p16INK4a Impairs Homologous Recombination−Mediated DNA ...scholar.harvard.edu/.../p16ink4a_impairs_homologous_recombinatio… · impaired the recruitment of RAD51 to damaged DNA

2014;74:1739-1751. Published OnlineFirst January 28, 2014.Cancer Res Rüveyda Dok, Peter Kalev, Evert Jan Van Limbergen, et al.

Positive Head and Neck Tumors−Repair in Human Papillomavirus Mediated DNA−p16INK4a Impairs Homologous Recombination

Updated version

10.1158/0008-5472.CAN-13-2479doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2014/01/28/0008-5472.CAN-13-2479.DC1.html

Access the most recent supplemental material at:

Cited Articles

http://cancerres.aacrjournals.org/content/74/6/1739.full.html#ref-list-1

This article cites by 49 articles, 17 of which you can access for free at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

[email protected]

To request permission to re-use all or part of this article, contact the AACR Publications Department at

on August 28, 2014. © 2014 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 28, 2014; DOI: 10.1158/0008-5472.CAN-13-2479



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-13-2479

http://cancerres.aacrjournals.org/content/suppl/2014/01/28/0008-5472.CAN-13-2479.DC1.html

http://cancerres.aacrjournals.org/content/74/6/1739.full.html#ref-list-1

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

mailto:[email protected]

http://cancerres.aacrjournals.org/


Molecular and Cellular Pathobiology

p16INK4a Impairs Homologous Recombination–MediatedDNA Repair in Human Papillomavirus–Positive Head andNeck Tumors

R€uveyda Dok1, Peter Kalev2,5, Evert Jan Van Limbergen1,6, Layka Abbasi Asbagh3, Iria V�azquez2,5,Esther Hauben4,7, Anna Sablina2,5, and Sandra Nuyts1,6

AbstractThe p16INK4a protein is a principal cyclin-dependent kinase inhibitor that decelerates the cell cycle.

Abnormally high levels of p16INK4a are commonly observed in human papillomavirus (HPV)–positive headand neck squamous cell carcinomas (HNSCC). We and others found that p16INK4a overexpression is associatedwith improved therapy response and survival of patients with HNSCC treated with radiotherapy. However, thefunctional role of p16INK4a in HNSCC remains unexplored. Our results implicate p16INK4a in regulation ofhomologous recombination–mediated DNA damage response independently from its role in control of the cellcycle. We found that expression of p16INK4a dramatically affects radiation sensitivity of HNSCC cells. p16INK4aoverexpression impairs the recruitment of RAD51 to the site of DNA damage in HPV-positive cells by down-regulating of cyclin D1 protein expression. Consistent with the in vitro findings, immunostaining of HNSCCpatient samples revealed that high levels p16INK4a expression significantly correlated with decreased cyclin D1expression. In summary, these findings reveal an unexpected function of p16INK4a in homologous recombi-nation–mediated DNA repair response and imply p16INK4a status as an independentmarker to predict responseof patients with HNSCC to radiotherapy. Cancer Res; 74(6); 1739–51. �2014 AACR.

IntroductionHead and neck squamous cell carcinoma (HNSCC) is the

sixth most common cancer worldwide and responsible forapproximately half a million cases every year (1). Tobacco andalcohol consumption are two environmental risk factors asso-ciated with the development of this group of cancers. Recentepidemiologic studies have also revealed that infection withhigh-risk humanpapillomavirus (HPV) is etiologically linked toHNSCC pathogenesis (2, 3). HPV-positive HNSCCs, which arepredominantly found within the oropharyngeal regions withan occurrence of 25% to 47% (4), are nowwidely recognized as adistinct HNSCC entity due to demographic, histologic, clinical,and molecular differences (5).

The HPV oncogene products, E6 and E7, play a key role inHPV-associated carcinogenesis by inactivating of p53 andretinoblastoma tumor suppressor functions. The functionalinhibition of retinoblastoma by HPV-E7 results in high expres-sion levels of the cyclin-dependent kinase (CDK) inhibitorp16INK4a. The p16INK4a, encoded by the CDKN2A tumorsuppressor gene, is one of the core components of the cellcycle that acts by disrupting the complex between cyclin D1and CDK4/6. The cyclin D1/CDK4/6 complex phosphorylatesthe retinoblastoma protein that results in the release of thetranscription factor E2F and the initiation of the cell-cycleprogression. Conversely, the retinoblastoma/E2F repressorcomplex inhibits transcription of several genes, includingCDKN2A (6, 7). By inactivating retinoblastoma, HPV-E7releases the CDKN2A gene from its transcriptional inhibitionthat results in p16INK4a overexpression. p16INK4a overex-pression has been recently proposed to serve as a surrogatemarker of HPV infection in clinical samples because of a strongcorrelation between HPV and p16INK4a expression status inHNSCCs (8).

Despite the differences between HPV-positive and HPV-negative HNSCC tumors, currently all locally advanced pati-ents with HNSCC are treated uniformly with radiotherapy andconcurrent chemotherapy or surgery followed by (chemother-apy) radiotherapy (9, 10). Recent clinical trials have revealedthat patients with HPV-infected tumors show markedlyimproved treatment response to conventional therapy andhave a better prognosis compared with patients with HPV-negative HNSCC (3, 8). Although recent studies showed thatHPV-positive HNSCC cells are more sensitive to ionizing

Authors' Affiliations: 1Department of Oncology, Laboratory of Experimen-tal Radiotherapy; 2Department of Human Genetics, Laboratory forMechanisms of Cell Transformation; 3Department of Oncology, Molecularand Digestive Oncology; 4Department of Imaging and Pathology, Trans-lational Cell and Tissue Research, KU Leuven, University of Leuven; 5VIBCenter for the Biology of Disease; Departments of 6Radiation Oncology;and 7Pathology, UZ Leuven, Leuven, Belgium

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Current address for Peter Kalev: Department of RadiationOncology, Dana-Farber Cancer Institute, Boston, Massachusetts.

Corresponding Author: R€uveyda Dok, Department of Oncology, KU Leu-ven, Herestraat 49 Box 815, 3000 Leuven, Belgium. Phone: 32-1634-6280/32-1637-2133; Fax: 32-1634-7623; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-13-2479

�2014 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org 1739




radiation (11, 12), the molecular mechanism and the functionof p16INK4a in the enhanced sensitivity to ionizing radiation ofthese tumors remains largely unknown.

