Cancer resistance in the blind mole rat is mediatedby concerted necrotic cell death mechanismVera Gorbunovaa,1, Christopher Hinea,2, Xiao Tiana, Julia Ablaevaa, Andrei V. Gudkovb, Eviatar Nevoc,1,and Andrei Seluanova,1

aDepartment of Biology, University of Rochester, Rochester, NY 14627; bDepartment of Cell Stress Biology, Roswell Park Cancer Institute, Buffalo, NY 14263;and cInstitute of Evolution, University of Haifa, Haifa 31905, Israel

Contributed by Eviatar Nevo, October 3, 2012 (sent for review August 28, 2012)

Blind mole rats Spalax (BMR) are small subterranean rodents com-mon in the Middle East. BMR is distinguished by its adaptations tolife underground, remarkable longevity (with a maximum docu-mented lifespan of 21 y), and resistance to cancer. Spontaneoustumors have never been observed in spalacids. To understand themechanisms responsible for this resistance, we examined thegrowth of BMR fibroblasts in vitro of the species Spalax judaeiand Spalax golani. BMR cells proliferated actively for 7–20 popu-lation doublings, after which the cells began secreting IFN-β, andthe cultures underwent massive necrotic cell death within 3 d. Thenecrotic cell death phenomenon was independent of culture con-ditions or telomere shortening. Interestingly, this cell behaviorwas distinct from that observed in another long-lived and can-cer-resistant African mole rat, Heterocephalus glaber, the nakedmole rat in which cells display hypersensitivity to contact inhibi-tion. Sequestration of p53 and Rb proteins using SV40 large Tantigen completely rescued necrotic cell death. Our results suggestthat cancer resistance of BMR is conferred by massive necrotic re-sponse to overproliferation mediated by p53 and Rb pathways,and triggered by the release of IFN-β. Thus, we have identifieda unique mechanism that contributes to cancer resistance of thissubterranean mammal extremely adapted to life underground.

aging | necrosis

Blind mole rats, Spalax (BMR), belong to a family of sub-terranean rodents of the Muroidea superfamily, prevalent in

the Middle East, ranging in the Eastern Mediterranean andNorth Africa. BMRs are solitary mammals that spend their livesin underground burrows. Living under extreme hypoxic con-ditions, BMRs evolved strong hypoxic tolerance (1, 2).BMRs are very long-lived for their size. The maximum lifespan

documented for the animals kept in our animal facility is 21 y (3).In comparison, mice and rats belonging to the same superfamilyhave a maximum lifespan of 4 y (4, 5). Furthermore, BMRs showa striking resistance to cancer. Our observation of thousands ofcaptive animals did not show a single case of spontaneous tumordevelopment over a 40-y period. Cancer accounts for ∼23% ofhuman mortality (6). In mice and rats, cancer mortality is veryhigh, reaching 90% in some strains (7, 8).Animals have evolved multiple mechanisms to protect them-

selves from cancer. These mechanisms include cell-cycle check-points, DNA repair, programmed cell death, and replicativesenescence controlled by a network of tumor-suppressor genes,such as p53 and Rb. Anticancer adaptations differ between spe-cies, which may explain the differences in cancer susceptibility (9–11). Mice have been used extensively as models for cancer re-search. However, mice are more prone to cancer and are likely topossess fewer anticancer defenses, thus limiting a potential fordiscovery of novel anticancer pathways.Hence, there is great valuein studying anticancer mechanisms in cancer-resistant species.Subterranean mammals are good candidates for these studies.We showed previously that the BMR p53 gene contains an

arginine to lysine substitution at a site corresponding to humanp53 position 174, where mutations are often found in human

