0 0   M o n t h   2 0 1 7   |   V o L   0 0 0   |   n A t U R E   |   1

LEttER

doi:10.1038/nature21706

The lung is a site of platelet biogenesis and  

a reservoir for haematopoietic progenitors

Emma Lefrançais

1

*, Guadalupe ortiz-Muñoz

1

*, Axelle Caudrillier

1

, Beñat Mallavia

1

, Fengchun Liu

1

, David M. Sayah

2

Emily E. thornton

3

, Mark B. headley

3

, tovo David

4

, Shaun R. Coughlin

4

, Matthew F. Krummel

3

, Andrew D. Leavitt

1

Emmanuelle Passegué

1

 & Mark R. Looney

1,5

Platelets are critical for haemostasis, thrombosis, and inflammatory 

responses

1,2

, but the events that lead to mature platelet production 

remain incompletely understood

3

. The bone marrow has been 

proposed to be a major site of platelet production, although there 

is indirect evidence that the lungs might also contribute to platelet 

biogenesis

4–7

. Here, by directly imaging the lung microcirculation 

in mice

8

, we show that a large number of megakaryocytes circulate 

through the lungs, where they dynamically release platelets. 

Megakaryocytes that release platelets in the lungs originate from 

extrapulmonary sites such as the bone marrow; we observed 

large megakaryocytes migrating out of the bone marrow space. 

The contribution of the lungs to platelet biogenesis is substantial, 

accounting for approximately 50% of total platelet production or  

10 million platelets per hour. Furthermore, we identified populations 

of mature and immature megakaryocytes along with haematopoietic 

progenitors in the extravascular spaces of the lungs. Under conditions 

of thrombocytopenia and relative stem cell deficiency in the bone 

marrow

9

, these progenitors can migrate out of the lungs, repopulate 

the bone marrow, completely reconstitute blood platelet counts, and 

contribute to multiple haematopoietic lineages. These results identify 

the lungs as a primary site of terminal platelet production and an 

organ with considerable haematopoietic potential.

Platelets are released from megakaryocytes; however, even 

though they were discovered in the nineteenth century, we do 

not completely understand the mechanisms by which platelets 

are produced. On the basis of previous work showing the presence 

of megakaryocytes in the lungs

10

 and demonstrating that blood  

leaving the lungs contains more platelets and fewer megakaryocytes 

than blood entering the lungs

4,11

, we hypothesized that the lungs 

could have a major role in platelet biogenesis, and directly investi-

gated this process using 2-photon intravital microscopy (2PIVM) 

of the lungs and fluorescent reporter mouse strains. We used  

PF4-Cre ×  Gt(ROSA)26Sor

tm4(ACTB-tdTomato,-EGFP)Luo

 (mTmG) (hereafter 

called PF4-mTmG) reporter mice, in which PF4-Cre

12

 drives mem-

brane GFP expression in megakaryocytes and platelets, while all other 

cells are labelled with membrane tomato. We observed large circulating 

GFP

+

 cells that passed through the lung microcirculation, where they 

produced GFP

+

 extensions in a flow-dependent manner (Fig. 1a, b  

and Supplementary Video 1). These events resembled proplatelet 

and preplatelet formation from cultured megakaryocytes

3,13,14

. In the 

lungs, the duration of these events varied from approximately 20 to 

60 min (Fig. 1a, b and Supplementary Video 1). Many of the GFP

+

 

cells contained large nuclei (more than 10 μ m), which appeared as 

unlabelled dark holes that remained intact during this process (Fig. 1b  

and Supplementary Video 2) and resulted in naked intravascular nuclei 

after platelets were released (Supplementary Video 2). We confirmed 

that we labelled large mobile nucleated cells by imaging the lung 

microcirculation of PF4-Cre ×  Gt(ROSA)26Sor

tm1(CAG-tdTomato

*,-EGFP*)Ees

 

(nTnG) (hereafter called PF4-nTnG) reporter mice, in which a fluores-

cence switch allows GFP

+

 nuclei to be tracked (Extended Data Fig. 1a  

and Supplementary Video 3).

We next quantified the GFP

+

 megakaryocytes and proplate-

lets in PF4-mTmG lungs by assigning surface volumes (Fig. 1c an

Supplementary Video 4). The putative megakaryocytes (large GFP

+

 

cells undergoing platelet release) had median volumes of 10,000 μ m

3

 

and diameters of more than 25 μ m  (Fig. 1d, e), whereas the putative 

platelets (small circulating GFP

+

 events) had median volumes of below 

10 μ m

3

 and diameters of 2–3 μ m  (Fig. 1d, e). These values are consistent 

with previous estimates of megakaryocytes and platelet sizes

3

. For each 

large GFP

+

 cell undergoing platelet release, we calculated the number 

of platelets that could be liberated into the lung circulation, and this 

ranged from fewer than 500 platelets for small megakaryocytes or pro-

platelets to more than 1,000 platelets for larger megakaryocytes (Fig. 1f),  

with a median of around 500 platelets per megakaryocyte. Previous studies  

have produced widely varying estimates of the number of platelets  

produced from a single megakaryocyte (200–10,000 platelets)

15–17

. Our 

method uses direct measurement for each event, and therefore is likely 

to yield more accurate estimates. In total, we analysed 20 h of footage 

from 10 mice, and observed an average of 2.2 ±  0.26  (n =  10)  mega-

karyocytes per hour in an imaged lung volume of 0.07 mm

3

 (Fig. 1g  

and Supplementary Video 5). When extrapolated to the entire lung 

volume, this equals more than 10 million platelets produced per hour 

from the lungs (Fig. 1h, Methods and Extended Data Table 1). Overall, 

when adjusted for platelet lifespan and splenic sequestration, we  

estimate that the lung is responsible for approximately 50% of total platelet 

production in the mouse (Fig. 1i, Methods and Extended Data Table 1).  

Blood platelet counts were unchanged after 2PIVM (Extended Data 

Fig. 1b). Platelet production by the lungs is also biologically tunable,  

as the administration of the megakaryocyte growth factor thrombopoi-

etin (TPO) increased blood platelets threefold (Fig. 1j) and the number  

of megakaryocytes undergoing proplatelet formation observed per 

hour twofold (3.9 ±  0.38,  n =  9)  (Fig. 1k). We conclude from these 

experiments that the lungs are a primary site of platelet biogenesis.

To investigate the origin of the intravascular megakaryocytes and 

proplatelets in the lungs, we adoptively transferred lung resident cells 

using the orthotopic single-lung transplant model in mice

18

. We trans-

planted a lung from an mTmG mouse (with no Cre or GFP expression)  

into a PF4-mTmG recipient mouse and vice versa (Extended Data  

Fig. 1c–e and Supplementary Video 6). Using 2PIVM, we observed 

proplatelet formation from GFP

+

 megakaryocytes in the lung vascu-

lature following transplantation of mTmG lungs to PF4-mTmG mice, 

but not following the reverse transplant (PF4-mTmG lungs to mTmG 

mice). These experiments confirmed that megakaryocytes releas-

ing platelets in the lung circulation originate from outside the lungs.  

1

Department of Medicine, University of California, San Francisco (UCSF), San Francisco, California 94143, USA. 

2

Department of Medicine, University of California, Los Angeles (UCLA), Los Angeles, 

California 90095, USA. 

3

Department of Pathology, University of California, San Francisco (UCSF), San Francisco, California 94143, USA. 

4

Cardiovascular Research Institute, University of California, 

San Francisco (UCSF), San Francisco, California 94143, USA. 

5

Department of Laboratory Medicine, University of California, San Francisco (UCSF), San Francisco, California 94143, USA.

* These authors contributed equally to this work.

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

2   |   n A t U R E   |   V o L   0 0 0   |   0 0   M o n t h   2 0 1 7

letter

reSeArCH

We hypothesized that the bone marrow

19

, spleen, or liver could be 

the source of these intravascular megakaryocytes and proplatelets. 

We imaged the calvarial bone marrow (Extended Data Fig. 1f) and 

the spleen in PF4-mTmG mice and observed extravascular megakar-

yocytes releasing proplatelets into the sinusoids of the bone marrow 

(Fig. 1l, Extended Data Fig. 1g and Supplementary Video 7) and spleen 

(Extended Data Fig. 1i and Supplementary Video 9). We also observed 

large megakaryocytes exiting the bone marrow space (Fig. 1m and 

Supplementary Video 8). In contrast to observations in the lung, we 

did not observe any intravascular megakaryocytes undergoing pro-

platelet formation. We did not observe any megakaryocytes in the liver 

(Extended Data Fig. 1h).