Here, we assessed the impact of p16INK4a on DNA damagerepair response in HNSCC. We found that p16INK4a hampersDNA damage response independently from its ability to inhibitCDK4/6 activity or control cell-cycle progression. Decreasedexpression of cyclin D1 in p16INK4a-overexpressing cellsimpaired the recruitment of RAD51 to damaged DNA andimpeded the homologous recombination–mediated DNArepair. These findings may result in novel therapeutic applica-tions for p16INK4a-overexpressing and HPV-positive tumors.

Material and MethodsImmunohistochemistry and HPV detection

For immunostaining, routine formalin-fixed paraffin-embedded (FFPE) tumor tissues were available of 87 patientswith oropharyngeal obtained before treatment with (chemo-therapy) radiotherapy at the University Hospitals Leuven(Leuven, Belgium). The human tumor samples were acquiredaccording protocols approved by the Ethic Board of theUniversity Hospitals (Leuven, Belgium). Briefly, 5-mm FFPEtumor sections were deparaffinized with the DAKO PT linkmodule and treated with primary antibodies against p16INK4a(G175-405; BD Pharmingen) or cyclin D1 (clone EP12; Dako).Immunodetection was performed with EnVision detectionsystems peroxidase/DAB (Dako), and sections were counter-stained with hematoxylin.

DNA was extracted from FFPE tumor tissues using theQIAMP DNA FFPE Kit according to the manufacturer's pro-tocol. The quality of extracted DNA to use for HPV PCR waschecked by control amplification of 167 basepair fragment ofthe isocitrate dehydrogenase 2 (IDH2) gene using PCR. HPVpositivity was determined by high-risk HPV–DNA-specific PCRreaction using GP5þ/GP6þ primer set as described previously(13).

Cell lines and reagentsSCC154 cell line (received 2011) was purchased from the

German collection of microorganisms and cell cultures(DSMZ), authentication was performed by DSMZ by shorttandem repeat testing. Cells were used within 6 months ofreceipt after resuscitation per the DSMZ recommendation.SCC090 cell line (received 2010) was a generous gift of Dr. S.Collin, the University of Pittsburgh (Pittsburgh, PA). SCC154and SCC090 cells were cultured and maintained in minimumessential media supplemented with 10% FBS, 1% L-glutamine,and 1% nonessential amino acids. SQD9, SC263, and CAL27 celllines (received 2008) were a generous gift of Dr. A. Begg, theNetherlands Cancer Institute (Amsterdam, the Netherlands)and were cultured as previously described (14). No authenti-cation was performed in our laboratory for these cells.

Lentiviral short hairpin RNAs (shRNA) against luciferase orp16INK4a (TRCN0000255853 and TRCN0000265833) were pur-chased from Sigma-Aldrich. Lentiviral infections were per-formed as previously described (15). Infected cells were select-ed by treatment with 4-mg puromycin (Sigma-Aldrich) for 4days. siRNA-targeting p16INK4a was 50AGAACCAGAGAGG-

CUCUGAtt-30 (sip16-1) and 50CGCACCGAAUAGUUACGGUtt-30 (sip16-2; Applied Biosystems). Scrambled siRNA (SilencerNegative Control; Applied Biosystems) was used as control.

Cyclin D1 plasmid (pCMV–cyclin D1) was obtained fromAddgene (plasmid 19927), and pBabe HPV-E7 was a generousgift from Dr. K. Munger, Harvard University (Cambridge, MA).pLA-p16INK4a was obtained by cloning INK4a in a lentiviralpLA vector. Transient transfections were performed usinglipofectamine 2000 (Life technologies) according to the man-ufacturer's protocol.

DNA damage was induced by ionizing radiation using alinear accelerator (6 MV photons; Varian Medical Systems).PD-0332991 (Pfizer) was obtained from Selleckchem.

Colony formation and cell viability assayTwenty-four hours upon drug (PD-0332991) treatment

where after the drug is replaced by fresh full media or 72 hoursafter transfectionwith siRNAs, cells were exposed to increasingdose of ionizing radiation (2–10 Gy) and plated into 10-cmdishes. After 2 to 3 weeks cells were fixed with 2.5% glutaral-dehyde in PBS and stainedwith 0.2% crystal violet. The coloniescontaining 50 cells or more were counted with ColCountcolony counter (Oxford Optronix). Survival fractions werecorrected for the plating efficiencies.

For cell viability assay cells were seededwith 20% confluencein 96-well plates and treated with drug or indicated doses ofionizing radiation and after 7 days a short-term survival assay(sulforhodamine B assay) was performed as previouslydescribed (14).

Nuclear extraction, immunoblotting, andcoimmunoprecipitation

For nuclear extraction and immunoprecipitation experi-ments a radiation dose of 5Gywas used as previously described(16). Whole-cell lysates were prepared in a lysis buffer (CellSignaling Technology) containing protease and phosphataseinhibitors (Roche). Nuclear extracts were prepared in radio-immunoprecipitation assay (RIPA) buffer containing proteaseand phosphatase inhibitors (Roche). Protein concentrationswere determined byBradford assay (Bio-Rad) and 10 to 50mg ofproteins were subjected to SDS–PAGE. Immunoblotting wasperformed with antibodies recognizing b-actin (#4967; CellSignaling Technology), RAD51 (clone 213; Abcam), cyclin D1(clone A-12; Santa Cruz Biotechnology), p16INK4a (cloneG175-405; BD Pharmingen), vinculin (clone hVIN-1; Sigma-Aldrich), and a-tubulin (clone 6-11B; Sigma-Aldrich).

The soluble nuclear fraction and insoluble chromatin frac-tion were extracted as previously described (17, 18). Thechromatin fraction was sonicated before gel electrophoresisand immunoblotting.

For coimmunoprecipitations, cells were lysed with RIPAbuffer. Two micrograms of rabbit anti-cyclin D1 antibody(clone H295; Santa Cruz Biotechnology) was incubated with1.5mg of cells lysate overnight at 4�C, and then protein A beadswere added for 1 hour at 4�C. Immunoprecipitated proteinswere detected by immunoblotting using antibodies specific toRAD51 (clone ab213; Abcam) and CDK4 (clone DCS156; CellSignaling Technology).