tumors (12). The R174K substitution affects the DNA-bindingdomain of p53; the resulting protein is capable of inducing cell-cycle arrest but is defective in initiating apoptosis. We hypoth-esized that R174K substitution evolved in BMR as an adaptationto the hypoxic environment in underground tunnels to preventhypoxia-induced apoptosis (12). Mice with the correspondingp53 mutation are defective in p53-dependent apoptosis (13).These mice show longer tumor latency than p53−/− mice, but aremore tumor-prone than the wild-type mice (13). Thus, the ex-treme tumor resistance of BMR is intriguing.Our previous studies of another long-lived subterranean rodent,

the naked mole rat (Heterocaphalus glaber), identified a novelanticancer mechanism termed early contact inhibition (ECI) (14).Naked mole rat cells proliferate slowly in culture and display hy-persensitivity to contact inhibition (14). Abrogation of ECI makesnaked mole rat cells more prone to malignant transformation.Naked mole rats inhabit East Africa and belong to the Hys-tricognath group of rodents, which also includes guinea pigs. Na-kedmole rats are evolutionarily distant fromBMR, which is closerto Old World mice and rats. Both mole rats, however, lead sub-terranean lifestyles protected from predators and the above-ground extreme climatic fluctuations, which allowed these speciesto evolve extreme longevity and cancer resistance.Here we set out to investigate cancer resistance in the BMR.

The analysis of BMR-cultured cells reveals that, unlike the nakedmole rat, BMR fibroblasts do not display early contact inhibition,but instead proliferate rapidly for 7–20 population doublings(PDs) after which the cells begin to secrete IFN-β and the cultureundergoes concerted cell death (CCD), which is characterized bya large fraction of necrotic cells. Inactivation of Rb and p53tumor suppressors rescued the cell-death phenotype. Our resultssuggest that the BMR has evolved a unique anticancer mecha-nism mediated by strong induction of the necrotic cell-deathresponse to hyperproliferation. Furthermore, we conclude thatthe two long-lived cancer-resistant rodents have achieved theircancer resistance by two distinct mechanisms.

ResultsBMR Fibroblasts Do Not Show ECI. To understand the molecularmechanism responsible for the cancer resistance of BMR, weisolated primary fibroblasts from BMR lung and skin and ex-amined their proliferation in culture. The experiments wereperformed on fibroblasts isolated from two animals of eachspecies: Spalax judaei and Spalax golani (2). All fibroblast linesshowed similar growth characteristics; therefore, we will refer tothem as BMR fibroblasts. BMR fibroblasts proliferated rapidly

Author contributions: V.G., A.V.G., E.N., and A.S. designed research; C.H., X.T., J.A., andA.S. performed research; A.V.G. and E.N. contributed new reagents/analytic tools; V.G.,E.N., and A.S. analyzed data; and V.G., E.N., and A.S. wrote the paper.

The authors declare no conflict of interest.1To whom correspondence may be addressed. E-mail: vera.gorbunova@rochester.edu,nevo@research.haifa.ac.il, or andrei.seluanov@rochester.edu.

2Present address: Department of Genetics and Complex Diseases, Harvard School of PublicHealth, Boston, MA 02115.

19392–19396 | PNAS | November 20, 2012 | vol. 109 | no. 47 www.pnas.org/cgi/doi/10.1073/pnas.1217211109

and reached high cell density on a plate (Fig. 1A). Cell densitywas similar to the density attained by mouse fibroblasts. Weconcluded that BMR fibroblasts, unlike naked mole rat fibro-blasts, do not display the ECI phenotype. Therefore, BMR islikely to have evolved a distinct anticancer mechanism.