In addition to intravascular megakaryocytes, we also observed  

large cells in the perivascular lung interstitium in PF4-Cre ×   

Gt(ROSA)26Sor

tm9(CAG-tdTomato)Hze

 (ROSA26-tdTomato) (hereafter 

called PF4-tomato) and PF4-mTmG mice (Fig. 2a, bExtended Data 

Fig. 2a and Supplementary Video 10), and in mTmG mice that had 

received PF4-mTmG lung transplants (Extended Data Fig. 2b). These 

extravascular cells were sessile during our imaging (up to 4 h) and con-

tained large nuclei (Fig. 2c). Although they were comparatively large, 

these extravascular cells were on average one-third of the volume and 

approximately half of the diameter of intravascular megakaryocytes 

(Fig. 2d) and also smaller than resident megakaryocytes in the bone 

marrow and spleen (Extended Data Fig. 2c–e). Using image analysis, 

we detected around 2,000 PF4-tomato cells per cubic millimetre of 

lung tissue or more than 1 million cells per lung (Fig. 2e). The nuclear 

diameters of these cells were significantly larger than those of non-

GFP

+

 cells (Fig. 2f) and the nuclei had more complex shapes (Fig. 2g). 

We used a method of intravascular labelling before lung digestion

20

 to 

determine the relative proportions of intravascular and extravascular 

megakaryocytes (Fig. 2i) and found that 85% of PF4-tomato events 

were extravascular and 15% intravascular (Fig. 2j).

To further characterize lung megakaryocytes, we sorted PF4-tomato

+

 

and CD41

+

 cells from perfused and digested lungs and stained them 

with the megakaryocyte and platelet marker von Willebrand factor 

(vWF). The large, PF4-tomato

+

 CD41

+

 cells with complex nuclei also 

stained positive for vWF in a granular pattern, which is consistent with 

megakaryocytes (Fig. 2h). To avoid cell aggregation with PF4-tomato

+

 

platelets during flow cytometry staining, we prepared digested lungs 

from PF4-nTnG mice, in which GFP is targeted to the cell nucleus and 

thus does not stain anucleate platelets. We gated on nGFP

+

 CD41

+

 

PF4-mTmG BM 2PIVM 

0

2

4

6

Platelet diameter 

m) 

m

3

S M L All

0

500

1,000

1,500

2,000

2,500

No. platelets per MK

GFP size

Image analysis 

50 

μm

c

i

0

10

20

30

40

50

MK volume 

 

(×10

3

 μ

m

3

)

0

20

40

60

80

Platelet volume 

 

0

10

20

30

40

MK diameter 

m)

0

1

2

3

4

Lung MK frag 

 

observed per hour 

0

5

10

15

20

Lung platelets 

 

10

6

 per hour 

0

20

40

60

80

100

Lung platelet

 

 production (%) 

– TPO

0

1,000

2,000

3,000

4,000

****

Blood platelets 

 

(10

6

 per ml)

– TPO

0

2

4

6

**

Lung MK frag 

 

observed per hour 

a

PF4-mTmG lung 2PIVM 

0 min 

20 min 

34 min 

40 min 

0 min 

4 min 

13 min 

19 min

30 

μm

PF4cre mGFP

Rosa mTom

b

40 

μm

PF4-mTmG BM 2PIVM 

0 min 

2 min 

21 min 

0 min 

23 min

33 min 

50 

μm

Figure 1 | The lungs are an important site of megakaryocyte circulation 

and platelet production. ac, Visualization of megakaryocytes and 

platelet production in the lung circulation by 2PIVM in PF4-mTmG  

mice. ab, Sequential images show a large megakaryocyte (green) in the 

lung capillaries (red) where it undergoes proplatelet formation (arrows).  

b, Dark hole in the cytoplasm (dashed outline) indicates the nucleus.  

Time elapsed is indicated. cf, Characterization of PF4

+

 events by  

image analysis. c, Representative image of surface analysis of the GFP 

channel. de, Volume distribution (d) and equivalent diameter (e) of 

megakaryocytes (MKs, n =  35) and platelets (n =   492).  f, Number of 

platelets produced by one megakaryocyte according to its size: small  

(< 500  platelets,  n =  18), medium (500–1,000 platelets, n =  7) and large 

(> 1,000  platelets,  n =   10).  df, Minimum-to-maximum boxplots are 

presented. gi, Quantification of lung platelet production. g, Number 

of megakaryocytes releasing platelets observed per hour in imaged lung 

volume (2-h movies, n =   10).  hi, Estimation of the number (h) and the 

percentage (i) of platelets produced by the lung. jk, Platelet counts in  

the blood (j) and number of megakaryocytes releasing platelets in the 

lungs (k) 5 days after TPO treatment. n ≥  5 mice per group. Unpaired  

t-test: 

* * * * P <  0.0001,  * * P <  0.005.  gk, Mean ±  s.d. are presented.  

lm, Visualization of proplatelet release (arrow) and megakaryocyte 

migration (circled) in bone marrow (BM) sinusoids by 2PIVM in  

PF4-mTmG mice.

0.0

0.1

0.2

0

5

10

15

Volume (x10

3

 μ

m

3

0

10

20

30

40

Diameter 

m) 

0

1,000

2,000

3,000

4,000

Cells per mm

3

0

1

2

Cells per lung (10

6

)

nTom

+

nGFP

+

nTom

+

nGFP

+

4

6

8

10

12

****

Nuclear diameter 

 (μ

m) 

0.4

0.6

0.8

1.0

****

Circularity 

nGFP

+

 cells (10

6

)

nGFP

nTom

Live lung cells 

nGFP

0.39% 

CD41 

nGFP

CD41

19% 

c-Mpl

GPVI 

Count

 

F4/80 

Gated on nGFP

+

 CD41

PF4-nTnG lungs

i

Intravascular 

CD41-APC

5 min

Lung 

digestion 

CD41–FITC 

staining

PF4-tomato lungs

Hoescht Tom 

PF4

+

 11% 

SSC

 

CD41-FITC 

CD41

23% 

Hoescht

CD41-APC

IV 

e.v. 

85% 

i.v.

12% 

WT lung

Hoescht Tom 

Live lung cells 

PF4 tomato

PF4 tomato

+

, CD41

Live lung cells 

Lu

ng

s

BM

Bloo

d

Lu

ng

s

BM

Bloo

d

0

5

10

15

nGFP

+

/CD41

+

 cells (10

4

)

m

g

20 

μm

DAPI 

vWF

PF4-tomato

Lung sorted
PF4-tomato

CD41

+

 cells

h

PF4-nTnG

PF4-tomato lung 2PIVM 

a

PF4-mTmG

PF4-nTnG

PF4cre mGFP

Rosa mTom

PF4cre nGFP

Rosa nTom

50 

μm

15 

μm

FITC dextran 

PF4-tomato 

25 

μm

50 

μm

FITC dextran 

PF4-tomato 

e.v

.

i.v.

0

20

40

60

80

100

Percentage (%)

Figure 2 | Resident megakaryocytes are present in the extravascular 

spaces of the lung. ac, Visualization of resident or static megakaryocytes 

in the lung by 2PIVM of PF4-tomato (a), PF4-mTmG (b), and PF4-

nTnG mice (c). (d) Size characterization of PF4

+

 cells (red, > 10  μ m)  by 

quantitative image analysis of PF4-tomato lungs (n =   312).  Minimum-

to-maximum boxplots are presented. e, Quantification of PF4

+

 cells (red, 

>

 10  μ m)  (n =   6).  fg, Comparison of nuclear size (f) and circularity (g

between PF4

+

 cells (green, n =  17) and all other lung cells (red, n =   91)   

by quantitative image analysis of PF4-nTnG lungs. Unpaired t-test:  

* * * * P <  0.0001.  h, Representative immunofluorescence images of PF4

+

 

and CD41

+

 cells sorted from perfused PF4-tomato lungs and stained 

with anti-vWF antibodies (green) and 4′ ,6-diamidino-2-phenylindole 

(DAPI, blue). ij, Intravascular (i.v.) or extravascular (e.v.) localization of 

PF4

+

 and CD41–FITC

+

 cells. i, Experimental schema and representative 

fluorescence-activated cell sorting (FACS) plots. j, Percentage of cells 

located intravascularly or extravascularly (mean of four experiments, n =   8 

mice). k, FACS gating strategy and surface expression of nucleated PF4

+

 

and CD41

+

 cells from lungs of PF4-nTnG mice. lm, FACS quantification 

of nucleated PF4

+

 (l) and nucleated PF4

+

/CD41

+

 (m) cells in PF4-nTnG  

whole lung (n =  28), bone marrow (BM, n =  24) (two femurs, two 

tibias ×  6.6) and blood (1.5 ml, n =   8).  e–glm, Mean ±   s.d.  presented.

