Homozygous loss of BHD causes early embryoniclethality and kidney tumor developmentwith activation of mTORC1 and mTORC2Yukiko Hasumia, Masaya Babaa, Rieko Ajimab, Hisashi Hasumia, Vladimir A. Valeraa, Mara E. Kleina, Diana C. Hainesc,Maria J. Merinod, Seung-Beom Honga, Terry P. Yamaguchib, Laura S. Schmidta,e, and W. Marston Linehana,1

aUrologic Oncology Branch, Center for Cancer Research and dLaboratory of Pathology, National Cancer Institute, National Institutes of Health, Bethesda,MD 20892; and bCell Signaling in Vertebrate Development Section Cancer and Developmental Biology Laboratory and eBasic Research Program andcPathology/Histotechnology Laboratory, SAIC-Frederick, Inc., National Cancer Institute-Frederick, Frederick, MD 21702

Edited by Richard D. Klausner, The Column Group, Seattle, WA, and approved September 17, 2009 (received for review August 6, 2009)

Germline mutations in the BHD/FLCN tumor suppressor gene pre-dispose patients to develop renal tumors in the hamartoma syn-drome, Birt-Hogg-Dube (BHD). BHD encodes folliculin, a proteinwith unknown function that may interact with the energy- andnutrient-sensing AMPK-mTOR signaling pathways. To clarify BHDfunction in the mouse, we generated a BHD knockout mousemodel. BHD homozygous null (BHDd/d) mice displayed early em-bryonic lethality at E5.5–E6.5, showing defects in the visceralendoderm. BHD heterozygous knockout (BHDd/�) mice appearednormal at birth but developed kidney cysts and solid tumors asthey aged (median kidney-lesion-free survival � 23 months, me-dian tumor-free survival � 25 months). As observed in human BHDkidney tumors, three different histologic types of kidney tumorsdeveloped in BHDd/� mice including oncocytic hybrid, oncocytoma,and clear cell with concomitant loss of heterozygosity (LOH),supporting a tumor suppressor function for BHD in the mouse. ThePI3K-AKT pathway was activated in both human BHD renal tumorsand kidney tumors in BHDd/� mice. Interestingly, total AKT proteinwas elevated in kidney tumors compared to normal kidney tissue,but without increased levels of AKT mRNA, suggesting that AKTmay be regulated by folliculin through post translational or post-transcriptional modification. Finally, BHD inactivation led to bothmTORC1 and mTORC2 activation in kidney tumors from BHDd/�

mice and human BHD patients. These data support a role forPI3K-AKT pathway activation in kidney tumor formation caused byloss of BHD and suggest that inhibitors of both mTORC1 andmTORC2 may be effective as potential therapeutic agents forBHD-associated kidney cancer.

Birt-Hogg-Dubé syndrome � kidney cancer � mouse model � mTOR �tumor suppressor

B irt-Hogg-Dube (BHD) syndrome is an inherited kidneycancer syndrome which predisposes patients to develop hair

follicle tumors, lung cysts, spontaneous pneumothorax, and anincreased risk of renal neoplasia (1–3). We previously identifiedgermline mutations in the BHD (FLCN) gene in patients withBHD (4). About one-third of BHD patients develop bilateralmultifocal renal tumors that are most frequently chromophoberenal tumors and renal oncocytic hybrid tumors with features ofchromophobe renal carcinoma and renal oncocytoma (5). So-matic mutations in the wild-type copy of BHD and loss ofheterozygosity at chromosome 17p11.2 have been identified inhuman BHD tumors, indicating that BHD is a classical tumorsuppressor gene (6). The BHD protein folliculin (FLCN) is a64-kDa protein with no known functional domains (4). Wereported two FLCN binding proteins FNIP1 and FNIP2, whichinteract with 5�-AMP-activated protein kinase (AMPK), animportant energy sensor in cells that negatively regulates mam-malian target of rapamycin (mTOR), the master switch for cellgrowth and proliferation (7–9). These findings suggest thatFLCN may play a role in cellular energy and nutrient sensing

through interactions with the AMPK-mTOR signaling pathway.Mutations in several other tumor suppressor genes, includingLKB1 (10), PTEN (11), and TSC1/2 (12), have been shown tolead to dysregulation of PI3K-AKT-mTOR signaling and to thedevelopment of other hamartoma syndromes. We and otherspreviously reported the generation of a conditionally targetedBHD allele and kidney-directed BHD inactivation in the mouseusing the cadherin16 (KSP)-Cre transgene (13, 14). AlthoughBHD homozygous deletion in kidney epithelial cells was suffi-cient to cause uncontrolled cell proliferation and hyperplasticcell transformation, the kidney-targeted BHD-knockout micelived only approximately 3 weeks and did not produce kidneytumors. A BHD heterozygous knockout mouse model thatdevelops tumors with age will more accurately reflect tumordevelopment in the human BHD patient and may be a bettermodel for understanding how BHD inactivation leads to tumorinitiation and progression. Here we report the analysis of anembryonic lethal phenotype that occurs in a BHD homozygousknockout mouse model and characterize and compare thekidney tumors that develop in a BHD heterozygous knockoutmouse model with human BHD kidney tumors.