BMR Cells Undergo Concerted Cell Death. BMR fibroblasts prolif-erated rapidly for 7–20 PDs, and then the cells arrested proliferationfor ∼3 d, after which all cells on the plate detached and died,

with no attached cells remaining within 4 d of the onset of celldeath. Our laboratory routinely cultures primary fibroblastsfrom 20 rodent species, and we have never observed sucha synchronous death of cell cultures. We termed this phenom-enon “concerted cell death” (CCD). Our standard growth mediafor all primary rodent fibroblasts is EMEM supplemented with15% (vol/vol) FBS. In an attempt to find growth conditionswhere the cells would not die, we have also cultured BMR cellsin low-serum and serum-free fibroblast growth media. However,variations of cell culture media changed the lifespan by afew PDs, after which all cultures invariably underwent CCD(Fig. 1B).To identify the type of cell death in BMR cultures, we col-

lected the dying cells, stained them with Annexin-V and propi-dium iodide (PI), and then analyzed by flow cytometry. The celldeath was preferentially necrotic, with 46% of the cells dying bynecrosis and 28% by apoptosis (Fig. 2B, control samples).

CCD Is Abrogated by SV40 Large T Antigen. We next tested whetherCCD is controlled by the major tumor suppressor pathways,

Sparse Confluent

Spalax

golani

Mus

musculus

Day 2 Day 7

A

0 20 40 60 80

20

16

12

8

4

0

Time, days

Po

pu

la

tio

n D

ou

blin

g

BMR

B 24

28 Mouse

NMR

1

2

3 4

5

6

7

Spalax

judaei

Fig. 1. BMR fibroblasts display a unique concerted cell death phenotype.All cells were cultured at 3% oxygen concentration. (A) Fibroblasts of bothS. judaei and S. golani species become confluent at high cell density anddo not undergo early contact inhibition typical of naked mole rat cells.Cells were seeded at 5 × 105 cells per 10-cm dish and photographed atindicated time points. (Magnification: 100×.) (B) BMR fibroblast culturesundergo CCD irrespective of growth conditions. Mouse, denotes mousefibroblasts, and NMR denotes naked mole rat fibroblasts cultured underour standard conditions of 15% FBS and EMEM media from ATCC. Bluegrowth curves correspond to BMR fibroblasts cultured in different con-centrations of FBS and either EMEM media from ATCC or low-serum fi-broblast growth media (FGM) from Lonza. FGM media was used with dif-ferent concentrations of fibroblast growth factor indicated by (no. inblue) -F or 1/2F (1), 15% FBS, FGM (2), 15% FBS, FGM-F (3), 2% FBS, FGM(4), 15% FBS, EMEM (5), 2% FBS, FGM 1/2F (6), 5% FBS, EMEM (7), 0% FBS,FGM. BMR cells grew rapidly for several PDs, then invariably stopped pro-liferation and underwent CCD.

A Large T

Control

Large T K1

0 20 40 60 80

30

20

10

0

Time, days

Po

pu

latio

n D

ou

blin

g

100 120

B

Day 47

Co

ntro

l

Larg

e T

60

40

20

0

Cell d

eath

, %

80

Day 67

Larg

e T

K

1

Co

ntro

l

Larg

e T

Larg

e T

K

1

Fig. 2. CCD requires Rb and p53 pathways. (A) At day 27 BMR cells werestably transfected with plasmids encoding SV40 LT antigen, Large T K1 (amutant version of LT antigen that inactivates p53 but not Rb), and a controlplasmid. Cells transfected with the wild-type LT did not undergo CCD. (B)CCD occurs by necrosis and apoptosis. Cell death was analyzed by Annexin-V/PI staining and flow cytometry in actively growing (day 47) and dying (day67) BMR cultures shown in A.

Gorbunova et al. PNAS | November 20, 2012 | vol. 109 | no. 47 | 19393

MED

ICALSC

IENCE

S

Rb and p53. We transfected BMR cells with SV40 large T an-tigen (LT) or its mutant derivatives and selected chromosomalintegrants. Wild-type LT is a viral oncoprotein that binds andinactivates both p53 and pRb. The mutant derivative LTK1inactivates only p53, and LTΔ434–444 inactivates only pRb andits family members (p107 and p130) (15). BMR cells expressingLT proliferated continuously for over 50 PDs and did not un-dergo CCD (Fig. 2). LTK1 did not rescue CCD, but rather ac-celerated its onset (Fig. 2), but LTΔ434–444 was toxic to the cellsand did not yield any chromosomally integrated clones despitemultiple attempts. This result indicates that both Rb and p53pathways must be inactivated to abrogate CCD.