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

0 0   M o n t h   2 0 1 7   |   V o L   0 0 0   |   n A t U R E   |   3

letter

reSeArCH

events as the putative lung megakaryocyte pool (Fig. 2k and Extended 

Data Fig. 3a) and this population stained for the megakaryocyte and 

platelet-specific markers glycoprotein VI (GP VI) and the TPO receptor 

c-Mpl (Fig. 2k), but did not stain for markers of other lung resident cells, 

such as F4/80 for macrophages (Fig. 2k and Extended Data Fig. 3a–d)  

confirming that these lung-derived cells were megakaryocytes. The 

majority of the nGFP

+

 cells in the lung were CD41

 (Extended Data 

Fig. 3b), but both CD41

+

 and CD41

 cells co-stained for GPVI and 

c-Mpl, confirming their megakaryocyte lineage (Extended Data  

Fig. 3c). The nGFP

+

 CD41

 cells had a more immature profile with 

lower CD61 and CD42b expression, a smaller size and lower DNA con-

tent than the nGFP

+

 CD41

+

 cells (Extended Data Fig. 3e–i). Overall, 

lung megakaryocytes were more immature than bone marrow mega-

karyocytes (Extended Data Fig. 3j–m), but the total number of meg-

akaryocytes in the lungs was comparable to that in the bone marrow 

(Fig. 2l, m). We also used RNA-sequencing (RNA-seq) to characterize 

lung and bone marrow megakaryocytes (nGFP

+

 CD41

+

 cells) in PF4-

nTnG mice. We identified more than 700 genes that were expressed 

differentially in the lungs and bone marrow (Extended Data Fig. 4a 

and Supplementary Table 1); many megakaryocyte and platelet path-

ways were represented in both groups, but there was less expression 

of mature megakaryocyte markers in the lung group (Extended Data  

Fig. 4b, c), consistent with our profiling by immunostaining.

Lung 2PIVM indicated that only intravascular megakaryocytes 

released platelets. To test the function of extravascular megakaryocytes 

in the lungs, we used the orthotopic single-lung transplant model to 

adoptively transfer lung-resident cells. PF4-tomato donor lungs were 

extensively perfused and the left lungs were immediately transplanted 

into wild-type or c-mpl

/

 thrombocytopenic recipient mice (Fig. 3a).  

These transplanted mice were injected with TPO at 3 and 40 days 

post-transplantation and bled weekly to track the number of platelets 

(Fig. 3a, b). Peripheral blood tomato

+

 events after lung transplantation 

are, by definition, of donor lung origin. In wild-type recipients, we 

detected low-level (1–2% of total CD41

+

 events) and transient produc-

tion of tomato

+

 events (Fig. 3d and Extended Data Fig. 5i). However, 

in the majority (70%) of c-mpl

/

 recipients, we detected large and sus-

tained (90 days) production of tomato

+

 events (Fig. 3c, d and Extended 

Data Fig. 5c, d) that fully reconstituted platelet counts (Fig. 3b an

Extended Data Fig. 5g, h). We observed the same response in two out 

of five c-mpl

/

 recipients not treated with TPO after transplantation 

(Extended Data Fig. 5a, b, e, f). The tomato

+

 CD41

+

 events were also 

positive for CD42, GPVI and c-Mpl (Extended Data Fig. 6a, b) and 

expressed CD62P when stimulated with thrombin (Extended Data  

Fig. 6c–e), confirming that they were platelets.

In selected experiments, c-mpl

/

 lung transplant recipients were 

followed for up to 10 months after transplantation. In these mice, we 

observed sustained production of tomato

+

 platelets and sustained 

reconstitution of platelet counts (Extended Data Fig. 5j, k). Because 

platelet lifespan in mice is 3–5 days

21

, the persistence of donor-origin  

platelets for more than 3 months suggested that the transplanted lungs 

contained a progenitor population capable of long-term reconstitu-

tion of mature megakaryocytes and platelets. Indeed, the fact that the 

extravascular megakaryocytes were smaller than the intravascular meg-

akaryocytes in the lungs and the extravascular megakaryocytes in the 

bone marrow and spleen could point to the presence of megakaryocyte 

progenitors.

We imaged lungs 3 months after transplantation and confirmed 

the persistence of PF4-tomato cells (Extended Data Fig. 7a). We also 

detected the presence of tomato

+

 CD41

+

 cells in the bone marrow 

of c-mpl

/

 mice that had received PF4-tomato lung transplants 

using flow cytometry (Fig. 3e) and immunofluorescence (Fig. 3f an

Extended Data Fig. 7b). To test for lung megakaryocyte progenitors and 

to track donor cells in recipient mice, we transplanted mTmG lungs, in 

which all cells and platelets are tomato

+

, into c-mpl

/

 mice (Extended 

Data Fig. 7c–e). We next quantified the megakaryocyte progenitor 

(MkP) population in the bone marrow of c-mpl

/

 mice transplanted 

with mTmG lungs by staining myeloid progenitors (Lin

 Sca-1

 c-Kit

+

for CD41 and CD150

22

 (Fig. 3g). We found more myeloid progenitors 

and MkPs in the bone marrow of the lung transplant recipients than in 

c-mpl

/

 bone marrow; the numbers in the transplant recipients were 

similar to the numbers of myeloid progenitors and MkPs normally 

found in the bone marrow of wild-type (mTmG) mice (Fig. 3h an

Extended Data Fig. 7f, g). One-third of the MkPs in the bone marrow 

of c-mpl

/

 mice transplanted with mTmG lungs expressed tomato 

(Fig. 3i, j).

We next tested whether the haematopoietic stem cell (HSC) defi-

ciency characteristic of c-mpl

/

 bone marrow

9

 could be reversed by 

lung transplantation. We gated on the bone marrow LSK (Lin

 Sca-1

+

 

Time (days) 

0

30

60

90

0

20

40

60

80

n = 7/10  

70% 

n = 3/10  

30% 

Tomat

o

+

 platelets (%) 

0

30

60

90

0

500

1,000

Blood platelet

s

(10

6

 per ml)

Time (days) 

Donor mTmG 

CD150 

MPP3/4 

76% 

MPP2 

3% 

LT 6% 

ST 15% 

93% 

1% 

3% 

71% 

9% 

12% 

 7% 

Recipient c-mpl

–/–

mTmG

c-mpl

–/–

3% 

CD48

 

BM LSK (Lin

 Sca-1

+

 c-Kit

+

)  

BM MP (Lin

 Sca-1

 c-Kit

+

CD41 

CD150 

Recipient c-mpl

–/–

  mTmG c-mpl

–/–

Donor mTmG 

MkP 3% 

MkP 1.5% 

MkP 2.8 % 

Lung transplant experiments : mTmG   c-mpl

–/–

Lung transplant experiments: PF4-tomato    c-mpl

–/–

CD41 

Tomato 

Tom

+

29% 

BM MkP

Recipient c-mpl

–/–

mTmG

c-mpl

–/–

Tomato 

Tom

+

45% 

BM LT 

Recipient c-mpl

–/–

 

mTmG

c-mpl

–/–

CD45

 

PF4-tomato 

donor   

c-mpl

–/–

recipient 

Perfused 

donor 

lungs

 

Tracking of Tom+ 

platelets

 

TPO 250 mg kg

–1

 

Days 3 and 40 

Day 7 

Day 30 

Day 90 

CD41 

Tomato (PF4) 

Tom

2% 

Tom

21% 

Tom

72% 

Platelets (FSC

low

/CD41

+

D R Tx

0

20

40

60

80

100

** *

Tom

+

 cells (% of LT) 

D R Tx

0

20

40

60

80

100

**

*

Tom

+

 cells (% of MkP) 

D R Tx

0

1

2

3

4

5

**

*

n.s.

MkP

 (% of MP) 

D R Tx

0

5

10

15

20

**

*

n.s.