ResultsRole of BHD during Early Embryogenesis. We have analyzed mouseBHD mRNA expression levels by qRT-PCR in wild-type em-bryos and adult tissues (Fig. S1). We detected consistent BHDmRNA expression from E8.5 to E12.5 with 4-fold elevation atE19 and high expression in adult heart, pancreas, and prostatewith moderate expression in adult brain, kidney, liver, and lung.BHD mRNA expression was further analyzed during earlyembryogenesis by whole mount in situ hybridization (Fig. S2).BHD mRNA was expressed consistently throughout embryo-genesis. At E5.5, BHD expression was restricted to extraembry-onic tissues; however, by E6.5, BHD was expressed in bothembryonic and extraembryonic tissues. We saw strong expres-sion in certain tissues including neural ectoderm, headfold, andlimb buds, but the signal was relatively weak in the surroundingendoderm and heart.

Next we evaluated BHD homozygous knockout (BHDd/d)embryos from intercrosses of BHD heterozygous knockout(BHDd/�) mice. BHDd/� mice appeared normal at birth and

Author contributions: Y.H., M.B., L.S.S., and W.M.L. designed research; Y.H., M.B., R.A.,H.H., V.A.V., M.E.K., and L.S.S. performed research; Y.H., M.B., D.C.H., M.J.M., S.-B.H., T.P.Y.,L.S.S., and W.M.L. analyzed data; and Y.H., M.B., L.S.S., and W.M.L. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.

1To whom correspondence should be addressed. E-mail: wml@nih.gov.

This article contains supporting information online at www.pnas.org/cgi/content/full/0908853106/DCSupplemental.

18722–18727 � PNAS � November 3, 2009 � vol. 106 � no. 44 www.pnas.org�cgi�doi�10.1073�pnas.0908853106

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0

developed normally, but no BHDd/d mice were born in 75neonates indicating embryonic lethality (Fig. 1A). No BHDd/d

embryos were found at E9.5. BHDd/d embryos were found beforeE8.5 with lower frequency (9.3%, 15/161) than expected (25%).There were many empty deciduas suggestive of early embryonicdeath and resorption (21.8%, 26/119 at E5.5 and E6.5). Themorphologies of all BHDd/d embryos were abnormal with oneexception at E5.5. By gross appearance BHDd/d embryos (Fig. 1B and D) were thinner or smaller than BHDd/� (Fig. 1 C and E)or BHD�/� embryos with occasional bleeding. There were nohistological abnormalities in BHD�/� or BHDd/� embryos (Fig.1 F and G). Most BHDd/d embryos lacked an organized cell layerwith only a small cell mass (Fig. 1I). We could occasionally seeBHDd/d embryos consisting of two types of cells, clear visceralendoderm-like cells and ectoderm cells (Fig. 1H). Immunohis-tochemical staining with the visceral endoderm (VE) markerDAB2 confirmed that the outer cell layer was VE (Fig. 1 J–M).Interestingly, VE of the BHDd/d embryos showed distortedenlarged cytoplasm and a disorganized structure (Fig. 1 K andM) instead of a cuboidal epithelial layer as seen in the BHDd/�

embryos (Fig. 1 J and L). Also the nuclei of the VE cells weredisorientated in BHDd/d embryos (Fig. 1H’), whereas nuclei werealigned along the basal membrane in an orderly fashion inBHD�/� embryos (Fig. 1F’). BHDd/d embryos also displayed a

disorganized ectoderm structure and did not show cavitation ora polarized epithelial cell layer (Fig. 1 H and K).

Spontaneous Kidney Tumor Development in BHDd/� Mice with Loss ofHeterozygosity. BHDd/� mice spontaneously developed cysts,complex cysts and solid tumors in their kidneys after the age of10 months (Fig. 2 A and B). Two of 23 BHD�/� mice displayedsmall, isolated simple cysts in their kidneys. The median age ofkidney-lesion-free survival for BHDd/� mice was 23 months (n �65), compared with an undefined kidney-lesion-free survival forBHD�/� littermate controls (n � 28, P � 0.0001) (Fig. 2C). Nosolid kidney tumors developed in BHD�/� mice (BHDd/� mediantumor-free survival � 25 months, n � 65; BHD�/� mediantumor-free survival � undefined, n � 28; P � 0.0026) (Fig. 2D).The number of kidney lesions in BHDd/� mice between the agesof 20 and 25 months is shown in Fig. 2E (n � 35; no. cysts peranimal, mean � 3.43, SD � 3.13; no. mixed lesions per animal,mean � 0.51, SD � 0.85; no. solid tumors per animal, mean �0.51, SD � 0.82). Histological examination showed that kidneycysts that develop in BHDd/� mice were lined by hyperplasticcells with enlarged cytoplasm and nuclei (Fig. 3A). The cystsfound in BHD�/� mice were lined by flat cyst cells characteristicof simple cysts, which were distinct from the hyperplastic cyststhat developed in BHDd/� mice. Complex cysts, defined as cysts