Dok et al.

Cancer Res; 74(6) March 15, 2014 Cancer Research1740




Immunofluorescence and light microscopyImmunofluorescence was performed as previously descri-

bed (18). Briefly, the cells were plated on mClear 96-well plates(Greiner Bio-One) and fixed with 4% paraformaldehyde. Afterpermeabilization with methanol, primary antibodies againstRAD51 (clone ab213; Abcam or clone H-29; Santa Cruz Bio-technology) were followed by a secondary antibody conjugatedto fluorescein isothiocyanate (Life Technologies). The DNAwas counterstainedwith 40,6-diamidino-2-phenylindole (DAPI;Sigma-Aldrich). The immunofluorescence images wereacquired using In Cell Analyzer 2000 (GE Healthcare).For micronuclei analysis cells were irradiated with 6 Gy and

fixed with 4% paraformaldehyde. Cells were stained with DAPIand analyzed by In Cell Analyzer 2000 (GE Healthcare). Thefrequencies of micronuclei were scored by counting the per-centage of micronuclei in at least 1,000 cells per experiment.For immunohistochemical analysis a light microscope

(Olympus) with �200 and �400 magnification was used andpictures were taken by Cell^D software (Olympus).

Cell-cycle analysisFor cell-cycle analysis cells were irradiated with doses of 4 to

6 Gy as previously described (19), cells were fixed with 70%ethanol and stained with 10 mg/mL propidium iodide contain-ing 100 mg/mL RNase A. Bromodeoxyuridine (BrdUrd) analysiswas performed with the BD Pharmingen BrdUrd Flow Cyto-metry Kit and the Cell Proliferation ELISA BrdUrd Kit (RocheApplied Science) according to the manufacturer's instruction.Cell-cycle distribution was assessed by FACSCanto (Becton-Dickinson) and FlowJo (Treestar).

Statistical analysisSurvival rates were determined on the basis of Kaplan–

Meier estimates using Statistica software version 11. Thelog-rank test was used to compare survival intervals. Toexamine association in categorical data the c2 test and thenonparametric-rank test were used. For continuous vari-ables, the unpaired Student t test or the nonparametric-ranktest were used. All tests were considered statistically signif-icant for P � 0.05.

ResultsHigh levels of p16INK4a expression correlate with bettertherapy response in patients with HNSCCRecent studies have demonstrated that HPV positivity in

HNSCC is strongly associated with p16INK4a overexpression.To expand these observations, we performed a parallel analysisof HPV status and p16INK4a nuclear staining in a cohort of 87patients with HNSCCs treated with radiotherapy or chemor-adiotherapy at the University Hospitals of Leuven (Leuven,Belgium). We assessed p16INK4a expression by immunohis-tochemical analysis in 87 tumor biopsies obtained beforetreatment (Fig. 1A). All tumors were classified dichotomouslyas either p16INK4a-positive (strong, diffuse nuclear staining inmore than 10% ofHNSCC cells) or p16INK4a-negative (Fig. 1A).Consistent with previous reports (8, 20), 37% of the analyzedpatients demonstrated high levels of nuclear p16INK4a immu-nostaining (Fig. 1B).

In parallel, we determined HPV status by PCR detection ofhigh-risk HPV–DNA using GP5þ/GP6þ primers in all 87patients of the Leuven cohort. However, we were not able todetect HPV status in 33 patients due to insufficient material orpoor quality of the material (see Materials and Methods).About 25% of HNSCCs were found to be HPV-positive, whichis the average HPV prevalence in Europe. We observed asignificant association betweenHPVpositivity and overexpres-sion of p16INK4a (P value < 0.001; Yates c2 test; Fig. 1B). Thesame tendencywas observed for the available HNSCC cell lines.HPV-positive HNSCC cell lines, SCC154 and SCC090, had highlevels of p16INK4a expression, whereas we were not able todetect p16INK4a expression inHPV-negative SQD9, SC263, andCAL27 cell lines (Fig. 1C).

A recent analysis of a cohort of patients treated solely withconventional radiotherapy in the Danish Head and NeckCancer Group 5 trial revealed that high levels of p16INK4aexpression are associated with better treatment response andsurvival rate of patientswithHNSCC treatedwith conventionalradiotherapy. The p16INK4a expression has been shown to bean even stronger independent prognostic factor than theclassical clinical parameters of tumor stage and nodal status(20). Consistent with this report, analysis of the Leuven cohortconfirmed that high levels of p16INK4a nuclear expression areassociated with improved locoregional control of patients withHNSCC (Fig. 1D). Together, these observations indicate thatp16INK4a overexpression has a major impact on treatmentresponse and survival in patients with HNSCC treated withconventional radiotherapy.

p16INK4a sensitizes HPV-infected HNSCCs to ionizingradiation independent from its ability to regulate cellcycle

We next examined the effect of p16INK4a overexpression oncell sensitivity to ionizing radiation. A colony formation assayof irradiated cells normalized for plating efficiencies revealedthat p16INK4a-positive SCC154 and SCC090 HNSCC cell linesare more sensitive to ionizing radiation compared withp16INK4a-negative SQD9 HNSCC cells (Fig. 2A and Supple-mentary Fig. S1A). A short-term cell survival assay furtherconfirmed that cell lines with lower levels of p16INK4a expres-sion were more resistant to ionizing radiation than cell lineswith high levels of p16INK4a (Fig. 2B).

To assess the contribution of p16INK4a to radiosensitivity,we suppressed p16INK4a expression in p16INK4a-positiveSCC154 HNSCC cell line by lentiviral shRNA against p16INK4a.We found that p16INK4a knockdown did not alter the colonyformation capacity of SCC154 cells (Supplementary Fig. S1B).In contrast, suppression of p16INK4a led to a significantincrease in survival of the irradiated cells (Fig. 2C and D). Wealso observed an increased survival of HPV-positive SCC090cells upon ionizing radiation whenwe suppressed p16INK4a inthese cells by the introduction of siRNA-targeting p16INK4a(Fig. 2E and Supplementary Fig. S1C).