LT (% of LSK) 

PF4-tom

c-mpl

–/–

(sustained response) 

PF4-tom

WT 

c-mpl

–/–

c-mpl

–/–

PF4-tom

c-mpl

–/–

(transient response) 

D R Tx

0

0.1

0.2

0.3

0.4

0.5

**

**

To

m

+

 CD4

1

+

 cells

(per cent of Lin

)

20 

μm

Tomato 

CD41-FITC 

BM cells

 

Figure 3 | Lung-derived progenitors reconstitute platelet counts  

and haematopoietic stem cell deficiency in thrombocytopenic mice.  

af, Transplantation of PF4-tomato lungs to c-mpl

/

 mice. a, Experimental 

schema. b, Blood platelet counts (n =  4–6 mice per group). cd, Percentage 

of donor-derived platelets analysed by counting tomato

+

 events in the 

CD41

+

 FSC

low

 gate. bd, Mean ±   s.e.m.  presented.  ef, Bone marrow 

cells from donor (PF4-tom), recipient (c-mpl

/

) or transplanted mice 

with 10 months sustained donor-derived platelet production (Lung Tx) 

were analysed. e, FACS analysis of bone marrow cells reveals Tomato

+

 

cells (CD41

+

 and CD41

 populations). Percentage of lineage-negative 

bone marrow cells positive for Tomato and CD41. f, Representative 

immunofluorescence image of Tomato

+

 cell (red) in the bone marrow of 

a transplanted mouse stained with anti-CD41 (green) and Syto60 nucleic 

acid stain (blue). gn, Transplantation of mTmG lungs into c-mpl

/

 mice. 

Bone marrow cells from donor (D, mTmG), recipient (R, c-mpl

/

), or 

transplanted (Tx) mice with sustained donor-derived platelet production 

(3 months) were analysed. g, Representative FACS analysis of the myeloid 

progenitor compartment and the MkP population. h, Percentage of  

the MkP population within the myeloid progenitor compartment.  

ij, Percentage of donor origin Tomato

+

 cells in the MkP population.  

k, Representative FACS analysis of the LSK compartment showing MPP2, 

MPP3/4, ST-HSC and LT-HSC population frequencies within the LSK 

compartment. l, Percentage of the LT-HSC population within the LSK 

compartment. mn, Percentage of donor origin Tomato

+

 cells in the  

LT-HSC population. hjln, Mean ±  s.d. are presented (n =   2–4  mice   

per group). Unpaired t-test: 

* * P <  0.01,  * P <  0.05; n.s., not significant.

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

4   |   n A t U R E   |   V o L   0 0 0   |   0 0   M o n t h   2 0 1 7

letter

reSeArCH

c-Kit

+

) population and probed for the following subsets: long-term 

HSCs (LT-HSCs; CD48

 CD150

+

), short-term HSCs (ST-HSCs; 

CD48

 CD150

), multipotent progenitor 2 (MPP2) cells (CD48

+

 

CD150

+

), and MPP3/4 cells (CD48

+

 CD150

)

23

 (Fig. 3k). Lung trans-

plantation from TPO-competent (mTmG) donors reversed the LSK 

population deficiencies (Fig. 3l anExtended Data Fig. 7h–l). Lung–

derived MkPs, LT-HSCs, ST-HSCs, MPP2s and MPP3/4s were also 

found in the bone marrow, spleen and recipient lungs (Fig. 3m, n and 

Extended Data Fig. 7n–p). A non-specific post-lung transplantation 

response was ruled out because transplantation of c-mpl

/

 lungs into 

c-mpl

/

 mice did not produce increased platelet counts (Fig. 3b) or 

alter the bone marrow composition (Extended Data Fig. 7m). Together, 

these results suggest that a haematopoietic progenitor population resid-

ing in the lungs can migrate to the bone marrow and reverse HSC 

defects and associated cytopenias.

We tested for haematopoietic progenitors in dispersed wild-type 

lungs using similar gating on live CD45

+

 Lin

 Sca-1

+

 c-Kit

+

 cells as 

in the bone marrow. We discovered that the lungs contain an array of 

haematopoietic progenitors, including ST-HSCs, MPP2s, MPP3/4s, 

myeloid progenitors and MkPs (Fig. 4a, b), which were morphologically 

indistinguishable from bone marrow LSK cells (Fig. 4e). These cells 

were present at lower numbers than in the bone marrow (Fig. 4c, d) and 

spleen (Extended Data Fig. 8a, b), except that there were more ST-HSCs 

in the lungs than in the spleen. These cells were extravascular, because 

they were not removed by perfusion and were not stained by intravas-

cular CD45-APC antibodies (Extended Data Fig. 8c–h). To our knowl-

edge, this is the first description of haematopoietic progenitor cells in 

the adult lungs, and we reasoned that these cells could be the source 

of the reconstituting effects of lung transplantation. To test this hypo-

thesis, we isolated LSK and ST-HSCs from perfused wild-type lungs  

(and the bone marrow for comparison), injected these cells intrave-

nously into c-mpl

/

 recipients, and tested for the presence of c-Mpl

+

 

platelets in the peripheral blood (Fig. 4f). Injection of lung LSK cells 

and ST-HSCs increased peripheral c-Mpl

+

 platelets and partially 

restored platelet counts, and injection of bone marrow cells had a  

similar effect (Fig. 4g–i). These results show that the adult lungs contain  

functional haematopoietic precursors capable of migration, bone  

marrow engraftment, and reconstitution of haematopoietic defects.

Finally, we tested whether lung haematopoietic progenitors are  

capable of multi-lineage bone marrow reconstitution. We transplanted 

lungs from mTmG donors to allow us to track mature lineages in the 

peripheral blood and bone marrow (Extended Data Fig. 7c), and 

detected sustained production of donor-derived (tomato

+

 CD41

cells in the peripheral blood of the c-mpl

/

 recipient mice (Extended 

Data Fig. 8i, j). The donor-derived cells included platelets (Extended 

Data Fig. 7d, e), neutrophils, and B and T cells (Extended Data  

Fig. 8k). Considering that there are no neutrophil, B cell or T cell defects  

in c-mpl

/

 mice and therefore no impetus for donor-derived recon-

stitution, these results demonstrate an important contribution of the 

lungs to overall haematopoiesis.

Our results provide direct evidence that the lungs are a major site of 

platelet biogenesis, which involves a distinct mechanism of proplatelet 

release from intravascular megakaryocytes (of extrapulmonary origin) 

in the lung microcirculation (Extended Data Fig. 9a). These results 

open new lines of investigation to improve our approach to treating 

thrombocytopenia, which affects millions of patients worldwide and 

causes substantial morbidity and mortality. We propose that the lungs 

are an ideal bioreactor for the production of mature platelets from meg-

akaryocytes, and could advance studies of the treatment of thrombocy-

topenia with cell-based therapies

16

. Beyond the mechanical forces that 

MP

P3/

4

MP

P2 ST

LT

MkP

MPP3

/4

M

PP2 S

T

LT

Mk

P

0

1,000

2,000

3,000

10,000

20,000

30,000

0

10,000

20,000

30,000

60,000

90,000

c-mp

l

Lu

ng

LS

K

BM

LS

K

Lu

ng

ST

BM

ST

0

200

400

600

***

***

**

**

c-mpl

Lu

ng

LS

K

BM

LS

K

Lu

ng

ST

BM

ST

0

5

10

15

20

****

***

**

**

i

Cells per BM 

Cells per lung 

c-mp

l

+

 platelets (%)

Blood platelets (10

6

 per ml) 

 

BM 

Lung 

Sorted LSK

Donor

c-mpl

–/–

recipient

Perfused 

lungs

Tracking of c-Mpl

+

platelets

TPO 250 mg kg

–1

 

Days 3 and 40 

FACS sorted     

LSK and ST 

i.v. injection 

2000 LSK  
or 150 ST 

10 

μm

MkP

FSC 

Lin/PI 

CD41 

CD150 

SSC 

CD45 

C-Kit

 

Sca-1 

CD48 

CD150 

Bone marrow

Single cells 

Lin/PI

CD45

+

LSK 

MP 

MP 

LSK 

MPP3/4  MPP2 

LT 

ST 

CD45

Lin/PI

MkP

FSC 

Lin/PI 

CD41 

CD150 

SSC 

CD45 

C-Kit

 

Sca-1 

CD48 

CD150 

Lung

Single cells 

Lin/PI

CD45

LSK 

MP 

MP  LSK  

MPP2 

LT 

ST 

MPP3/4 

CD45

+

Lin/PI

MkP

c-Mpl

Lung LSK 

Lung ST 

c-mpl

–/–

FSC 

CD41 

Blood cells

Platelets

c-mpl

2% 

c-mpl

9% 

c-mpl

11% 

Figure 4 | The lung contains haematopoietic progenitors, including 

megakaryocyte progenitors. ac, Representative lung (a) and bone 

marrow (c) FACS plots of haematopoietic progenitors within the LSK  

and myeloid progenitor compartments. bd, Cell counts of haematopoietic 

progenitor (MPP2, MPP3/4, ST-HSC, LT-HSC, and MkP) populations  

in whole lungs (n =   12)  (b) and bone marrow (n =   5)  (legs,  d).  

e, Representative image of Wright-Giemsa staining of LSK cells sorted 

from bone marrow or lung. f, Experimental schema. gh, Percentage of 

donor-derived platelets analysed by FACS counting of c-Mpl

+

 events in  

the CD41

+

 FSC

small

 gate. i, Blood platelet counts. hi, Mean ±   s.d. 

presented (n =  4–5 mice per group). Unpaired t-test: 

* * P <   0.01,   

* * * P <  0.001,  * * * * P <  0.0001.