A

243

7505223neonates

72041E9.5

3232(2)22(1)5E8.5

1031(1)3(1)3E7.5

54138(8)2310E6.5

65134(3)3018E5.5

totalEmptydecidua

d/d(abnormal)d/++/+

7505223

72041E9.5

3232(2)22(1)5E8.5

1031(1)3(1)3E7.5

54138(8)2310E6.5

6134(3)3018E5.5

totalEmptydecidua

d/d(abnormal)d/++/+

E5.5 d/d E5.5 d/+.5 d/d E6.5 d/+

F IHG

H'B C D E

F'

E6.5 +/+

E5.5 d/d E5.5 d/+ E6.5 d/d E6.5 d/+

ML

J K

E6.5 +/+

E6.5 d/d E6.5 d/+

E6.5 d/d

E6.5 d/d E6.5 d/d E6.5 +/+

Fig. 1. Characterization of BHDd/d mouse embryo phenotype. (A) Embryos were isolated from BHD d/� intercrosses, observed under a dissection microscope andgenotyped by PCR. Numbers in parentheses represent embryos with abnormal appearance, as shown in (B and D). Gross appearance of BHDd/d (B and D) or BHD d/�

(C and E) embryos. H&E images of BHD �/� (F), BHDd/� (G), and BHDd/d (H and I) embryos at E6.5. Immunohistochemical staining of DAB2 was performed to seevisceral endoderm of BHDd/� (J), and BHDd/d (K) embryos. Magnified images of visceral endoderm of BHD�/� (F’ and L) and BHDd/d (H’ and M) with H&E stainingand DAB2 staining, respectively. Magnification; �20 (F–I), �40 (J and K), �63 (F’ and H’), and �100 (L and M). [Scale bar, 100 �m (F–I) and 50 �m (J and K).]

cyst mixed lesion solid tumor0

5

10

15

mixed lesionsolid tumor

cyst

lesi

on

num

ber

/ a

nim

al

kidn

ey tu

mor

free

sur

viva

l

0 10 20 30 400

50

100

150+/+

d/+

month

n=28

n=65

Median survival +/+ Undefinedd/+ 25month

Log-rank TestP value = 0.0026

kidn

ey le

sio

n fr

ee s

urvi

val

0 10 20 30 400

50

100

150+/+

d/+

month

n=28

n=65

Median survival +/+ Undefinedd/+ 23month

Log-rank TestP value <0.0001

EC D

B

A

**

*

Fig. 2. Spontaneous kidney tumor development in BHDd/� mice. BHDd/� mice and BHD�/� mice were aged and dissected randomly at different time points whenthey were moribund or had to be euthanized due to dermatitis. The renal capsule was removed from isolated kidneys and observed under dissection microscopy.(A) Gross appearance of BHDd/� mouse kidney. (B) Solid tumor (**) in BHDd/� kidney(*). (C) Kaplan-Meier analysis of BHDd/� (n � 65) and BHD�/� (n � 28) micefor kidney-lesion-free survival. Log-rank test (two-sided), P � 0.0001. Dotted lines, SEM. (D) Kaplan-Meier analysis of BHDd/� (n � 65) and BHD�/� (n � 28) micefor kidney-tumor-free survival (complex cysts and solid tumors). Log-rank test (two-sided), P � 0.0026. (E) The number of kidney lesions was counted in BHDd/�

mice between the ages of 20 and 25 months (n � 35). (Scale bar, SEM.)

Hasumi et al. PNAS � November 3, 2009 � vol. 106 � no. 44 � 18723

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0

with some structure or solid portions in the lumen, wereobserved. These complex cysts showed papillary protrusions intothe lumen (Fig. 3B) as well as regions containing solid tumor(Fig. 3C). Interestingly the solid tumors displayed histologiessimilar to human kidney tumors that develop in BHD patients,including clear cell (Fig. 3D), oncocytic hybrid consisting of amixture of chromophobe renal carcinoma and oncocytic cells(Fig. 3E), and oncocytoma (Fig. 3F). Southern blot analysis oftumors that formed in BHDd/� mice showed a severely reducedwild-type BHD allele signal compared to adjacent normal kidneytissue, supporting LOH at the BHD locus (Fig. 3G). Western blotanalysis showed only very weak FLCN expression in the tumorscompared with adjacent BHDd/� kidney or wild-type kidneytissue (Fig. 3H), thus confirming the inactivation of both BHDalleles. Loss of endogenous FLCN expression was also con-firmed by immunofluorescent staining using the Duolink in situproximity ligation assay (PLA), which enables visualization ofendogenous FLCN with two different antibodies (Fig. 3 I and J).Importantly all five kidney tumors that were analyzed fromBHDd/� mice showed very low FLCN expression compared tonormal mouse kidneys. These results strongly support thepremise that kidney tumors, which developed in the BHDd/�

mouse model, were produced as a consequence of loss of FLCNfunction.

