

well

s g

en

era

tin

g

Na

no

g-G

FP

+ c

ell

s

Pro

bab

ilit

y o

f

we

lls

ge

ne

rati

ng

Na

no

g-G

FP

+ c

ells

Cu

mu

lati

ve %

Na

no

g-G

FP

+ W

ell

s

=5.4±0.6

=15±2

R2 =0.99

a

NGFP1 NGFP1-p53KD

Cd, Population-averaged # of Cell


Cu

mu

lati

ve %

Na

no

g-G

FP

+ W

ell

s

f (Cd, , ) Clonal CD11b+

cell population Cell population

with iPS cells

=3.8±0.8

=20±5

R2 =0.98

f (Cd, , ) Clonal pre-B

cell population

Cell population

with iPS cells

b

Gamma Fit

NGFP1 NGFP1-p53KD

Gamma Fit







0

20

40

60

80

100

0 50 100 150



NGFP1-p21KD

#1 (w3)

NG

FP

1-p

21K

D #

21

(w

5)

40 XY

NG

FP

1-p

21K

D #

3

(w5

)

N

GF

P1-p

21K

D #

1

(w

3)

40 XY

40 XY

a b

c

NGFP1-p21KD

#1

(w3)

NGFP1-p21KD

#3

(w5)

NGFP1-p21KD

#21 (w5)

Supplementary Figure 14 | Characterization of NGFP1-p21KD B derived iPSC clones. a, Randomly selected

NGFP1-p21KD B cell derived secondary iPSC lines showed normal karyotype when tested at passage 5. Clone # is indicated

for each panel and time on DOX required to derive each line is indicated in parenthesis (w; weeks on DOX). b, Well

differentiated teratomas were derived from all lines tested with evident formation of endoderm, mesoderm and ectoderm

structures. Table on the right summarizes the results of immuno-staining / fluorescent detection of the indicated markers on

the DOX independent iPSC clones derived. Clone # is indicated for each panel and time on DOX required to derive each line

is indicated in parenthesis (w; weeks on DOX). c, Chimera generated following injecting NGFP1-p21KD #1 iPSC clone

derived following 3 weeks on DOX induction in vitro. Chimerism is evident by the agouti coat color.



NG

FP

1-N

OE

#80

(w

5)

NG

FP

1-N

OE

#12

(w4)

NG

FP

1-N

OE

#3

(w4

)

a b

c NGFP1-NOE #3

(w4)

Supplementary Figure 15 | Characterization of NGFP1-NanogOE B cell derived iPSC clones. Clone # is

indicated for each panel and time on DOX required to derive each line is indicated in parenthesis (w; weeks on DOX). a,

Randomly selected NGFP1-NOE derived iPSC lines maintained ES-like morphology and remained Nanog-GFP+

independent of DOX. b, Well differentiated teratomas were derived from all lines tested with evident formation of endoderm,

mesoderm and ectoderm structures. Table on the right summarizes the results of immuno-staining / fluorescent detection of

the indicated markers on the DOX independent iPSC clones derived. Clone # is indicated for each panel and time on DOX

required to derive each line is indicated in parenthesis (W; weeks on DOX). c, Chimera generated following injecting

NGFP1-NOE #3 iPSC clone derived following 4 weeks on DOX induction in vitro. Chimerism is evident by the agouti coat

color.

Phase

contrast

Nanog-

GFP

NG

FP

1-N

OE

#80

(w

5)

NG

FP

1-N

OE

#12

(w4)

NG

FP

1-N

OE

#3

(w4)

NG

FP

1-N

OE

#41

(w

7)



0.00.51.01.52.02.53.03.54.04.55.05.56.06.5

0.00.51.01.52.02.53.03.54.04.55.05.56.06.5

0.00.51.01.52.02.53.03.54.04.55.05.56.06.5

0.00.51.01.52.02.53.03.54.04.55.05.56.06.5

0.00.51.01.52.02.53.03.54.04.55.05.56.06.5

0

10

20

30

40

NGFP1 NGFP1-p53KD NGFP1-p21KDNGFP1-ctrl hairpin NGFP1-Lin28OE

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

NGFP1-NanogOE

0

1

2

3

4

5

6

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

NGFP1 NGFP1-p53KD

Latency (Weeks on DOX)

%G

FP

+ C

ell

s i

n e

ach

we

ll a

t

init

ial G

FP

dete

cti

on

tim

e p

oin

t a b

Nu

mb

er

of

we

lls

in

ea

ch

bin

%GFP+ Cells in each well at initial GFP detection (Labels indicate center of bin)

c

NG

FP1

NGFP

1-ct

rl ha

irpin

NG

FP1-

p53K

D

NG

FP1-

p21K

D

NG

FP1-

Lin2

8OE

NG

FP1-

Nan

ogO

E0

5

10

15

20

1.271 1.515

11.65

1.8311.376

1.813

%G

FP

+ C

ell

s i

n e

ach

we

ll a

t

init

ial G

FP

de

tec

tio

n

Supplementary Figure 16 | Fraction of iPSCs detected for each B cell population at first detection of GFP

signal. a, Percentage of %Nanog-GFP+ cells for each individual well at the week when GFP was initially detected (>0.5%