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

0 0   M o n t h   2 0 1 7   |   V o L   0 0 0   |   n A t U R E   |   5

letter

reSeArCH

promote proplatelet formation and extension, the lung may contain 

unique signalling partners for megakaryocytes that promote platelet 

release. Interactions between megakaryocytes and endothelial cells 

through glycoprotein 1b (GPIb)–vWF signalling have been shown to 

promote proplatelet formation in vitro

24

. Considering that vWF levels 

are particularly high in the pulmonary arteries

25

, this pathway could 

finely regulate megakaryocytes for platelet production.

The lungs are a reservoir for resident megakaryocytes and haemato-

poietic progenitor cells (Extended Data Fig. 9b), which raises questions 

about the factors responsible for the homing of these cells into and 

out of the lungs, the function of these cells in the lung niche, and the 

roles of these cells in host defence

26

. Additionally, megakaryocytes are 

a rich source of cytokines and growth factors that have the potential to 

influence inflammatory or fibrotic lung diseases. Our RNA-seq analysis  

revealed that lung megakaryocytes were skewed towards an innate 

immunity function (Extended Data Fig. 4d–f and Supplementary  

Table 1), which may reflect the unique environment of the exposed lung 

versus the bone marrow. Indeed, we detected changes in the resident 

and circulating lung megakaryocyte populations in mice with bacterial 

pneumonia (Extended Data Fig. 4g–k). Our findings may also be appli-

cable to the field of lung transplantation, in which post-transplantation  

chimaerism could affect acute and chronic allograft rejection. Our 

results add to the growing evidence that the lungs are a sophisticated 

organ that is capable of regeneration after major injury, are a major site 

of platelet production, and have untapped potential as a contributor to 

haematopoiesis.

Online Content Methods, along with any additional Extended Data display items and 

Source Data, are available in the online version of the paper; references unique to 

these sections appear only in the online paper.

received 24 April 2016; accepted 14 February 2017. 
Published online 22 March 2017.

1.  Semple, J. W., Italiano, J. E., Jr & Freedman, J. Platelets and the immune 

continuum. Nat. Rev. Immunol. 11, 264–274 (2011).

2.  Versteeg, H. H., Heemskerk, J. W., Levi, M. & Reitsma, P. H. New fundamentals 

in hemostasis. Physiol. Rev93, 327–358 (2013).

3.  Machlus, K. R. & Italiano, J. E. Jr. The incredible journey: from 

megakaryocyte development to platelet formation. J. Cell Biol201, 

785–796 (2013).

4.  Howell, W. H. & Donahue, D. D. The production of blood platelets in the lungs. 

J. Exp. Med. 65, 177–203 (1937).

5.  Levine, R. F. et al. Circulating megakaryocytes: delivery of large numbers of 

intact, mature megakaryocytes to the lungs. Eur. J. Haematol51, 233–246 

(1993).

6.  Zucker-Franklin, D. & Philipp, C. S. Platelet production in the pulmonary 

capillary bed: new ultrastructural evidence for an old concept. Am. J. Pathol

157, 69–74 (2000).

7.  Weyrich, A. S. & Zimmerman, G. A. Platelets in lung biology. Annu. Rev. Physiol

75, 569–591 (2013).

8.  Looney, M. R. et al. Stabilized imaging of immune surveillance in the mouse 

lung. Nat. Methods 8, 91–96 (2011).

9.  Alexander, W. S., Roberts, A. W., Nicola, N. A., Li, R. & Metcalf, D. Deficiencies in 

progenitor cells of multiple hematopoietic lineages and defective 

megakaryocytopoiesis in mice lacking the thrombopoietic receptor c-Mpl. 

Blood 87, 2162–2170 (1996).

10.  Aschoff, L. Uber capillare Embolie von riesenkemhaltigen Zellen. Arch Pathol 

Anat Phys 134, 11–14 (1893).

11.  Kallinikos-Maniatis, A. Megakaryocytes and platelets in central venous and 

arterial blood. Acta Haematol42, 330–335 (1969).

12.  Tiedt, R., Schomber, T., Hao-Shen, H. & Skoda, R. C. Pf4-Cre transgenic mice 

allow the generation of lineage-restricted gene knockouts for studying 

megakaryocyte and platelet function in vivoBlood 109, 1503–1506 (2007).

13.  Patel, S. R., Hartwig, J. H. & Italiano, J. E., Jr. The biogenesis of platelets from 

megakaryocyte proplatelets. J. Clin. Invest. 115, 3348–3354 (2005).

14.  Thon, J. N. et al. Cytoskeletal mechanics of proplatelet maturation and platelet 

release. J. Cell Biol191, 861–874 (2010).

15.  Trowbridge, E. A. et al. The origin of platelet count and volume. Clin. Phys. 

Physiol. Meas5, 145–170 (1984).

16.  Fuentes, R. et al. Infusion of mature megakaryocytes into mice yields functional 

platelets. J. Clin. Invest. 120, 3917–3922 (2010).

17.  Kaufman, R. M., Airo, R., Pollack, S. & Crosby, W. H. Circulating megakaryocytes 

and platelet release in the lung. Blood 26, 720–731 (1965).

18.  Sayah, D. M. et al. Neutrophil extracellular traps are pathogenic in primary 

graft dysfunction after lung transplantation. Am. J. Respir. Crit. Care Med. 191, 

455–463 (2015).

19.  Junt, T. et al. Dynamic visualization of thrombopoiesis within bone marrow. 

Science 317, 1767–1770 (2007).

20.  Pereira, J. P., An, J., Xu, Y., Huang, Y. & Cyster, J. G. Cannabinoid receptor 2 

mediates the retention of immature B cells in bone marrow sinusoids.  

Nat. Immunol10, 403–411 (2009).

21.  Liu, Z. J. et al. Expansion of the neonatal platelet mass is achieved via an 

extension of platelet lifespan. Blood 123, 3381–3389 (2014).

22.  Pronk, C. J. et al. Elucidation of the phenotypic, functional, and molecular 

topography of a myeloerythroid progenitor cell hierarchy. Cell Stem Cell 1, 

428–442 (2007).

23.  Pietras, E. M. et al. Functionally distinct subsets of lineage-biased multipotent 

progenitors control blood production in normal and regenerative conditions. 

Cell Stem Cell 17, 35–46 (2015).

24.  Dunois-Lardé, C. et al. Exposure of human megakaryocytes to high shear rates 

accelerates platelet production. Blood 114, 1875–1883 (2009).

25.  Yamamoto, K., de Waard, V., Fearns, C. & Loskutoff, D. J. Tissue distribution and 

regulation of murine von Willebrand factor gene expression in vivoBlood 92, 

2791–2801 (1998).

26.  Haas, S. et al. Inflammation-induced emergency megakaryopoiesis driven by 

hematopoietic stem cell-like megakaryocyte progenitors. Cell Stem Cell 17, 

422–434 (2015).

Supplementary Information is available in the online version of the paper.