Activation of the PI3K-AKT-mTOR Signaling Pathway in Kidney Tumorsfrom BHDd/� Mice and Human BHD Patients. Total AKT proteinlevels, AKT1, and AKT2 as well as total AKT, were dramaticallyelevated in kidney tumors compared with normal kidney tissue(Fig. 4A). Levels of AKT phosphorylation were also dramaticallyelevated at both the PDK1 phosphorylation site (Thr308) andthe mTORC2 phosphorylation site (Ser473) (15) (Fig. 4 A andB). AKT mRNA levels were measured by qRT-PCR in the samesamples (Fig. 4C). Contrary to the Western blot results, mRNAlevels of AKT1 and AKT2 were not significantly different be-tween tumors and normal kidneys. Increased phosphorylation ofdownstream effectors of AKT signaling would support AKTactivation in kidney tumors that developed in BHDd/� mice.Indeed, p-GSK3, p-FOXO1, and p-FOXO3a were elevated intumors compared with normal kidneys (Fig. 4 D–F). We also sawCyclin D1 elevation, which might be a consequence of GSK3phosphorylation (Fig. 4D). mTOR phosphorylation on serine2448, which reflects mTORC1 activation (16), was higher intumors than in normal kidneys. Furthermore, mTOR phosphor-ylation on serine 2481, indicating mTORC2 activation (17), wasalso higher in tumors compared with normal kidney tissue (Fig.5A). Interestingly Rictor was more highly expressed in tumorsfrom BHDd/� mice, which may lead to mTORC2 activation andresult in higher AKT phosphorylation on serine 473. Levels ofphospho(p)-S6 kinase (Thr421/Ser424) and phospho S6 ribo-somal protein (Ser240/244), readouts of mTORC1 activation,were higher in tumors than in normal kidney tissue, althoughtotal protein levels of S6K and S6R were also elevated (Fig. 5A).Consistent with the Western blot results, we were able to seestronger p-mTOR (Ser2448) and p-S6R (Ser240/244) staining incells lining the cysts and in tumors from BHDd/� mice whencompared to adjacent normal kidney (Fig. 5 B–G).

Finally we evaluated these protein levels in human BHDtumors and normal kidneys. Western blotting and immunoflu-orescent staining showed higher levels of p-AKT(Ser473) inkidney tumors compared to normal kidney controls (Fig. 6 A, E,and F). We quantified p-AKT(Ser473), p-S6K(Thr421/Ser424),and p-S6R(Ser240/244) using a Meso Scale Discovery multiplexmicrotiter plate assay, a quantitative assay system using acombination of electrochemiluminescence detection and pat-terned arrays (Fig. 6 B–D). The protein levels of p-AKT (Ser473)were more than 30 times higher in human BHD tumors than innormal kidneys. Levels of p-S6K (Thr421/Ser424) protein andp-S6R protein were also higher in the BHD kidney tumors (Fig.6 C and D). We compared the genotype and phenotype of theBHD patients whose kidney tumors were evaluated in this study(Fig. 6G). The kidney tumors were surgical specimens from threedifferent female patients with different types of germline BHDmutations: missense, splicing defect and frameshift. All of thepatients with different BHD mutations showed the classic triadof BHD phenotypic features- fibrofolliculomas, pulmonary cystsand renal tumors (Fig. S3). Importantly, the molecular pheno-type was similar among the tumors derived from these threedifferent types of BHD alterations, underscoring the importanceof PI3K-AKT-mTOR activation for kidney tumorigenesis inBHD patients.

DiscussionThe early embryonic lethality of BHD homozygous knockoutmice supports an essential role for BHD in mouse development.Embryonic ectoderm-like cells of BHDd/d embryos did not formthe proamniotic cavity or bilayered ectoderm structure. Thevisceral endoderm (VE) cell layer was disorganized with mis-aligned nuclei suggesting loss of polarity. Interestingly, VE cellsdisplayed swollen cytoplasm with enlarged vacuoles. In addition

Nor

mal

kid

ney1

(+/+

)

Nor

mal

kid

ney2

(+/+

)

Nor

mal

kid

ney3

(d/+

)

Nor

mal

kid

ney4

(d/+

)

Kidn

ry tu

mor

1Ki

dney

tum

or2

Wild allele (7.0kb)

Deleted allele (4.9kb)

FLCN

β−actin

Tissue genotype +/+ +/+ d/+ d/+ d/d d/d

G

H

I

J

Adj. kidney

Tumor

A C

D E F

B

Fig. 3. Histological analysis of BHDd/� mouse kidney lesions. (A) H&E stainingon formalin fixed paraffin embedded kidney samples. Cells lining the cystshow proliferative tubular epithelium unique to BHD cysts. (B) Cyst withpapillary projection. (C) Micro solid tumors adjacent to BHD cyst. (D–F) Dif-ferent histological features were observed in tumors, resembling clear cell (D),oncocytic hybrid (E), and oncocytoma (F). Magnification; �5 (B), �10 (C), and�20 (A and D–F). [Scale bar, 400 �m (B); 200 �m (C); and 100 �m (A and D–F).](G) Southern blotting was performed on genomic DNA isolated from normalkidneys and tumors. Loss of wild-type BHD alleles was observed in the mousekidney tumors. (H) Western blotting was performed on the samples corre-sponding to (G), showing loss of FLCN protein expression. (I and J) Loss of FLCNprotein in tumors was detected by immunostaining with the Duolink system.Representative staining of FLCN in adjacent kidney (I) and tumor (J) in thesame BHDd/� mouse. Magnification; �100. (Scale bar, 20 �m.)