The presence ofHPV,which inhibits retinoblastoma activity,in p16INK4a-overexpressing cells makes it unlikely thatp16INK4a affects cell survival by affecting the cell-cycle pro-gression upon DNA damage. Indeed, we found that

p16 Impairs HR-Mediated DNA Repair in Head and Neck Cancers

www.aacrjournals.org Cancer Res; 74(6) March 15, 2014 1741




suppression of p16INK4a did not alter the cell-cycle distribu-tion of HPV-positive SCC154 cells either under untreatedconditions or after ionizing radiation, confirming thatp16INK4a does not affect the cell-cycle progression of HPV-positive cells upon DNA damage (Fig. 3A). In addition, we didnot observe any difference in BrdUrd incorporation at differenttime points after ionizing radiation between cells that expressshRNAs against shluc or shp16INK4a (Fig. 3B). These resultssuggest that p16INK4a affects radiation response independent-ly from its ability to regulate cell cycle.

To confirm these observations,we overexpressedp16INK4a inHPV-negative SQD9 cells. To exclude the effect of p16INK4a onDNA damage response due to its impact on cell-cycle progres-sion, we inhibited the retinoblastoma pathway by introducingHPV-E7 to SQD9 cells.We confirmed that p16INK4a overexpres-sion did not alter cell-cycle distribution of SQD9E7 cells afterDNA damage (Fig. 3C). On the other hand, we found thatoverexpression of p16INK4a in SQD9E7 cells resulted in adecreased cell survival upon DNA damage (Fig. 3D and E andSupplementary Fig. S1D). Taken together, these results stronglyindicate that p16INK4a overexpression contributes to radiosen-sitivity of HPV-positive HNSCCs. This also suggests a novel

function of p16INK4a in control of DNA damage response thatis independent from its ability to regulate cell-cycle progression.

p16INK4a affects radiosensitivity independent of itsability to inhibit CDK4/6 activity

Previous reports demonstrated that CDKs could modulateDNA damage response and affect cell survival upon DNAdamage by phosphorylating DNA repair proteins (21–23).Therefore, we hypothesize that p16INK4a overexpressionaffects radiosensitivity due to direct inhibition of CDK4/6activity. To test this idea, we examined whether a specificCDK4/6 inhibitor, PD-0332991 (24, 25), has the same effect onsurvival of HPV-positive cells upon ionizing radiation as weobserved in case of p16INK4a overexpression.

Consistent with previous reports (24, 26), the presence ofPD-0332991 did not affect cell-cycle distribution of HPV-pos-itive SCC154 cells with inactivated retinoblastoma (Supple-mentary Fig. S2A). In contrast, we observed G1–S cell-cyclearrest and decreased proliferation of HPV-negative SQD9 cellsafter treatment with increasing doses of PD-0332991 (Supple-mentary Fig. S2A and S2B). Also, a slight increase in toxicitywas seen with higher doses of PD-0332991 in HPV-negative

p16INK4a-Negative p16INK4a-Positive

HPV-

Negative

HPV-

Positive

A

C D

B

0 1 2 3 4 5

Time since treatment (y)

0

10

20

30

40

50

60

70

80

90

100

Locore

gio

nal tu

mor

contr

ol (%

)

p16INK4a-Negative n = 58

p16INK4a-Positive n = 29 96%

68%

P = 0.01

HPV p16INK4a-

Positive

Positive 11 (20%) 3 (6%)

Negative 9 (17%) 31 (57%)

Total (n = 54) 20 (37%) 34 (63%)

p16INK4a-

Negative

Vinculin

p16INK4a

SC

C1

54

SC

C0

90

SQ

D9

CA

L2

7

SC

26

3

Figure 1. High levels of p16INK4a expression correlates with better therapy response in patients with HNSCC. A, representative examples of p16INK4aimmunostaining in HNSCCs. Pictures were taken with a magnification of �200 and �400. B, p16INK4a overexpression correlates with HPV positivityin HNSCC samples. p16INK4a andHPV statuses were assessed in the same tumor samples. C, immunoblot analysis of p16INK4a expression in the indicatedcell lines using an anti-p16INK4a antibody. D, survival of patients with HNSCC with different status of p16INK4a as measured by Kaplan–Meier curves withlocoregional tumor control as endpoint.

Dok et al.





SQD9 cells (Supplementary Fig. S2C). These findingsconfirm that the used doses of PD-0332991 efficiently inhibitedCDK4/6 activity. Nonetheless, we did not find any differences

in cell-cycle progression after ionizing radiation between cellsexpressing shRNAs against either luciferase (shluc) orp16INK4a (shp16) in the presence of PD-0332991 (Fig. 4A).

Clo

nogenic

surv

ival (%

)

Radiation dose (Gy)

shlu

c

shp16-1

shp16-2

p16INK4a

β-Actin

shluc

shp16-1

shp16-2

A B

C D

0

0.5

1

1.5

shluc shp16-1 shp16-2

Rela

tive c

ellu

lar

surv

ival

*

0 Gy

2 Gy

4 Gy

6 Gy

0 Gy 2 Gy 4 Gy 6 Gy

P = 0.04

*P = 0.04 *P = 0.03

0.1

1

10

100

0 2 4 6

Clo

no

ge

nic

su

rviv

al (%

)

*

Radiation dose (Gy)

P = 0.04

*P = 0.03SQD9

SCC154

SCC090

0 Gy 2 Gy 4 Gy 6 Gy

SQD9

SCC154

SCC090

0.1

1

10

100

0 2 4 6

shluc

shp16-1

shp16-2

Rela

tive c

ellu

lar

surv

ival

0

0.5

1

1.5

SC263 CAL27 SQD9 SCC154 SCC090

0.1

1

10

100

0 2 4 6

Control

sip16-1

sip16-2 Clo

nogenic

surv

ival (%

)

Radiation dose (Gy)