Acknowledgements We thank the UCSF BIDC for assistance with 2PIVM and 3D 

printing; A. Hérault, E. Verovskaya and S. Y. Zhang from the Passegué laboratory 

for assistance with hematopoietic progenitor isolation and transplantation; 

and D. Erle and the UCSF SABRE Functional Genomics Facility for assistance 

with the RNA-sequencing experiments. This work was supported in part by 

NIH grants HL092471 to E.P., HL107386 and HL130324 to M.R.L., the UCSF 

Nina Ireland Program in Lung Health (M.R.L.), and the UCSF Program for 

Breakthrough Biomedical Research (M.R.L.).

Author Contributions E.L. designed and conducted most of the experiments, 

analysed the data, and wrote the manuscript. G.O.-M. designed and conducted 

experiments and analysed the data. A.C. and B.M. conducted experiments and 

analysed data. F.L. performed the lung transplantation experiments. D.M.S., 

E.E.T., M.B.H. and T.D. assisted in designing and conducting experiments. 

S.R.C, M.F.K. and A.D.L. assisted in designing experiments and provided 

editorial support on the manuscript. E.P. assisted in designing experiments, 

provided technical expertise with haematopoietic progenitor analyses, and 

provided editorial support on the manuscript. M.R.L. designed the experiments, 

conducted experiments, analysed data, and wrote the manuscript.

Author Information Reprints and permissions information is available at 

www.nature.com/reprints. The authors declare no competing financial 

interests. Readers are welcome to comment on the online version of the paper

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional 

claims in published maps and institutional affiliations. Correspondence and 

requests for materials should be addressed to M.R.L. (mark.looney@ucsf.edu).

reviewer Information Nature thanks F. Ginhoux, S. Morrison, G. Zimmerman 

and the other anonymous reviewer(s) for their contribution to the peer review 

of this work.

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

letter

reSeArCH

MethOdS

Mice. Mice were housed and bred under specific pathogen-free conditions at 

the University of California, San Francisco (UCSF) Laboratory Animal Research 

Center and all experiments conformed to ethical principles and guidelines 

approved by the UCSF Institutional Animal Care and Use Committee. Male and 

female mice between 8 and 12 weeks of age were used for experiments. C57BL/6, 

PF4-Cre, Rosa26-LSL-tdTomato, mTmG, nTnG and Ptprc

a

 Pepc

b

 BoyJ mice were 

purchased from Jackson Laboratories. To track platelets or megakaryocytes,  

PF4-Cre expressing mice were crossed with Rosa26-LSL-tdTomato, mTmG 

or nTnG reporter strains, in which the fluorophore of PF4-expressing cells  

(tdTomato or GFP) is localized to the cytoplasm, the cell membrane, or the nucleus, 

respectively. c-mpl

/

 mice (C57BL/6 background) were obtained from a Material 

Transfer Agreement from Genentech.

Lung intravital imaging. We used 2PIVM to observe megakaryocyte and platelet 

production in real time in mouse lungs. A modified version

27

 of the previously 

published method of stabilized lung imaging was used

8

. Mice were anaesthetized 

with ketamine and xylazine and secured with tape to a custom heated microscope 

stage. A small tracheal cannula was inserted, sutured into place, and attached to 

a MiniVent mouse ventilator (Harvard Apparatus). Mice were ventilated with a 

tidal volume of 10 μ l compressed air (21% O

2

) per gram of mouse weight, a respira-

tory rate of 130–140 breaths per minute, and a positive-end expiratory pressure of 

2–3 cm H

2

O. Isoflurane was delivered continuously to maintain anaesthesia and 

mice were injected with 300 μ l of 0.9% saline solution intraperitoneally every hour. 

The mice were placed in the right lateral decubitus position and a small surgical 

incision was made to expose the rib cage. A second incision was then made into 

the intercostal space between ribs 4 and 5, through the parietal pleura, to expose 

the surface of the left lung lobe. A flanged thoracic window with an 8-mm coverslip 

was then inserted between the two ribs and secured to the stage using a set of two 

optical posts and a 90° angle post clamp (Thor Labs)

27,28

. We applied 20–25 mm Hg 

of suction (Amvex Corporation) to gently immobilize the lung. The two-photon  

microscope objective was then lowered into place over the thoracic window. 

In selected experiments, to permit identification of the lung vasculature, FITC  

dextran (50 μ l of 25 mg/ml; Life Technologies) was injected intravenously into the 

tail vein before imaging.

Spleen and liver intravital imaging. Mice were anaesthetized and ventilated as 

noted above. To expose the spleen, a skin incision was made in the left flank. An 

incision was made along the costal margin to expose and externalize the liver. 

The same window as was used for lung imaging was used to facilitate imaging 

of the spleen and liver. The mouse was placed on a 37 °C temperature-controlled 

heated stage for the duration of the imaging and saline solution was administered 

intraperitoneally every hour.

Bone marrow intravital imaging. Mice were anaesthetized with an initial dose of 

ketamine and xylazine and anaesthesia was maintained with isoflurane delivered 

through a nose cone. Hair and the underlying subcutaneous tissue were removed 

to expose the calvarium. The periosteum was removed using a microsurgical knife. 

To stabilize the skull, we 3D printed an apparatus that was fixed to the mouse skull 

with Vetbond and attached to the heated stage below. The microscope objective 

was then lowered into a 5-mm bevelled hole filled with saline.

Two-photon microscopy. Intravital imaging was performed using a Nikon A1R 

Multi-photon microscope equipped with a Mai Tai DeepSee IR Laser (Spectra 

Physics) (UCSF BIDC). The MaiTai laser was tuned to 920 nm for simultane-

ous excitation of GFP or FITC and tdTomato. Emitted light was detected using 

a 25×  water lens (Nikon) with green (500–550 nm) and red (570–620 nm) filters. 

Images were captured with a high-resolution galvano scanner (1 frame per second, 

512 ×  512 pixels). The microscope was controlled using NIS Element AR software 

(4.50). We captured a 1,052 μ m ×  1,578 μ m  xy surface area (1.66 mm

2

) and z-stack 

images were acquired with z-depths of 5 μ m (total of 40 μ m  z-depth). We captured 

a complete image every 1 min for 120 min.

Image analysis. Images were analysed using Imaris 7.6.1 (Bitplane) or NIS-Element 

(Nikon) software (UCSF BIDC). Surface analysis was performed to quantify and 

characterize the volume, diameter, or circularity of resident and circulating mega-

karyocytes or platelets. Megakaryocytes or megakaryocyte fragments were defined 

as PF4

+

 events with a diameter > 15 μ m. Platelets were defined as PF4

+

 events with 

a diameter between 0.5 and 5 μ m. To calculate the number of platelets released by 

each megakaryocyte observed, the ratio of the megakaryocyte to platelet volumes 

was calculated for each of the 35 fragmenting megakaryocytes or proplatelets 

observed during lung imaging. Megakaryocytes were divided into three groups 

according to the number of platelets that can be produced by one megakaryocyte:  

small (< 500  platelets,  n =  18), medium (500–1,000 platelets, n =  7) and large  

(> 1,000  platelets,  n =  10).

Quantifying platelet production in the lung (see variables in Extended Data 

Table 1). For each movie (~ 2 h), the megakaryocytes observed to be undergoing  

fragmentation in the lung (LungMK

frag

) were quantified: LungMK

frag

 per 

hour =  (LungMK

frag

)/(Acquisition time in min) ×  60. The number of platelets 

released by each megakaryocyte undergoing fragmentation in the lung was calcu-

lated using the volume ratio of the megakaryocyte volume to the average platelet 

volume: N

Plat/MK

 =  Volume

MK

/Volume

Plat

. The number of platelets produced in the 

lung was then calculated: Lung

Platelets

 per hour =  LungMK

frag

 per hour ×  N

Plat/MK

  

×

  Lung fraction, where the lung fraction is the ratio of the mouse total lung  

volume

29

 to the observed lung volume: Lung fraction =  Volume

lung

/Volume

observed

Finally, we estimated the contribution of the lung to overall thrombopoiesis: % lung 

platelet production =  (Lung platelets per hour ×  24)/(Total Platelets per day) ×  100.

The total number of platelets produced per day was calculated according to 

the number of circulating platelets in the mouse blood divided by the life span of 

platelets, and takes into account the fact that one-third of the produced platelets 

are sequestered by the spleen: total platelets per day =  (Plat

Blood

(1 +  Plat

Spleen

))/

(Life

Plat

). In selected experiments, mice were treated with recombinant human 

thrombopoietin (rhTPO, Genentech) intraperitoneally (250 mg/kg), 5 days before 

lung imaging.