18724 � www.pnas.org�cgi�doi�10.1073�pnas.0908853106 Hasumi et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0

to nutrient uptake and transport, the anterior VE plays an activerole in formation of the proamniotic cavity during the postim-plantation period (18). The existence of swollen vacuoles maysuggest a defect in phagocytosis, trafficking, or digestion. Theloss of polarity of VE cells resulting from BHD inactivation maycompromise their function. LKB1, working upstream of AMPK,regulates cell polarity (19) and loss of LKB1 or AMPK functionis associated with a defect in cell polarity. The BHD encodedprotein, folliculin (FLCN), interacts with FNIP1/2, which asso-ciate with AMPK. Loss of BHD may lead to a defect in cellpolarity by altering LKB1/AMPK signaling resulting in embry-onic lethality of BHDd/d mice.

Kidney tumor development in BHDd/� mice mimics the kidneytumor phenotype found in humans with BHD. Previously wereported a kidney-targeted conditional BHD knockout mousemodel, which produced enlarged highly cystic kidneys displaying

profoundly increased cell proliferation and hyperplastic mor-phologic changes. However, the animals died of renal failure at21 days of life and therefore did not live long enough to developkidney tumors. Furthermore, additional genetic and/or epige-netic events may be required for tumor formation. BHDd/� micedeveloped hyperplastic kidney cysts, complex cysts and solidtumors at different frequencies and with different latency peri-ods, providing evidence to support a multistep process in BHD-associated kidney carcinogenesis. Loss of both copies of BHD inall analyzed tumors arising in BHDd/� mice supports BHDinactivation as the initiating step for kidney tumorigenesis inBHD.

Although other naturally-occurring animal models for BHDhave been described (20, 21), they may harbor additional geneticchanges that could confound studies of the functional conse-quences of BHD inactivation. The genetically engineered mousemodel in this report provides a ‘‘clean’’ system with which topursue FLCN functional studies. Hartman et al. (22) has de-scribed another BHD heterozygously targeted mouse model inwhich a gene trap cassette inserted in intron 8 encodes aFLCN-�geo fusion protein. These investigators reported a lowerincidence of kidney tumors than in our BHDd/� model, (threetumors in 31 BHD�/� mice and 0 tumors in 15 wild-type mice),which may possibly be due to a shorter observation period (17 vs.30 months). Reduced phospho-S6R (Ser235/236) immunostain-ing of paraffin-embedded tumors led these investigators toconclude that mTOR activity was suppressed in kidney tumorsthat developed in Bhd�/� mice. The inconsistencies between theresults of Hartman et al. and our results may be due todifferences in gene targeting strategy. In our model, mRNAtranscribed from the BHDd allele creates a frameshift resultingin a premature termination codon at the beginning of exon 8 andwill be degraded by nonsense mediated decay (NMD). In fact, notruncated forms of FLCN protein were detected by Westernblotting in our BHDd/� mouse kidneys. However, althoughHartman et al. did not evaluate the presence of the FLCN-�geofusion protein, successful selection of targeted embryonic stemcells by G418 screening would necessitate the expression of aFLCN-�geo fusion protein that retains the N-terminal half ofFLCN in the embryonic stem cells and, presumably, also in theBhd�/� mouse tumors. Since Hartman et al. did not confirmLOH of BHD, it is not clear if the kidney tumors that developedin the Bhd�/� mice were caused by homozygous inactivation of

Adj

. kid

ney

Kid

ney

tum

or

p-AKT(Ser473) IHC

pGSK3/DAPIpGSK3/DAPI

Kid

ney

tum

or

Adj

. kid

ney

Normal TumorB D

AKT

p-AKT(Thr308)

p-AKT(Ser473)

AKT1

AKT2

β-actin

Normal Tumor

SE

LE

CA

p-FOXO1

p-FOXO3a

β-actin

p-GSK3α/β

E F

cyclin D1

Fig. 4. Elevated total AKT and phosphorylated AKT protein in kidney tumors arising in BHDd/� mice. (A) Western blotting was performed on the protein lysatesisolated from normal kidneys and kidney tumors. Both total AKT and phospho-AKT expressions were up-regulated in BHDd/� tumors. Isoforms of AKT, AKT1,and AKT2, were overexpressed in tumors. (B) Immunohistochemical staining showed highly expressed p-AKT (Ser-473) in kidney tumors consistent with Westernblot results. (C) qRT-PCR was performed on total RNA isolated from frozen tissue samples corresponding to (A). mRNA expression for AKT1 and AKT2 was notsignificantly different between tumors and normal kidneys. (D) Western blotting on tissue lysates corresponding to (A). AKT activation resulted in phosphor-ylation of downstream effectors of the AKT pathway in tumors. (E and F) Immunofluorescent staining of p-GSK3�/� was consistent with Western blotting in (D).Magnification; �63. (Scale bar, 20 �m.)