CCD Is Not Caused by Telomere Shortening.We then tested whetherCCD is induced by rapid telomere shortening. Importantly,BMR cultures expressed high levels of telomerase activity (Fig.3A), as expected for a small-sized rodent (10). In addition, time-resolved fluorescence (TRF) assay showed that BMR has telo-meres with a mean length of 50 kb, which is much longer thantelomeres in the species with replicative senescence, such ashumans (Fig. 3B). Furthermore, telomere length did not changein cultured cells undergoing CCD. These results indicate thatCCD is not triggered by telomere shortening.

Concerted Cell Death of BMR Cells Is Mediated by Release of IFN-β.We then set out to identify the mechanism of CCD. We hy-pothesized that the pattern of cells dying synchronously is con-sistent with a response to IFN. We tested whether BMR cellcultures secrete IFN-β using two assays. In the first assay, HT1080cells were infected with vesicular stomatitis virus encoding a GFPgene (VSV-GFP). The number of cells containing green fluores-cent virus particles was quantified by flow cytometry. HT1080 cellswere incubated overnight with the media conditioned by BMRcells at different stages of growth. HT1080 cells were then infectedwith VSV-GFP, and the presence of IFN was quantified by theability of the conditioned media to halt VSV-GFP infection andreduce the number ofGFP+ cells (Fig. 4A). The number of units of

IFN-β was determined via a calibration curve built by adding hu-man IFN to VSV-GFP–infected HT1080 cells (Fig. 4B). Young,growing BMR cultures did not release any detectable IFN, but justbefore the onset of CCD BMR cells secreted high levels of IFN-β(Fig. 4A). In the second assay, BMR-conditionedmedia was addedto HEK-blue cells containing a β-galactosidase gene under anIFN-β–responsive promoter. Young, growing BMR cultures didnot induce β-galactosidase above the control level, but conditionedmedia from the dying cultures showed a threefold elevated ex-pression of the reporter gene. Notably, transfection with SV40 LTantigen that rescues CCD also abrogated IFN-β release (Fig. 4C).Finally, we tested whether BMR-conditioned media from

dying cells would induce CCD in young BMR cells or in cells forother species. Fresh media, media conditioned by young, grow-ing BMR cells, and media conditioned by dying BMR cells wereadded to young BMR, mouse, or human cells. Media condi-tioned by dying BMR cells caused massive necrotic cell death ofyoung BMR cells, and massive apoptotic death of mouse cells(Fig. 4D). Media conditioned by dying BMR cells did not causemassive death of human cells, which is likely explained by thedivergence between rodent and human interferons. In summary,these results indicate that CCD is triggered by IFN-β release.

DiscussionIn this study we show that cells of a cancer-resistant rodent,BMR, display an unusual cell death mechanism that rapidlywipes out entire cultures of cells, leaving no survivors. CCDoccurs by necrosis and is triggered by the release of INF-β. Thisresponse required functional p53 and Rb pathways. When cellsare cultured in vitro they are subjected to growth factors fromFBS and are forced to proliferate because of frequent subcultureat low cell density. These strong progrowth signals may be rec-ognized by BMR cells as a potentially oncogenic signaling im-balance and trigger a cell death response. This mechanismresembles oncogene-induced senescence of human cells, wherecells enter irreversible arrest in response to hyperproliferativesignals, such as expression of activated oncogenes (16, 17). An-other similarity with oncogene-induced senescence is that CCDof BMR cells is independent of telomere shortening and occursin cells with long telomeres. We hypothesize that in vivo CCDefficiently clears away premalignant cells contributing to thecancer resistance of the BMR. Cell culture experiments revealedthat a similar necrotic cell death takes place in primary BMR cellcultures upon extensive in vitro proliferation.Strikingly, BMR cells use necrosis rather than apoptosis to on-