0 Gy

2 Gy

4 Gy

6 Gy

E

SCC154

SCC090

0 Gy 2 Gy 4 Gy 6 Gy

Control

sip16-1

sip16-2

P = 0.05*

P ≤ 0.01*P ≤ 0.01*

p16INK4a

β-Actin

Contr

ol

sip

16-1

sip

16-2

1 0.5 0.4

1 0.4 0.5

Figure 2. p16INK4a sensitizes HPV-infected HNSCCs to ionizing radiation. A, colony formation assay of p16INK4a-positive SCC090 and SCC154 cells andp16INK4a-negative SQD9 cells after exposure to increasing doses of ionizing radiation. B, cell survival of the indicated cell lines 7 days after treatmentwith increasing doses of ionizing radiation determined by sulforhodamine B assay. C, colony formation assay of SCC154 cells expressing shRNAstargeting luciferase (shluc) or p16INK4a treated with increasing doses of ionizing radiation. Suppression of p16INK4a was confirmed by immunoblotting.Knockdown efficiency was assessed by densitometry in arbitrary units (below the immunoblots). D, cell survival of the indicated cell lines determinedby sulforhodamine B assay 7 days after treatment with increasing doses of ionizing radiation. E, colony formation assay of SCC090 cells expressingsiRNA for p16INK4a or scrambled siRNA (control) treated with increasing doses of ionizing radiation. Suppression of p16INK4a was confirmed byimmunoblotting. Knockdown efficiency was assessed by densitometry in arbitrary units (below the immunoblots). For A, C, and E, survival is expressed asa percentage � SEM of colonies formed relative to untreated cells over three experiments. Representative images (left) of the colony assays. For B andD, survival is expressed as �SEM relative to untreated cells in three independent experiments.






Next, we performed a colony formation assay and ashort-term cell survival analysis upon ionizing radiationof SCC154 cells pretreated with PD-0332991 for 24 hours.Inhibition of CDK4/6 activity with PD-0332991 did notaffect radiosensitivity of p16INK4a knockdown cells (Fig.4B and C and Supplementary Fig. S2D). Taken together,these results indicate that increased sensitivity ofp16INK4a-overexpressing/HPV-positive cells to ionizing

radiation is not due to the ability of p16INK4a to inhibitCDK4/6 activity.

p16INK4a inhibits recruitment of RAD51 to the sites ofDNA damage via cyclin D1 downregulation

It is also well established that p16INK4a inhibits CDK4/6activity by displacing cyclin D1 from cyclin D1–CDK4/6 com-plexes that, in turn, leads to cyclin D1 degradation (27). In fact,

10

100

0 2 4 6

SQD9E7 p16

SQD9E7 V

0

0.5

1

1.5

SQD9E7 V SQD9 E7 p16

Rela

tive c

ellu

lar

surv

ival

0 Gy2 Gy

4 Gy

6 Gy

Clo

no

ge

nic

su

rviv

al (%

)

Radiation dose (Gy)

D E

P = 0.04*

P = 0.02*

p16INK4a

Vinculin

p16

V

0

20

40

60

80

100

p16 V p16 V p16 V p16 V

0 Gy 4 h 14 h 24 h

6 Gy

Cell-

cycle

phase (

%)

0 Gy

DNA content

shluc

2 n 2 n2 n2 n 4 n 4 n 4 n4 n

2 n 2 n2 n2 n 4 n 4 n 4 n4 n

2 n 2 n2 n2 n 4 n 4 n 4 n4 n

Cell

count

shp16-1

shp16-2

0

20

40

60

80

100

shlu

c

shp16-1

shp16-2

0 Gy 4 h 14 h 24 h

shlu

c

shp16-1

shp16-2

shlu

c

shp16-1

shp16-2

shlu

c

shp16-1

shp16-2

Cell-

cycle

phase (

%)

G2–M S

G1

G2–M S

G1

6 Gy 4 h 14 h 24 h

6 Gy

0

20

40

60

80

100

shlu

c

shp16-1

shp16-2

shlu

c

shp16-1

shp16-2

shlu

c

shp16-1

shp16-2

shlu

c

shp16-1

shp16-2

0 Gy 4 h 14 h 24 h

6 Gy

Brd

Urd

-positiv

e c

ells

(%

)

B

A

C

SQD9E7

Figure 3. p16INK4a affects radiosensitivity independent of its ability to regulate cell cycle. A, cell-cycle distribution of propidium iodide–stained SCC154cells expressing shRNAs specific to luciferase (shluc) or p16INK4a after ionizing radiation with 6 Gy. B, percentage of BrdUrd-positive cells inSCC154 cells expressing shRNA for luciferase or p16INK4a after ionizing radiation with 6 Gy. C, cell-cycle distribution of propidium iodide–stainedSQD9E7 cells expressing p16INK4a or empty vector (V) after ionizing radiation with 6Gy. D, colony formation assay of SQD9E7 cells expressing p16INK4a orempty vector (V) with increasing doses of ionizing radiation. Overexpression of p16INK4a was confirmed by immunoblotting. Survival is expressedas a percentage � SEM of colonies formed relative to untreated cells over three experiments. E, cell survival of the indicated cell lines determined bysulforhodamine B assay 7 days after ionizing radiation. Survival is expressed as �SEM relative to untreated cells in three independent experiments.A–C, percentage of cells is presented as a mean � SEM for three independent experiments.

Dok et al.





we found lower levels of cyclin D1 expression in p16INK4a-overexpressing SCC154 and SCC090 cells, whereaswe observedhigher levels of cyclin D1 in p16INK4a-negative SQD9, SC263,and CAL27 cells (Fig. 5A).

We also tested whether there is a correlation between cyclinD1 and p16INK4a and between cyclin D1 and HPV status inHNSCC samples by performing immunohistochemistry (IHC)analysis of cyclin D1 expression in HNSCC biopsies from the

A

DNA content

shluc

Cell

count

shp16-1

shp16-2

0 Gy 4 Gy 24 h

0

20

40

60

80

100

Cell-

cycle

phase (

%)

shlu

c

shp16-1

shp16-2

shlu

c

shp16-1

shp16-2

0 Gy 4 Gy 24 h

B C

0

0.5

1.0

1.5

shluc shp16-1 shp16-2

Rela

tive c

ellu

lar

surv

ival

G2–M

S

G1

1 µmol/L PD-0332991

2 n 2 n 4 n4 n

2 n 2 n 4 n4 n

2 n 2 n 4 n4 n

1 µmol/L PD-0332991

1 µmol/L PD-0332991

0 Gy

2 Gy

4 Gy

6 Gy

0 Gy 2 Gy 4 Gy 6 Gy

shluc+

1 µmol/L PD-0332991

shp16-1+

1 µmol/L PD-0332991

shp16-2+

1 µmol/L PD-0332991

Clo

no

ge

nic

su

rviv

al (%

)