Lung, bone marrow, spleen and blood single-cell preparation for flow cytom-

etry or cell sorting. Lung digestion. For lung HSC or megakaryocyte cell sorting, 

lungs were perfused before removal and digestion. Lungs were placed in 2 ml 

PBS with 5 μ l/ml DnaseI (Roche) and 0.5 mg/ml LiberaseTM (Roche), minced 

with scissors in 15-ml tubes, and digested for 30 min at 37 °C before filtration 

through a 100-μ m cell strainer and red blood cell lysis. Samples were then filtered 

through a 40-μ m filter and resuspended for subsequent FACS staining. For exper-

iments in which vascular localization was tested, mice were injected intravenously 

with CD41–APC (eBioscience) or CD45-APC (eBioscience) 5 min before lung  

collection.

Bone marrow isolation. Tibias and femurs from both legs were removed from mice 

following euthanasia. Bone marrow cells were flushed with PBS with 5 mM EDTA 

before filtration through a 70-μ m cell strainer and red blood cell lysis.

Spleen isolation. The spleen was removed and pressed with the end of a plunger 

from a 1-ml syringe into 1-ml PBS with 5mM EDTA before filtration through a 

70-μ m cell strainer and red blood cell lysis.

Flow cytometry and cell sorting. For surface staining, cells or platelets were incu-

bated with anti-Fc receptor antibodies (clone 2.4G2) and stained with antibodies 

in Hanks buffered salt solution (HBSS) with 2% fetal calf serum and 5 mM EDTA 

for 30 min.

Antibody clones used: CD41-APC (MWReg30, eBioscience), CD41-FITC 

(MWReg30, BD), CD41-BV421 (MWReg30, BioLegend), CD41-BV570 

(MWReg30, BioLegend), c-mpl-Biotin (AMM2, IBL), streptavidin PE-Cy7 

(BD), GPVI-FITC (JAQ1, emfret), F4/80-PB (CI:A3-1, BioLegend), CD45-

APC (30F11, BioLegend), CD42d-APC (1C2, eBioscience), CD62P-APC (Psel.

KO2.3, eBioscience). HSCs were stained with rat unconjugated Lin antibodies 

Gr-1 (RB6-8C5), Mac1 (M1/70), B220 (RA3-6B2), CD5 (53-7.3), CD4 (GK1.5), 

CD8 (53-6.7) (eBioscience), Ter-119 (BD), CD3 (17A2, BioLegend), goat anti-rat-

PE-Cy5 (Invitrogen), c-Kit-APC-eFluor780 (2B8, eBioscience), Sca-1-PB (D7, 

BioLegend), CD48-APC (HM48-1, BioLegend), CD150-PE-Cy7 (TC15-12F12.2, 

BioLegend) and CD45.2-FITC (104, eBioscience). HSC and progenitor popula-

tions were defined as follows: myeloid progenitor (Lin

 CD45

+

 Sca-1

 c-Kit

+

), 

megakaryocyte progenitor (CD150

+

 CD41

+

), LSK compartment (Lin

 CD45

+

 

Sca-1

+

 c-Kit

+

), MPP2 (LSK CD48

+

 CD150

+

), MPP3/4 (LSK CD48

+

 CD150

), 

ST-HSC (LSK CD48

 CD150

) and LT-HSC (LSK CD48

 CD150

+

). Mature blood 

cells were stained with CD19-PB (6D5, BioLegend), Gr-1 APC-Cy7 (RB6-8C5, 

eBioscience), CD3-PerCP710 (17A2, eBioscience), and CD11b-APC (M1/70, BD).  

For sorting HSCs, a c-kit

+

 cell enrichment step was done before staining the cells 

using c-kit antibody-conjugated magnetic beads and MACS separation columns 

(Miltenyi Biotec). Stained cells were re-suspended for final analysis in Hanks 

buffered salt solution (HBSS) with 2% serum (FCS) and 1 μ g/ml  propidium 

iodide for dead cell exclusion. Samples were analysed on a LSRII flow cytometer  

(BD Biosciences) and cell sorting was performed using a BD FACS Aria II  

(BD Biosciences) in the UCSF Flow Cytometry Core. Analysis of flow cytometry 

data was performed using Flowjo (Treestar)

Lung perfusion. Lungs were perfused before lung transplant experiments, 

FACS, or cell analysis. Mice were anaesthetized with an intraperitoneal injec-

tion of ketamine (50 mg/kg) and xylazine (10 mg/kg). A 20-gauge angiocath 

was inserted into the trachea and connected to a ventilator. Through a midline 

abdominal incision the diaphragm was incised circumferentially and 0.1 ml 

heparin was injected directly into the inferior vena cava. A thoracotomy was 

performed to expose the heart and lungs. Cold PBS (5 ml) with 0.1 ml heparin 

solution was perfused directly into the right ventricle using a 27-gauge needle. 

The trachea was then tied and the heart–lung block removed and placed in a 

small tube containing PBS. In selected experiments, we also performed retro-

grade lung perfusion by instilling 5 ml cold PBS with 0.1 ml heparin solution 

into the left atrium.

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

letter

reSeArCH

Lung transplant experiments. Left lung transplants in mice were performed as 

previously described

18

. Lungs from PF4-tomato or mTmG donor mice were per-

fused and the left lung was immediately transplanted into C57BL/6 wild-type or 

c-mpl

/

 recipients. Selected c-mpl

/

 recipients received recombinant human 

thrombopoietin (rhTPO, Genentech) intraperitoneally (250 mg/kg) on days 3 

and 40 after transplantation. Blood was collected from the submandibular vein 

every week to count blood platelets, donor-derived platelets (tomato

+

 platelets) 

and donor-derived mature blood cells. The transplanted mice were analysed for 

3 months and then euthanized for bone marrow, blood, and lung removal and 

analysis. In selected experiments, mice were allowed to survive for 10 months after 

transplantation before analysis. For 2PIVM after lung transplantation, a perfused 

mTmG lung was transplanted into a PF4-mTmG recipient or vice versa. The trans-

planted lung was imaged immediately after transplantation.

Blood collection and platelet counting. For whole blood analysis, blood was with-

drawn from the submandibular vein for survival bleeding or by cardiac puncture in 

terminal experiments. Blood was collected into either acid citrate dextrose (Sigma-

Aldrich) for flow cytometry analysis or EDTA tubes (BD microtainer) for platelet 

counting. Platelet counts in the peripheral blood were measured with a Hemavet 

950 FS complete blood counting instrument (Drew Scientific) and confirmed using 

manual platelet counts.

Cell immunofluorescence. Lung megakaryocytes. Nucleated cells that were double- 

positive for tomato and CD41 were sorted from digested lung tissue and cytospin 

slides were prepared. The cells were fixed with PFA 4% and permeabilized with 

0.5% Triton. After saturation with PBS/BSA 3%, cells were stained overnight at 4 °C 

with sheep anti-vWF antibody (Abcam) and 1 h with anti-Sheep AlexaFluor488 

antibody (Invitrogen). Slides were mounted with DAPI mounting medium 

(Molecular Probes) and analysed on a Nikon TI-E high-throughput epifluores-

cence microscope (UCSF Nikon Center).

Bone marrow cells after lung transplants. Bone marrow cells were isolated  

(see single-cell preparation methods) and resuspended in 1 ml PBS-EDTA 5 mM. 

Two-hundred and fifty microlitres of the suspension was stained for 20 min at room 

temperature in the dark with 1 μ m Syto60 (cell-permeant nucleic acid stain) and 

anti CD41-FITC (1 μ l). Cells were washed, resuspended in PBS, and seeded in a 

6-well plate for immediate imaging on a Nikon A1R Multi-photon microscope 

(UCSF BIDC). Cells were excited with a 920-nm laser for simultaneous excitation 

of FITC (green), tdTomato (red) and Syto60 nucleic acid stain (FarRed).

Platelet activation assay. To determine whether the tomato

+

 CD41

+

 events 

detected in the peripheral blood after lung transplantation were bona fide platelets,  

we stimulated whole blood with thrombin (10 nM) to induce the expression of 

P-selectin (RB40.34, BD)

30

, which was measured by flow cytometric analysis and 

compared to platelets from PF4-tomato mice.

MRSA infection. We used the SF8300 strain of MRSA (methicillin-resistant 

Staphylococcus aureus, obtained from C. Chambers at UCSF), which is a minimally 

passaged USA300 clinical strain representative of the epidemic clone USA300-0114. 