Normal Tumor

p-mTOR(ser2448)

β-actin

FLCN

p-mTOR(ser2481)

Total mTOR

Rictor

Raptor

Total S6K

p-S6K

p-S6R(Ser240/244)

Total S6R

A

pS6R/DAPI

Kid

ney

cys

t K

idn

ey t

um

or

Ad

j. ki

dn

ey

pmTOR/DAPI

pmTOR/DAPI

pmTOR/DAPI

pS6R/DAPI

pS6R/DAPI

p-mTOR(Ser2448) p-S6R(Ser240/244)

B

C

D

E

F

G(Thr421/Ser424)

Fig. 5. mTOR pathway activation in kidney tumors from BHDd/� mice. (A)Western blotting was performed on the protein lysates isolated from normalkidneys and kidney tumors. Both mTOR phosphorylation sites (Ser2448 andSer2481) were more phosphorylated in tumors than in normal mouse kidneys.Downstream effectors of the mTOR pathway were highly expressed in mousekidney tumors suggesting mTORC1 activity is up-regulated in these tumors.Immunofluorescence staining was performed on normal adjacent kidney,cysts, and tumors (B–G). Phospho-mTOR (Ser2448) (B–D) and its downstreamtarget phospho-S6R (E–G) were more highly expressed in tumors and cystslining cells than in adjacent kidney. Magnification, �63. (Scale bar, 20 �m.)

Hasumi et al. PNAS � November 3, 2009 � vol. 106 � no. 44 � 18725

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0

BHD or by another molecular mechanism. If the FLCN-�geofusion protein has partial function it could down-regulate mTORand explain the discrepancy between the mTOR activation seenin tumors that developed in our BHDd/� mice and the reducedmTOR activity reported by Hartman in tumors that developedin the gene trap mouse model for BHD. One additional differ-ence between the two studies was the method of tissue preser-vation for immunostaining, which may affect the antigenicity ofcertain proteins. In this report freshly frozen tissues in OCTcompound were analyzed, whereas paraffin embedded tissueswere used in the Hartman’s study.

We found PI3K-AKT-mTOR pathway activation in both kidneytumors from BHDd/� mice and human BHD tumors, consistentwith the kidney-targeted BHD knockout kidney results, supportinga role for PI3K-AKT-mTOR pathway in both BHD null kidneytumorigenesis and in the development of hyperplastic kidney cysts.Interestingly, total AKT protein levels were elevated in thosetumors without changes in AKT mRNA levels. Therefore FLCNmay be involved in regulation of total AKT protein levels throughpost-translational or post transcriptional modification. It is possiblethat elevated total AKT could result in higher AKT activation asindicated by elevated AKT phosphorylation (Thr308/Ser473). AKTdownstream target molecules were also highly phosphorylated intumors from BHDd/� mice, supporting the possibility that elevatedtotal AKT may be driving the activation of the AKT pathway.Downstream of AKT, we found mTORC1 was activated in tumorsfrom BHDd/� mice. The high levels of p-AKT (Ser473) suggestmTORC2 activation. As expected, we saw elevated levels of p-mTOR on Ser2481, a readout of mTORC2 activity (17), in themouse kidney tumors. Additionally we observed elevated Rictorexpression in tumors from BHDd/� mice. mTORC2 activity isregulated by Rictor expression level (23), and Facchinetti et al.reported that mTORC2 phosphorylated AKT on the turn motif andstabilized AKT (24). Taken together, the high expression level ofRictor and subsequent activation of mTORC2 may be the primarymechanism by which AKT activation occurs in BHD null tumors.Previously we found that FLCN phosphorylation was partiallyblocked by rapamycin (7), suggesting that FLCN function may beregulated by mTOR. This may support a hypothesis whereby FLCN

is on a negative feedback loop suppressing PI3K-AKT-mTORsignaling (Fig. S4).

Each kidney tumor from human BHD patients that wasanalyzed showed PI3K-AKT-mTOR activation, regardless oftype of BHD mutation. Most frameshift or splicing defectmutations found in BHD patients are predicted to produceaberrant mRNAs that would be degraded by nonsense mediateddecay. Two different missense mutations have been reported todate; however, the physiological significance of these mutationshas yet to be determined. Tumor 1 with the H255Y missensemutation (Fig. 6 A and G) showed the same molecular pheno-type as the tumors from patients with germline frameshift orsplicing mutations, suggesting loss of function of this mutantFLCN protein. Our data are consistent with a common conse-quence of BHD inactivation in mouse and man, regardless ofBHD mutation type, and support activation of AKT signaling asan important mechanism driving kidney tumorigenesis in BHDsyndrome. Rapamycin does not inhibit mTORC2 effectively,which may explain its partial effect on kidney-targeted BHDknockout mice (13). Taken together, data generated from ourtwo BHD mouse models suggest that mTORC2 as well asmTORC1 inhibition may be needed for the development of aneffective form of therapy for patients with BHD-associatedkidney cancer.

Materials and MethodsDevelopment of BHD Knockout Mouse Model. The BHD heterozygous knockoutmice were generated as previously described (13). Details of the targetingstrategy are described in the SI Methods (Fig. S5). All mice which were used inthese experiments were housed in the National Cancer Institute (NCI)-Frederick animal facility according to the NCI-Frederick Animal Care and UseCommittee guidelines.

PCR-Based BHD Genotyping. Mouse genomic DNA was isolated from tails(weaned neonates), yolk sacs (E8.5 or later), and whole embryos (E7.5 orearlier). Primers and details are in the SI Methods.

Quantitative Real Time-PCR. The qRT-PCR for BHD, AKT1, and AKT2 wasperformed as described in the SI Methods.