cogenic insults. This finding can be explained by the unusual se-quence of BMR p53, which is deficient in activating an apoptosiscascade, and evolved as an adaptation to subterranean life underhypoxic conditions (12). Nevertheless, necrotic death of BMRcells requires a functional p53 pathway. Despite necrosis beingcommonly viewed as less precise or as an inefficient way of elim-inating unwanted cells, BMRs have evolved a highly efficient an-titumor mechanism based on necrotic response. An advantage ofnecrosis could be in eliminating all cells surrounding the pre-malignant lesion, which may provide an added antitumor effectby eliminating reactive tumor stroma, including tumor-activatedfibroblasts (18).In the future, it would be interesting to move these studies into

in vivo system by testing whether CCD protects BMRs fromchemically induced carcinogenesis. It would also be interesting tomix BMR cells with malignantly transformed mouse cells andtest whether BMR cells suppress carcinogenesis in a xenograftmouse model.Remarkably, cells of another cancer-resistant subterranean

rodent, the naked mole rat, do not display CCD, and achievehigh PDs in culture (11). Furthermore, naked mole rat cellseasily undergo apoptosis and do not favor necrosis in response tostress (14). This finding is consistent with the naked mole rat p53

HeL

a

G D

BMR A B

100 150 kb

50 33 25 19 15 12 10

8 7 6 5 4

G D

BMR

Y S

Human

Fig. 3. CCD of BMR cells is not caused by telomere shortening. (A) BMRfibroblasts have endogenous telomerase activity. Telomeric repeat amplifi-cation protocol assay was performed on extracts from growing (G) or dying(D) BMR cells, and HeLa cells as a positive control. (B) Telomeres do notshorten in dying BMR cells. Telomere length in growing (G) and dying (D)BMR cells was measured using TRF assay. Young (Y) and senescent (S) humanfibroblasts were used as a reference.

19394 | www.pnas.org/cgi/doi/10.1073/pnas.1217211109 Gorbunova et al.

gene, which has arginine in position 174 (19), and is thus similarto the mouse or human p53 in its propensity to induce apoptosis.Naked mole rat cells display early contact inhibition that acts asan additional tumor suppressor mechanism in this species. Incontrast BMR cells proliferate to high cell density. Thus, BMRsand naked mole rats have evolved two distinct anticancermechanisms that provide remarkable cancer resistance to thesesmall rodents.Our earlier comparative studies of 15 rodent species suggested

that small, long-lived rodents evolve novel anticancer mecha-nisms (10, 11). The current in-depth analysis of two small, long-lived species of subterranean rodents belonging to different

superfamilies have shown that these mechanisms are distinct foreach species, as they evolved phylogenetically independently ineach species, possibly associated with hypoxic conditions un-derground and the species longevity. We speculate that manyunique anticancer adaptations will be found in long-lived rodentspecies. This knowledge could then be used for treatment orprevention of cancer in humans.

Experimental ProceduresCell Isolation and Culture. Primary blind mole rat (BMR) and mouse fibroblastswere isolated from lung and under arm skin, as described previously (20). Theexperiments were performed on fibroblasts isolated from two animals of

0 10 30 40 50

12

8

4

0

Time, days

Po

pu

la

tio

n D

ou

blin

g

60 70 C

on

tro

l

hIN

T-

0.6

0.4

0.2

0 HE

K-b

lu

e

-g

al a

ctiv

ity, O

D 0.8

BM

R g

ro

win

g

BM

R d

yin

g

BM

R LT

16

20

120

80

40

0

In

te

rfe

ro

n-

, U

nits

/m

l 160

A

C

BMR

Co

ntro

l

Gro

win

g

60

40

20

0

Ce

ll d

ea

th

, %

80

Dyin

g

Mouse

Co

ntro

l

Gro

win

g

Dyin

g

Human

Co

ntro

l

Gro

win

g

Dyin

g

Conditioned

media:

D

20

80

0

20

60

80

10

0

30

20

10

0

GF

P, %

12

0

14

0

40

40

Interfiron- , Units/ml

B

50

16

0

18

0

60

Control Dying

BMR

Mouse

E

Fig. 4. CCD is triggered by release of IFN-β by BMR cells. (A) IFN release coincides with the onset of CCD. The growth curve for BMR cells is overlaid on the bargraph showing IFN levels in the media. IFN levels were measured by VSV-GFP assay. In this assay HT1080 cells are incubated with fresh media, or conditionedmedia containing INT-β; the cells are then infected with a GFP-encoding VSV, and the level of IFN in the test media corresponds to the reduction in thenumber of GFP positive HT1080 cells. (B) Calibration curve used to quantify the amount of BMR IFN in A. The curve was built by performing a VSV assay withknown amounts of human INT-β. (C) IFN release by dying BMR cells determined by HEK-blue assay, which measured the induction of β-gal reporter in HEK cellsunder IFN-inducible promoter. Control indicates untreated HEK-blue cells. hINT-β are HEK-blue cells to which 300 U/mL of human IFN-β was added. Other barsshow HEK-blue cells treated with BMR-conditioned media. (D) Conditioned media from dying BMR cells causes death of mouse cells. Media conditioned byBMR cells was added to growing BMR, mouse, or human fibroblasts. Cell death was quantified by Annexin-V/PI staining and flow cytometry. (E) Microscopicexamination of CCD. Photographs of early passage mouse and BMR cells before and after addition of conditioned media from dying BMR cells. (Magnifi-cation: 100×.) Upon addition of the BMR conditioned media mouse cells display typical apoptosis with membrane blebbing, but BMR cells show pre-dominantly necrotic cell death with cells detouching from the plate and breaking up without membrane bebbing.

Gorbunova et al. PNAS | November 20, 2012 | vol. 109 | no. 47 | 19395

MED

ICALSC

IENCE

S

each BMR species: Spalax judaei and Spalax golani (2) and three mice Musmusculus. All BMR fibroblast isolates showed similar growth characteristics.

Normal human fibroblasts were HCA2 neonatal foreskin fibroblasts.Human Fibrosarcoma (HT-1080) cells were purchased from American TypeCulture Collection (ATCC CCL-121).

Under standard conditions, all primary fibroblasts were cultured at 37 °C,5% CO2, 3% O2, on treated polystyrene culture dishes (Corning) in EMEMmedia (ATCC) supplemented with 15% FBS (Gibco), nonessential aminoacids, sodium pyruvate, 100 units/mL penicillin, and 100 μg/mL streptomycin(Gibco) or FGM-2 BulletKit (Lonza) supplemented with 15% FBS (Gibco).

For experiments with different types of media (Fig. 1B), BMR cells werecultured in different concentrations of FBS (0%, 2%, or 15%) and eitherEMEM media from ATCC or low-serum fibroblast growth media (FGM) fromLonza. FGM media was used with different concentrations of fibroblastgrowth factor. In addition, we unsuccessfully tried to grow BMR cells in lowserum growth medium 106 (Gibco) supplemented with different concen-trations of FBS (0%, 2%, or 15%) and with different concentrations of lowserum growth supplements (Gibco).

We did not observe significant differences in the growth of BMR cells ondifferent surfaces, including treated polystyrene culture dishes (Corning), BDPureCoat Amine (positively charged surface), BD PureCoat Carboxyl (nega-tively charged surface), collagen, fibronectin, gelatin, and glass.

Analysis of Cell Growth. Cells were seeded at 5 × 105 cells per 100-mm dish.When cells reached 80% confluence, they were harvested, counted, and thenumber of population doublings was calculated.