Stock solutions of SF8300 at the mid-logarithmic growth phase (10

10

 CFU/ml)  

were aliquoted and frozen at − 80 °C using standard techniques. On the day of the 

experiment, a vial of SF8300 was thawed and diluted with PBS to the concentration  

needed (5 ×  10

7

 CFU per mouse in 50 μ l) for direct tracheal instillation into 

anaesthetized PF4-nTnG mice. Lungs were harvested 24 h after inoculation for 

single-cell preparation and flow cytometry. To test for vascular localization, mice 

were injected intravenously with CD41–APC (eBioscience) 5 min before lung  

collection and CD41-FITC staining.

RNA-seq analysis. For RNA-seq experiments, we used PF4-nTnG mice and sorted 

PF4

+

 CD41

+

 cells from lung and bone marrow prepared for single-cell suspension  

and flow cytometry analysis. For each experiment, tissue from four mice 

was pooled and cells were sorted directly into lysis buffer. Total mRNA was  

isolated from between 1 ×  10

4

 and 4 ×  10

5

 purified cells using a Dynabeads mRNA 

DIRECT kit (Ambion-Life Technologies). Three independent replicates were 

used for each population. An mRNA library was prepared by the UCSF SABRE 

Functional Genomics Facility using low input Nugen Ovation plus Nextera kit  

(< 100 ng RNA) and sequencing was performed using a single-end 50-bp RNA-seq 

Illumina TruSeq Stranded PolyA library kit and an Illumina HiSeq 4000 machine. 

Sequencing yielded ~ 432 million reads in total with an average read depth of  

72 million reads per sample. Reads were then aligned to the mouse genome 

(aligner: STAR_2.4.2a aligner, alignment genome: Ensembl Mouse GRCm38.78) 

and genes that mapped uniquely to known mRNAs were used to assess differential 

expression between lung and bone marrow groups. Differential expression testing 

was carried out using DESeq2 v1.14.0.

Using a log ratio test and fold-change cutoffs (false discovery rate (FDR)  

<

 0.05), we found 705 genes that were differentially expressed: 543 were upregu-

lated in lung megakaryocytes and 162 in bone marrow megakaryocytes. Functional 

pathways representative of each gene signature were analysed for enrichment in 

gene categories from the Gene Ontology Biological Processes (GO-BP) database 

(Gene Ontology Consortium) using DAVID Bioinformatics Resources. For gene 

signatures and each GO category, the significance of the number of overlapping 

genes in the two sets was calculated using a Fisher’s exact test performed by the 

DAVID software. The P value resulting from this test reflects the probability of 

obtaining the observed overlap or greater by chance. GO-BP categories with at 

least three genes and P <  0.001 were identified.

Statistical analysis. For surface analysis, minimum-to-maximum boxplots are 

shown: the line in the middle of the box is plotted at the median, the box extends 

from the 25th to the 75th percentile and the whiskers go down to the smallest value 

and up to the largest. The +  indicates the mean. Other results plotted as histograms 

are reported as mean ±  s.d. and were analysed by t-test, and multi-group com-

parisons were performed using a one-way ANOVA test and Bonferroni post-hoc 

test (GraphPad PRISM version 5.0; GraphPad Software Inc.). In all cases where 

statistical significance is provided, variance was not statistically different between 

groups. Sample sizes were chosen on the basis of previous experience in the labora-

tory with respect to inherent variability in intravital imaging, lung transplantation, 

and cellular transplantation experiments. There is a 10% surgical failure rate with 

the mouse orthotopic lung transplantation surgery, and it was pre-established that 

these mice would be excluded from the analysis. Animals within each cohort were 

randomly assigned to treatment groups. Blinded analysis was not performed in 

these studies. P values of less than or equal to 0.05 were considered to be statistically 

significant, indicated by asterisks in figures, unless otherwise noted.

Data availability. The RNA-seq data that support the findings of this study have 

been deposited in the NIH SRA database with the accession code SRP097794.

27.  Headley, M. B. et al. Visualization of immediate immune responses to pioneer 

metastatic cells in the lung. Nature 531, 513–517 (2016).

28.  Ortiz-Muñoz, G. et al. Aspirin-triggered 15-epi-lipoxin A4 regulates neutrophil-

platelet aggregation and attenuates acute lung injury in mice. Blood 124, 

2625–2634 (2014).

29.  Canals, M., Olivares, R. & Rosenmann, M. A radiographic method to 

estimate lung volume and its use in small mammals. Biol. Res38, 41–47 

(2005).

30.  Cornelissen, I. et al. Roles and interactions among protease-activated receptors 

and P2ry12 in hemostasis and thrombosis. Proc. Natl Acad. Sci. USA 107, 

18605–18610 (2010).

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

letter

reSeArCH

Extended Data Figure 1 | Megakaryocytes and proplatelets observed in 

lung circulation are from an extrapulmonary source. a, Lung 2PIVM 

of a PF4-nTnG mouse (nuclear GFP). The presence of the mobile GFP

+

 

nucleated cells (circled) indicates the presence of a nucleus in circulating 

megakaryocytes. b, Platelet counts in the blood before and after imaging. 

n.s., not significant (n =   3).  c, Experimental schema of transplantation 

of lungs from mTmG mouse (perfused donor lung) to PF4-mTmG 

(recipient) mouse and vice-versa followed by 2PIVM. d, 2PIVM of a 

mTmG mouse lung showing no GFP signal. e, 2PIVM of an mTmG mouse 

lung transplanted into a PF4-mTmG recipient mouse showing GFP

+

 

cells from recipient and platelet production in the lung. f, Bone marrow 

2PIVM apparatus. g, Representative image of proplatelet release in bone 

marrow (BM) sinusoids (arrows). h, 2PIVM of PF4-mTmG mouse liver. 

Small platelets (GFP, green) were seen in the circulation but neither 

resident nor circulating megakaryocytes or proplatelets were observed. 

i, 2PIVM of PF4-mTmG mouse spleen. Sequential images show resident 

megakaryocytes (GFP, green) releasing proplatelets (arrows) in the spleen 

vasculature (in red).

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

letter

reSeArCH

Extended Data Figure 2 | Resident megakaryocytes in the lungs and 

other organs. a, Survey of PF4-tomato mouse lung visualized by 2PIVM. 

PF4-tomato-expressing cells (red) are found in high numbers in the lungs. 

Lung vasculature is labelled by intravascular injection of FITC dextran 

(green). A total area of 2.49 mm

2

 (1.6 mm ×  1.6 mm) was imaged.  

b, Resident (static) GFP

+

 cells are found in a PF4-mTmG lungtransplanted 

into an mTmG mouse. cd, 2PIVM images of bone marrow (c) and  

spleen (d) from PF4-mTmG mice. Many large megakaryocytes (GFP, green) 

are found in the bone marrow and spleen. e, Size characterization of resident 

(static) GFP

+

 cells by image analysis of PF4-mTmG lungs (n =  16),  bone 

marrow (n =  12), and spleen (n =  16). Minimum-to-maximum boxplots are 

shown: the line in the middle of the box is plotted at the median, the box 

extends from the 25th to 75th percentiles and the whiskers range from the 

smallest to the largest values. The +  indicates the mean.

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

letter

reSeArCH

Extended Data Figure 3 | Surface expression of lung megakaryocytes 

compared to bone marrow megakaryocytes. a, Flow cytometric analysis 

of nGFP

+

 cells from PF4-nTnG lungs. b, CD41 expression defines two 

populations of megakaryocytes: CD41

+

 (red) and CD41

 (green).  

c, Positive surface expression of GPVI, c-Mpl and CD45 was detected in 

both populations. Unstained cells are plotted in blue. d, Surface expression 

of F4/80, CD34, CD11b, Sca-1 and c-Kit was not detected. ei, The CD41

+

 

population has a higher percentage of CD61

+

 cells (e), CD42b

+

 cells (f)  

and larger cells (g) and had higher DNA content (h), as summarized 

in i (n =   3).  j, Flow cytometric analysis of nGFP

+

 cells from PF4-nTnG 

bone marrow. km, Compared to the lungs, the bone marrow nGFP

+

 

population has a higher percentage of CD41

+

 cells (n =   21–23)  (k), CD61

+

 

cells (n =   3)  (l), and CD42b

+

 cells (n =   3)  (m). Data are representative  

of three or more replicates. Mean ±  s.d. are presented. Unpaired t-test:  

P <  0.05,  * * P <  0.01,  * * * P <  0.001.

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.