B

p-AKT/DAPI

p-AKT/DAPI

P -AKT

(Thr308)

P -AKT

(Ser473)

AKT1

AKT2

β -actin

G

0

10

20

30

40

50

60

0

5

10

15

20

25

30

0

1

2

3

4

5

6

p-AKT (Ser473) p-S6R(Ser240/244)p-S6K(Thr241/Ser244)

Norm

al kid

ney 3

Norm

al kid

ney 4

BHD tum

or 3

BHD tum

or 4

D E

Norm

al kid

ney 1

Norm

al kid

ney 2

BHD tum

or 1

BHD tum

or 2

BHD tum

or 3

Norm

al kid

ney 3

Norm

al kid

ney 4

BHD tum

or 3

BHD tum

or 4

Norm

al kid

ney 3

Norm

al kid

ney 4

BHD tum

or 3

BHD tum

or 4

F BHD tumor #5

Adj. normal kidney

Chromophobe (#3) Oncocytic Hybrid (#5)#3, #5FrameshiftYesYesYesFemale503

#2, #4Splicing defect

YesYesYesFemale642

Oncocytic Hybrid

#1Missensec.763C>T, His255Tyr, Exon7YesYesYesFemale641

HistologyTumor

#Mutation

TypeGermline BHD Gene MutationRenal

TumorsPulmonary

CystsFibrofolli-culomaGenderAgePatient #

c.1300G>C, Glu434Gln, Exon 11

HistologyTumor

#Mutation

TypeGermline BHD Gene MutationRenal

TumorsPulmonary

CystsFibrofolli-culomaGenderAge

Rel

ativ

e p-

AK

T

Rel

ativ

e p-

S6K

Rel

ativ

e p-

S6R

Oncocytic Hybrid

A C

c.1285insC, Exon11

Fig. 6. Activation of the PI3K-AKT-mTOR signaling pathway in kidney tumors from human BHD patients. (A) Western blotting was performed on protein lysatesof normal kidneys and kidney tumors from BHD patients, showing elevation of p-AKT, total AKT1, and total AKT2 in BHD tumors 1–3. (B–D) Protein expressionlevels of p-AKT(Ser-473) (B), p-S6K (Thr-421/Ser-424) (C), and p-S6R (Ser-240/244) (D) were quantified using the Meso Scale Discovery multiplex array system andwere consistent with the Western blotting results. The protein expression levels are relative to the value of normal kidney 3, which is defined as 1. (E and F)Immunofluorescence staining of p-AKT (Ser-473) was performed on frozen sections of human BHD tumors (F) and adjacent normal kidneys (E). Magnification,�40. (Scale bar, 50 �m.) (G) Information of BHD patients whose tumors were analyzed in this study.

18726 � www.pnas.org�cgi�doi�10.1073�pnas.0908853106 Hasumi et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0

Whole Mount in Situ Hybridization. The embryos were collected from wild-typeC57BL/6 mice intercrossed at different stages of gestation and processed forwhole mount in situ hybridization as previously described (25). Negativestaining was confirmed using a sense probe with wild-type embryos. Probeinformation is in the SI Methods.

Histological and Immunohistochemical Analysis. Embryo sections were stainedwith hematoxylin and eosin (H&E) or with DAB2 antibody (BD Bioscience) forimmunohistochemical evaluation performed as previously described (26). Theslides were read by at least three persons, including two pathologists (M.J.M.and D.C.H.).

Western Blotting and Antibodies. Immunoblotting was performed as describedin the SI Methods.

Tissue Genotyping by Southern Blotting. Nonradioactive Southern blotting wasperformed with DIG OMNI System for PCR Probes according to the manufac-turer’s protocol (Roche). Probe information is in the SI Methods.

Endogenous FLCN Detection by Duolink System. Duolink in situ PLA wasperformed per manufacturer’s instruction (OLink Biosciences). For furtherinformation, see the SI Methods.

Human Sample Preparation and MSD Analysis. Renal tumors were obtainedfrom BHD patients surgically treated at the Urologic Oncology Branch, Na-tional Cancer Institute, National Institutes of Health, Bethesda, MD withpatient permission under a National Institutes of Health Institutional ReviewBoard (IRB)-approved protocol #97-C-0147. All patients signed informed con-sent. MSD (Meso Scale Discovery) 96-well multispot AKT signaling pathway(phospho-AKT [Ser-473]/total GSK3�/phospho-S6K [Thr-421/Ser-424]), andphospho-S6R (Ser-240/244) assays were carried out according to the manu-facturer’s protocol. For further information, see the SI Methods.

Immunofluorescence Imaging of the AKT-mTOR Pathway. Immunostaining forp-AKT, p-mTOR, and p-S6R was performed on frozen sections as described inthe SI Methods.

ACKNOWLEDGMENTS. We thank Louise Cromwell for excellent technicalsupport with the mouse studies, Serguei Kozlov for helpful discussions, YelenaGolubeva for her technical support regarding laser microdissection (LCM),Jaime Rodriguez-Canales and Jeffrey Hanson for their technical expertiseregarding LCM procedures, and Kristin K. Biris for her technical supportregarding in situ hybridization. This research was supported by the IntramuralResearch Program of the National Institutes of Health (NIH), National CancerInstitute (NCI), Center for Cancer Research; federal funds from the NCI, NIH,under Contract HHSN261200800001E.