Analysis of Telomeres. Telomeric repeat amplification protocol and time-re-solved fluorescence assays were performed as previously described (11).

Analysis of Cell Death. Cell death was analyzed using Annexin-V-FLUOS assaykit (Roche) according to the manufacturer’s instructions. Briefly, floating andadherent cells were harvested, stained with Annexin-V and propidium io-dide, and analyzed on a BD Biosciences FACS Canto flow cytometer.

Transfections. Fibroblasts were seeded at 5 × 105 cells/100-mm plate 2 d be-fore transfection. For transfection, cells were harvested, counted, and 106

cells were transfected with 5 μg of plasmid DNA using Amaxa Nucleofector IIon program U-020 and solution NHDF (Amaxa). Transfections includeda mock with no DNA, pSG5 Large T (Addgene 9053), pSG5 Large T K1(Addgene 9055), pSG5 Large TΔ434–444 (Addgene 9054), and no electro-poration control.

IFN-β Assays. Vesicular stomatitis virus encoding a GFP gene assay. Vesicularstomatitis virus encoding a GFP gene (VSV-GFP) was previously described (21).It consists of VSV with GFP linked to the cytoplasmic domain of the VSV-Gprotein. The addition of GFP to VSV-G does not cause a reduction in titer, isstable through multiple passages, and is incorporated into virions with al-most the same efficiency as wild-type VSV-G. The virus was expanded asfollows: 200 μL of the virus was added onto 1.5 × 106 HeLa cells plated2 d earlier; 48 d after infection, media with the virus was collected, filtered,and stored at −80 °C. The titer of the virus was determined for individualcell lines using standard plaque assay. Aliquots of VSV stock were stored at−80 °C until use.

For the experiments, the viruses were thawed slowly on ice and dilutedin EMEM medium to obtain a working stock of 100 pfu/μL. VSV was usedaccording to bio-safety procedures in a P2-level safety facility/room. The useof VSV-GFP in this project was reviewed and approved by the InstitutionalBiosafety Committee at the University of Rochester.

For the assays, media conditionedwith BMRs at different stages of growth,mouse, or human HCA2 fibroblasts were collected after 3–5 d of growingcells, spinned down to remove debris, and stored at +4 °C. For the VSV-GFPinfection we used HT1080 cells, that were seeded at a density of 60,000 cellsper well in six-well plates in EMEM with 15% FBS and incubated for 24 h.Then, the media was discarded and the conditioned media was added toeach well. In addition, we used human IFN-β human (100 U per well) asa positive control and human IFN-γ (100 U per well) as a negative control.After 3–4 h of incubation in conditioned media, cells were infected withVSV-GFP at a multiplicity of infection 0.1 or 1 for a 12–16 h period. GFP signalappears in infected cells after 12–16 h, and the cells die about 6 h later. FACSanalysis was performed 16 h after infection to quantify GFP+ cells.HEK-Blue cells assay. HEK-Blue IFN-α/β cells (Invivogen) allow the detection ofbioactive human type I IFNs by monitoring the activation of the ISGF3pathway. The assay was performed according to the manufactured protocol(Invivogen). Briefly, HEK-Blue cells were plated in 96-well plates and in-cubated overnight. Then, 20 μL of test medium was added to each well.Human IFN-β was used as a positive control and human IFN-γ was used asa negative control. Cells were incubated overnight. The next day, 30 μL ofthe supernatant from Hek-Blue cells was added to 170 μL of Quanti-BlueReagent (Invivogen) for 3 h at 37 °C. The colorimetric reaction was measuredat 650 nm on a plate reader.

ACKNOWLEDGMENTS. This work was supported by grants from the NationalInstitutes of Health and Ellison Medical Foundation (to V.G.); an EllisonMedical Foundation grant (to A.S.); and the Ancell-Teicher Research Foun-dation of Genetics and Molecular Evolution (E.N.).