GFP+). As shown here for NGFP1 and NGFP1-p53KD wells, the %Nanog-GFP+ cells at initial time of detection in a given

well of all populations was independent of the time on DOX treatment. b, The average value of GFP+ fraction in the

populations upon initial detection of a positive signal (defined >0.5% in the study; see Supplementary Fig. 2) is shown here

for all NGFP1 B cell types. The average value is listed and plotted in the center of the box. The whiskers delineate the 95%

confidence interval, while the edges of the box are 25% and 75% percentiles. Comparison indicates a significant difference

between NGFP1 and NGFP1-NanogOE lines (p<0.05), but no significant difference between NGFP1 and the other

populations (p>0.05). We attribute this increase in the NanogOE case to the higher selection of iPSCs possibly resulting from

1) enhanced growth of Nanog overexpressing stem cells (Darr et al. Stem Cells 2006) and 2) a possible delay between entry

into the pluripotent state by the Nanog transgene before the endogenous Nanog-alleles are reactivated (Silva et al. Cell 2009).

c, A histogram of the values shown in part a, as well as for the other NGFP1 B cell types used in this study. Note the change

in bins for the NGFP1-NanogOE wells.



Supplementary Figure 17 | Analytical modeling results with two parameter fits. a & b, Plots of the

experimental and modeling results in Fig. 4d & e respectively, but leaving the population-rescaled average proliferation time,

p, as an extra fit parameter (see Fig. 4a for analytical model details). Given these fitted rates, our conclusions remain

unchanged, as they are based on relative changes between the rates and not absolute values. Error bars indicate 95%

confidence intervals.

a

k,

Intr

insic

rate

per

cell

div

isio

n

[10

-8 C

d-1

]

k,

Intr

insic

rate

per

week

[10

-7 w

eeks

-1]

Cum

ula

tive f

raction o

f

Nanog-G

FP

+ W

ells

b

Population Rescaled Time, [106 weeks]

Cum

ula

tive f

raction o

f

Nanog-G

FP

+ W

ells

Population Rescaled Cell Divisions, /td [107 Cd]

0.89

0.53 0.58 0.37

0.49

0.0

0.5

1.0

0.90 0.91 1.0

0.47 0.42

0.0

0.5

1.0



Supplementary Figure 18 | Computational simulation of reprogramming. To take into account factors like

stochasticity in cell division times, fluctuations in the number of cells and potential loss of iPSCs during culturing in addition

to the stochasticity in the cell-intrinsic reprogramming process, we implemented a detailed computer simulation of each

experiment in this study. Comparison of the simulations and experimental data are shown in a in terms of population rescaled

time, and in b in terms of population rescaled cell divisions. Given these fitted rates, our conclusions remain unchanged, as

they are based on relative changes between the rates and not absolute values. Error bars indicate 95% confidence intervals.

a

k,

Intr

insic

rate

per

cell

div

isio

n

[10

-8 C

d-1

]

k,

Intr

insic

rate

per

week

[10

-7 w

eeks

-1]

b

Population Rescaled Time, [106 weeks]

Population Rescaled Cell Divisions, /td [107 Cd]

2.5 2.4 2.8

1.0 0.9 0

1

2

3

2.5

1.4 1.6 0.8 1.0

0

1

2

3 Cum

ula

tive f

raction o

f

Nanog-G

FP

+ W

ells

Cum

ula

tive f

raction o

f

Nanog-G

FP

+ W

ells



Supplementary Figure 19 | Summary of two distinct modes of accelerating reprogramming. Nearly all

secondary somatic donor cells in our system can give rise to iPSCs upon continued proliferation and induction of the four

reprogramming factors in vitro. Although the majority of clonal populations eventually produce iPSCs, the latency is not

constant for each clonal population. The top x-axis represents the average latency time during factor expression, while the

bottom x-axis is the average number of cell divisions during latency. The y-axes represent the effective number of cells

maintained during the experiment (see Fig. 4). Under our experimental conditions, Neff, remained nearly constant over the

course of the experiment, and therefore population rescaled time, , is simply the Nefft, the area of the square. In order to

describe the population rescaled number of cell divisions during latency, the time axis is rescaled by the doubling time, td. The

intrinsic rate of reprogramming is the inverse of the average . Blue scheme (left) summarizes the NGFP1 case. Red scheme

(middle) represents a cell doubling time dependent scenario of accelerating the reprogramming process. As exemplified in p53/

p21 pathway inhibition experiments, reprogramming kinetics are accelerated in a manner directly proportional to the

augmentation of cell population doubling time, however the average number of cell divisions required to achieve full

reprogramming is not influenced (still Cd,0). Orange scheme (right) represents a cell doubling time independent scenario (e.g.,

Nanog overexpression) of accelerating the reprogramming process where the occurrence of required unknown secondary

stochastic events is enhanced and achieved within a fewer number of cell divisions on average (cell divisions << Cd,0). Notably,

the different modes need not to be mutually exclusive as certain perturbations could enhance or inhibit reprogramming via both

cell proliferation dependent and independent effects.

Average Cell Divisions on DOX before iPSC formation, Cd

Cell-division

dependent

<

~

Eff

ecti

ve

Po

pu

lati

on

Siz

e, N

eff

t1

Cd, 1

t0

Cd, 0

Cell-division rate

independent

<

< t1

Cd, 2

t0

Cd, 0

Acceleration Reprogramming to pluripotency

with defined factors

Average Time on DOX before iPSC formation, t

Eff

ecti

ve

Po

pu

lati

on

Siz

e, N

eff

0 /td,0 1 /td,1 2 /td,2

0 1 2

k = 1Area

Area, = Neff t

Area = Neff C d = td



a b c d e - mitweb.mit.edu/biophysics/papers/nature2009c_si.pdf · only the c-myc transgene and...

Documents