1. Zbar B, et al. (2002) Risk of renal and colonic neoplasms and spontaneous pneumo-thorax in the Birt-Hogg-Dubé syndrome. Cancer Epidemiol Biomarkers Prev 11:393–400.

2. Toro JR, et al. (1999) Birt-Hogg-Dubé syndrome: A novel marker of kidney neoplasia.Arch Dermatol 135:1195–1202.

3. Birt AR, Hogg GR, Dubé WJ (1977) Hereditary multiple fibrofolliculomas with tricho-discomas and acrochordons. Arch Dermatol 113:1674–1677.

4. Nickerson ML, et al. (2002) Mutations in a novel gene lead to kidney tumors, lung walldefects, and benign tumors of the hair follicle in patients with the Birt-Hogg-Dubésyndrome. Cancer Cell 2:157–164.

5. Pavlovich CP, et al. (2002) Renal tumors in the Birt-Hogg-Dubé syndrome. Am J SurgPathol 26:1542–1552.

6. Vocke CD, et al. (2005) High frequency of somatic frameshift BHD gene mutations inBirt-Hogg-Dubé-associated renal tumors. J Natl Cancer Inst 97:931–935.

7. Baba M, et al. (2006) Folliculin encoded by the BHD gene interacts with a bindingprotein, FNIP1, and AMPK, and is involved in AMPK and mTOR signaling. Proc NatlAcad Sci USA 103:15552–15557.

8. Hasumi H, et al. (2008) Identification and characterization of a novel folliculin-interacting protein FNIP2. Gene 415:60–67.

9. Takagi Y, et al. (2008) Interaction of folliculin (Birt-Hogg-Dubé gene product) with anovel Fnip1-like (FnipL/Fnip2) protein. Oncogene 27:5339–5347.

10. Shaw RJ, et al. (2004) The LKB1 tumor suppressor negatively regulates mTOR signaling.Cancer Cell 6:91–99.

11. Eng C (2003) PTEN: One gene, many syndromes. Hum Mutat 22:183–198.12. Inoki K, Corradetti MN, Guan KL (2005) Dysregulation of the TSC-mTOR pathway in

human disease. Nat Genet 37:19–24.13. Baba M, et al. (2008) Kidney-targeted Birt-Hogg-Dubé gene inactivation in a mouse

model: Erk1/2 and AKT-mTOR activation, cell hyperproliferation, and polycystic kid-neys. J Natl Cancer Inst 100:140–154.

14. Chen J, et al. (2008) Deficiency of FLCN in mouse kidney led to development ofpolycystic kidneys and renal neoplasia. PLoS One 3:e3581.

15. Bozulic L, Hemmings BA (2009) PIKKing on PKB: Regulation of PKB activity by phos-phorylation. Curr Opin Cell Biol 21:256–261.

16. Sabatini DM (2006) mTOR and cancer: Insights into a complex relationship. Nat RevCancer 6:729–734.

17. Copp J, Manning G, Hunter T (2009) TORC-specific phosphorylation of mammaliantarget of rapamycin (mTOR): Phospho-Ser2481 is a marker for intact mTOR signalingcomplex 2. Cancer Res 69:1821–1827.

18. Bielinska M, Narita N, Wilson DB (1999) Distinct roles for visceral endoderm duringembryonic mouse development. Int J Dev Biol 43:183–205.

19. Amin N, et al. (2009) LKB1 regulates polarity remodeling and adherens junctionformation in the Drosophila eye. Proc Natl Acad Sci USA 106:8941–8946.

20. Lium B, Moe L (1985) Hereditary multifocal renal cystadenocarcinomas and nodulardermatofibrosis in the German shepherd dog: Macroscopic and histopathologicchanges. Vet Pathol 22:447–455.

21. Okimoto K, et al. (2004) A germ-line insertion in the Birt-Hogg-Dubé (BHD) gene givesrise to the Nihon rat model of inherited renal cancer. Proc Natl Acad Sci USA101:2023–2027.

22. Hartman TR, et al. (2009) The role of the Birt-Hogg-Dubé protein in mTOR activationand renal tumorigenesis. Oncogene 28:1594–1604.

23. Guertin DA, et al. (2009) mTOR complex 2 is required for the development of prostatecancer induced by Pten loss in mice. Cancer Cell 15:148–159.

24. Facchinetti V, et al. (2008) The mammalian target of rapamycin complex 2 controlsfolding and stability of AKT and protein kinase C. EMBO J 27:1932–1943.

25. Biris KK, Dunty WC, Jr, Yamaguchi TP (2007) Mouse Ripply2 is downstream of Wnt3aand is dynamically expressed during somitogenesis. Dev Dyn 236:3167–3172.

26. Fazili Z, Sun W, Mittelstaedt S, Cohen C, Xu XX (1999) Disabled-2 inactivation is an earlystep in ovarian tumorigenicity. Oncogene 18:3104–3113.

Hasumi et al. PNAS � November 3, 2009 � vol. 106 � no. 44 � 18727

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0