1. Nevo E (1999) Mosaic Evolution of Subterranean Mammals: Regression, Progression,and Global Convergence (Oxford Univ Press, London).

2. Nevo E, Ivanitskaya I, Beiles A (2001) Adaptive Radiation of Blind Subterranian MoleRats (Backhuys, Leiden, The Netherlands).

3. Edrey YH, et al. (2012) Sustained high levels of neuregulin-1 in the longest-lived ro-dents; A key determinant of rodent longevity. Aging Cell 11(2):213–222.

4. Turturro A, et al. (1999) Growth curves and survival characteristics of the animals usedin the Biomarkers of Aging Program. J Gerontol A Biol Sci Med Sci 54(11):B492–B501.

5. de Magalhães JP, Costa J, Toussaint O (2005) HAGR: The Human Ageing GenomicResources. Nucleic Acids Res 33(Database issue):D537–D543.

6. Heron M (2012) Deaths: Leading Causes for 2008, National Vital Statistics Reports60(6):9–11.

7. Lipman R, Galecki A, Burke DT, Miller RA (2004) Genetic loci that influence cause ofdeath in a heterogeneous mouse stock. J Gerontol A Biol Sci Med Sci 59(10):977–983.

8. Burek JD, Hollander CF (1977) Incidence patterns of spontaneous tumors in BN/Bi rats.J Natl Cancer Inst 58(1):99–105.

9. Wright WE, Shay JW (2000) Telomere dynamics in cancer progression and prevention:Fundamental differences in human and mouse telomere biology. Nat Med 6(8):849–851.

10. Seluanov A, et al. (2007) Telomerase activity coevolves with body mass not lifespan.Aging Cell 6(1):45–52.

11. Seluanov A, et al. (2008) Distinct tumor suppressor mechanisms evolve in rodentspecies that differ in size and lifespan. Aging Cell 7(6):813–823.

12. Ashur-Fabian O, et al. (2004) Evolution of p53 in hypoxia-stressed Spalax mimics

human tumor mutation. Proc Natl Acad Sci USA 101(33):12236–12241.13. Liu G, et al. (2004) Chromosome stability, in the absence of apoptosis, is critical for

suppression of tumorigenesis in Trp53 mutant mice. Nat Genet 36(1):63–68.14. Seluanov A, et al. (2009) Hypersensitivity to contact inhibition provides a clue to

cancer resistance of naked mole-rat. Proc Natl Acad Sci USA 106(46):19207–19208.15. Hahn WC, et al. (2002) Enumeration of the simian virus 40 early region elements

necessary for human cell transformation. Mol Cell Biol 22(7):2111–2123.16. Di Micco R, Fumagalli M, d’Adda di Fagagna F (2007) Breaking news: High-speed

race ends in arrest—How oncogenes induce senescence. Trends Cell Biol 17(11):

529–536.17. Serrano M, Lin AW, McCurrach ME, Beach D, Lowe SW (1997) Oncogenic ras provokes

premature cell senescence associated with accumulation of p53 and p16INK4a. Cell 88

(5):593–602.18. Mueller MM, Fusenig NE (2004) Friends or foes—Bipolar effects of the tumour stroma

in cancer. Nat Rev Cancer 4(11):839–849.19. Kim EB, et al. (2011) Genome sequencing reveals insights into physiology and lon-

gevity of the naked mole rat. Nature 479(7372):223–227.20. Seluanov A, Vaidya A, Gorbunova V (2010) Establishing primary adult fibroblast

cultures from rodents. J Vis Exp (44)pii:2033.21. Diaz RM (2007) Oncolytic immunovirotherapy for melanoma using vesicular stoma-

titis virus. Cancer Res 67(6):2840–2848.

19396 | www.pnas.org/cgi/doi/10.1073/pnas.1217211109 Gorbunova et al.