Radiation dose (Gy)

shluc+

PD-0332991shp16-1+

PD-0332991shp16-2

PD-03329911

10

100

0 2 4 6

P = 0.04*

P = 0.05*

P = 0.002*P = 0.03*

SCC154

SCC154SCC154

Figure 4. p16INK4a affects radiosensitivity independent of its ability to inhibit CDK4/6 activity. A, cell-cycle distribution of PD-0332991-treated SCC154 cellsexpressing shluc or shp16INK4a upon ionizing radiation with 4 Gy. Cell-cycle distribution (right) is presented as a mean � SEM for two independentexperiments. B, colony formation of PD-0332991-treated SCC154 cells expressing shluc or shp16INK4a upon ionizing radiation with increasing dosesof ionizing radiation. Cell survival is expressed as a percentage� SEM of colonies formed by irradiated cells versus nonirradiated cells for three independentexperiments. Representative images (below) of the colony assay. C, cell survival of PD-0332991-treated SCC154 cells expressing shluc or shp16INK4a 7days after treatment with increasing doses of ionizing radiation as detected by sulforhodamine B assay. Cell survival is expressed as �SEM relative tononirradiated cells. A–C, cells were pretreated with 1 mmol/L of PD-0332991 for 24 hours before ionizing radiation, after which the drug was removed.






Leuven cohort. All HNSCC tumors were classified dichoto-mously as either cyclin D1–positive (strong, diffuse staining inmore than 25% of carcinoma cells) or cyclin D1–negative (Fig.5B). We found that cyclin D1 overexpression correlates withHPV negativity and low p16INK4a expression (P values of �0.001; the Yates c2 test; and the nonparametric spearman-ranktest; Supplementary Fig. S3A), further confirming that HPVstatus and p16INK4a overexpression could affect the expres-

sion of cyclin D1 (Fig. 5C and D and Supplementary Fig. S3A).High levels of cyclin D1 in a number of p16INK4a-positivetumors (Fig. 5D) could be explained by cyclin D1 amplification,which is commonly observed in HNSCCs (28).

Previous studies revealed that elevated levels of cyclin D1confer radiation resistance of cancer cells (29, 30) and highlevels of cyclin D1 correlate with unfavorable response toradiotherapy (31, 32). Conversely, suppression of cyclin D1

B

Cyclin D1

Negative Positive A

D

0

20

40

60

80

100

shluc Cyclin D1

% o

f C

ells

with

>1

0

RA

D5

1 f

oci

E F

Cp16INK4a Cyclin D1–

negative

Cyclin D1–

positive

Negative 8 (10%) 39 (51%)

Positive 16 (21%) 13 (17%)

Total (n = 76) 24 (32%) 52 (68%)

HPV Cyclin D1–

negative

Cyclin D1–

positive

Negative 5 (11%) 28 (61%)

Positive 9 (19%) 4 (9%)

Total (n = 46) 14 (30%) 32 (70%)

0.01

0.1

1

10

100

0 2 4 6 8 10

Cyclin D1

shluc

0 Gy 4 Gy 8 Gy 10 Gy

Cyclin D1

shluc

Radiation dose (Gy)

Clo

no

ge

nic

su

rviv

al (%

)

*P = 0.02

*P < 0.001 *P < 0.001

*P = 0.02

Cyclin D1

Vinculin S

CC

154

SC

C090

SQ

D9

CA

L27

SC

263

Cyclin D1

Vinculin

shlu

c

Cyclin

D1

SCC154

SCC154

Figure 5. p16INK4anegatively correlateswith cyclin D1protein expression. A, immunoblot analysis of cyclinD1 expression in the indicatedcell lines using anti-cyclin D1 antibody. B, representative examples of cyclin D1 immunostaining in HNSCCpatient samples. Imageswere takenwith amagnification of�200 and�400. C, association between HPV status and cyclin D1 expression was assessed in the same tumor samples. D, association between p16INK4aoverexpression and cyclin D1 expression was assessed in the same tumor samples by IHC using antibodies against cyclin D1 and p16INK4a. E, colonyformation of SCC154 shluc cells transfected with an empty vector or cyclin D1 after exposure to increasing doses of ionizing radiation. Cell survival isexpressedasapercentage�SEMof colonies formedby irradiated versus nonirradiated cells. Representative imagesof the colony assay are shownbelow thegraph. Expression of cyclin D1 was confirmed by immunoblotting. F, RAD51 foci formation 4 hours after 2 Gy of ionizing radiation in SCC154 shluc cellsexpressing empty vector or cyclin D1. Data, mean � SEM percentage of cells containing more than 10 foci for three independent experiments.

Dok et al.





sensitizes cancer cells to DNA damage (29, 33). Our results alsodemonstrated that the response to ionizing radiation ofHNSCCcells strongly correlated with levels of cyclin D1 expression(Figs. 2A and B and 5A). Furthermore, when we overexpressedcyclin D1 in p16INK4a-positive SCC154 cells, it led to radiationresistance of these cells (Fig. 5E and Supplementary Fig. S3B).Given that p16INK4a overexpression leads to cyclin D1 down-regulation, this suggests that p16INK4a could affect radiosen-sitivity by modulating cyclin D1 protein expression.Recent reports have shown that cyclin D1 facilitates homol-

ogous recombination–mediated DNA repair through directrecruitment of RAD51 to the sites of DNA damage (34, 35). Infact, we found that overexpression of cyclin D1 in SCC154 cellsenhanced radiation-induced RAD51 foci formation, a markerof homologous recombination competence (Fig. 5F).We next examined how p16INK4a affects cyclin D1–RAD51

complex formation. We pulled down cyclin D1 from SCC154 orSCC090 cells expressing either shluc, or shp16. Consistent withour prior observations, we revealed higher expression levels ofcyclin D1 in p16INK4a-depleted cells compared with shluc-expressing cells (Fig. 6A and B). Notably, suppression ofp16INK4a led to accumulation of cyclin D1–RAD51 complexes(Fig. 6A and B). Inhibition of p16INK4a expression also led tothe increased amount of RAD51 associated with chromatin,whereas p16INK4a depletion did not affect RAD51 proteinlevels in the soluble fraction (Fig. 6C and D). We also assessedthe effect of p16INK4a overexpression in SQD9E7 cells oncyclin D1–Rad51 complex formation. In contrast withp16INK4a knockdown, overexpression of p16INK4a impairedthe interaction between cyclin D1 and Rad51. A lower amountof RAD51 associated with chromatin was seen in cells over-expressing p16INK4a comparedwith cells expressing an emptyvector (Fig. 6E and F). These results strongly indicate thatp16INK4a suppresses RAD51 recruitment to the chromatin bydownregulating cyclin D1 protein expression.

p16INK4a impairs homologous recombination–mediated DNA repairWenext assessedwhether p16INK4a affects the homologous

recombination–mediated DNA repair pathway in HPV-posi-tive cells by analyzing RAD51 foci formation. Consistent withthe observation that p16INK4a suppression blocks RAD51recruitment to the chromatin, we found decreased RAD51 fociformation in p16INK4a-positive SCC154 and SCC090 HNSCCcells compared with p16INK4a-negative SQD9 cells (Fig. 7A).On the other hand, depletion of p16INK4a resulted in adramatic increase of RAD51 foci formation, indicating thatp16INK4a contributes to control of the error-free homologousrecombination–mediated DNA repair pathway (Fig. 7B).We hypothesized that a decrease in the homologous recom-

bination–mediated DNA repair pathway could result in anincrease in the error-prone nonhomologous end-joining path-way (NHEJ), which commonly leads to misrepair of double-strand breaks in DNA. Increased rate of NHEJ DNA repairresults in formation of both dicentric chromosomes andacentric chromosome fragments, which could be detected bymicronuclei analysis (36, 37). In concordance with our idea, amicronuclei analysis revealed increased frequency of micro-

nucleated cells in p16INK4a-positive SCC154 cells upon ion-izing radiation compared with p16INK4a-depleted cells (Fig.7C). This suggests that p16INK4a overexpression in HPV-positive cells impairs the homologous recombination DNArepair but facilitates the NHEJ pathway, leading to increasedionizing radiation–mediated genomic instability and enhancedradiosensitivity of p16INK4a-positive cells.

DiscussionThe role of the tumor suppressor p16INK4a in the regulation

of cell cycle and senescence has been studied extensivelyduring last two decades (38, 39). On the other hand, cell-cycle–independent functions of p16INK4a are much less stud-ied. Because the HPV oncogenic proteins, E6 and E7, disruptcell-cycle regulation and lead to p16INK4a overexpression,HPV-infected tumors serve as an ideal platform to elucidatethe involvement of p16INK4a in regulation of cellular processesother than control of cell cycle. Here, we found that p16INK4a isdirectly involved in control of DNA damage response inde-pendently from its ability to inhibit CDK4/6 activity andregulate cell-cycle progression. Conversely, p16INK4a expres-sion impairs error-free homologous recombination–mediatedDNA repair by downregulating cyclin D1 protein expressionand thereby decreasing recruitment of RAD51 to the sites ofDNA damage.

HPV status is well known to be an independent prognosticfactor determining tumor control and survival after radio-therapy for head and neck cancer. Despite the prognosticvalue of HPV-positive HNSCCs, universal methods for strat-ifying patients with HPV status remain elusive (40) becausedirect methods of HPV detection are complex to implementin a clinical setup. Recent studies propose that p16INK4aexpression could serve as a surrogate marker for HPVpositivity (41, 42).

Our results strongly indicate that p16INK4a actively mod-ulates the radiation response and has a broader role than itscausal relationship with HPV infections. Consistent with ouranalysis of the Leuven cohort of patients with HNSCC, recentanalyses of clinical data (8, 42) revealed that p16INK4a expres-sion has a major impact on treatment response and survival inpatients with HNSCC and have a stronger prognostic valuethan HPV status.

Importantly, the prognostic value of p16INK4a expressionhas been shown to depend on its cellular localization.Nuclear p16INK4a gave a better correlation with prognosis(42, 43). In our study, we classified p16INK4a tumors aspositive only for samples with higher than 10% nuclearstaining. This cutoff is quite strict and it could explain whysome of HPV-positive tumors were considered as p16INK4a-negative. We also observed the high levels of p16INK4aexpression in some of HPV-negative tumors. It could be dueto other mechanisms of retinoblastoma pathway inactiva-tion. Although retinoblastoma mutations are quite rare inHNSCC, deletions of the retinoblastoma-containing regionare found in 15% to 50% of the tumors (44). Moreover, CCND1has been shown to be amplified in about 30% of HNSCCs(31). Taken together, our data strongly imply p16INK4a is an






independent marker for prediction of radiotherapy responsein patients with HNSCC.

In addition, we confirmed previous reports demonstrating adirect involvement of cyclin D1 in DNA damage response (34,35). There are conflicting studies on how cyclin D1 expression

affects response to ionizing radiation in HNSCCs. Whereassome groups reported a negative correlation between highcyclin D1 expression and radiosensitivity (30), other groupsobserved that high cyclin D1 expression is associated withincreased radiosensitivity in oral squamous cell cancer and

RAD51

Cyclin D1

CDK4

0 Gy 5 Gy

sh

luc

sh

luc

sh

p1

6-2

sh

p1

6-2

0 Gy 5 Gy

sh

luc

sh

luc

sh

p1

6-2

sh

p1

6-2

0 Gy 5 Gy

sh

luc

sh

luc

sh

p1

6-2

sh

p1

6-2

Vinculin

Cyclin D1

RAD51

p16INK4a

Vinculin

Cyclin D1

RAD51

0 Gy 5 Gy

p16

p16

V

V

IgG

RAD51

Cyclin D1

CDK4

0 Gy 5 Gy

p16

p16

V

V

Vinculin

Cyclin D1

RAD51

CDK4

p16INK4a

WCL:

0 Gy 5 Gy

p16

p16

V

V

0 Gy 5 Gy

p16

p16

V

V

Vinculin

Cyclin D1

RAD51

p16INK4a

Vinculin

Cyclin D1

RAD51

SCC154

SQD9E7

SCC154

SQD9E7

0 Gy 5 Gy

shlu

c

shlu

c

shp16-2

shp16-2

IgG

RAD51

Cyclin D1

IP: Cyclin D1

CDK4

Vinculin

Cyclin D1

RAD51

CDK4

p16INK4a

0 Gy 5 Gy

shlu

c

shlu

c

shp16

-2

shp16

-2

WCL:

SCC090SCC090

0 Gy 5 Gy

shlu

c

shlu

c

shp16

-2

shp16

-2

Soluble fraction

Vinculin

Cyclin D1

RAD51

p16INK4a

Vinculin

Cyclin D1

RAD51

0 Gy 5 Gy

shlu

c

shlu

c

shp16

-2

shp16

-2

A

B

C

D

E F

IgG

Chromatin fraction

Soluble fraction Chromatin fraction

IP: Cyclin D1

IP: Cyclin D1

0 Gy 5 Gy

sh

luc

sh

luc

sh

p1

6-2

sh

p1

6-2

Vinculin

Cyclin D1

RAD51

CDK4

p16INK4a

WCL:

Soluble fraction Chromatin fraction

1 1.5 0.9 1.1

1 1.8 3 4.8

1 2.1 1.3 1.7

1 1.5 3.9 6.4

1 1.1 0.8 1.5

1 0.7 0.4 2.7

1 2.2 0.5 1.2

1 0.5 1.5 9.9

1 1.2 1.5 4.7

1 1.2 2.0 6.4

1 2.5 0.6 1.8

1 1.1 0.9 35.4

Figure6. p16INK4a inhibits recruitment ofRAD51 to thesites ofDNAdamagebydownregulating cyclinD1protein expression. AandB, coimmunoprecipitationof cyclin D1 with RAD51 and CDK4 in SCC154 and SCC090 cells expressing shluc or shp16INK4a 4 hours after 5 Gy of ionizing radiation. Immunoblottingwas performed using the indicated antibodies. C and D, association of RAD51 with chromatin in SCC154 and SCC090 cells expressing shluc or shp16INK4a 4 hours after ionizing radiation with 5 Gy. E, coimmunoprecipitation of cyclin D1 with RAD51 and CDK4 in SQD9E7 cells expressingp16INK4a or empty vector (V) 4 hours after 5 Gy of ionizing radiation. Immunoblotting was performed using the indicated antibodies. F, associationof RAD51 with chromatin in SQD9E7 cells expressing shp16INK4a or empty vector (V) 4 hours after ionizing radiation with 5 Gy. For C, D, and F, solubleand chromatin fractions were isolated and immunoblotted with the indicated antibodies. A–F, the expression levels of indicated proteins were assessed bydensitometry (below the immunoblots).

Dok et al.





% o

f C

ells

with m

icro

nucle

i

A B

0 Gy 24 h 40 h 48 h

shluc

shp16-2C

6 Gy

SQD9

SCC090

SCC154

0 Gy 2 Gy 4 h 0 Gy 2 Gy 4 h

shluc

shp16-1

shp16-2

shluc shp16-2

48 h after

6 Gy

0

20

40

60

80

100

% o

f C

ells

with >

10 R

AD

51 foci

SQ

D9

SC

C090

SC

C154

*P < 0.001

*P < 0.001*P < 0.001

*P < 0.001

*P = 0.006

0

20

40

60

80

100

% o

f C

ells

with >

10 R

AD

51 foci

shlu

c

shp16-1

shp16-2

0

20

40

60

80

100

SCC154

SCC154

Figure 7. p16INK4a impairs homologous recombination–mediated DNA repair. A, RAD51 foci formation 4 hours after 2 Gy of ionizing radiation in theindicated cell lines. Data, mean � SEM percentage of cells containing more than 10 foci in three experiments. Representative images of RAD51immunostaining (supra). Scale bar, 100 mm. B, RAD51 foci formation 4 hours after 2 Gy of ionizing radiation in shluc and shp16INK4a-expressing cell lines.Data, mean � SEM percentage of cells containing more than 10 foci in three experiments. Representative images of RAD51 immunostaining (supra).Scale bar, 100 mm.C, relative number of micronucleated cells in expressing shluc or shp16INK4a after 6 Gy of ionizing radiation. Graph,mean� SEMof threeindependent experiments. Representative images of micronuclei detection (left). Arrows, micronucleated cells. Images were taken with a magnificationof �60.






early-stage of laryngeal cancer (45, 46). It is possible that thecontradictory observations in these reports are due to the lackof p16INK4a andHPVdetection. Moreover, the strong negativecorrelation between HPV status and cyclin D1 expression inthe Leuven cohort and the interaction between p16INK4a,cyclin D1, and RAD51 inHPV-positive cells suggests that cyclinD1 has only a prognostic value in HPV-positive HNSCCs.Therefore, cyclinD1 IHCcould be used in addition to p16INK4aas a biomarker for stratification of patients with HPV-positiveHNSCC. Unfortunately, the relatively low number of patientswith HPV-positive HNSCC did not allow us to make conclu-sions with certitude.

We believe that our data have important clinical implica-tions by highlighting the importance of p16INK4a as a markerfor stratification of patients with HNSCC. Current highlyintensive chemo- and radiation treatment schedules lead tosevere short- and long-term morbidity (47, 48). Clinical trialslooking at treatment deintensification and organ preservativetherapeutic procedures for HPV-positive tumors are ongoing.Our data suggest that p16INK4a status should be taken intoaccount for further clinical trials of novel treatment therapies.

We also found that p16INK4a overexpression dramaticallyimpairs the homologous recombination–DNA repair pathway,which suggests the potential use of PARP inhibitors, a prom-ising class of anticancer agents particularly effective againsthomologous recombination–deficient tumors (49), for treat-ment of p16INK4a-positive HNSCCs. However, further in vitroand in vivo studies are required to confirm the utility of thisapproach for patients with HNSCC.

In summary, we found that p16INK4a hampers homologousrecombination–mediated DNA repair machinery and sensi-tizes HPV-positive cells to radiation treatment, implicatingp16INK4a expression as a potent marker for stratification ofpatients with HNSCC.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: R. Dok, P. Kalev, E.J. Van Limbergen, A. Sablina,S. NuytsDevelopment of methodology: R. Dok, P. Kalev, S. NuytsAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): R. Dok, P. Kalev, E.J. Van Limbergen, L. AbbasiAsbagh, I. V�azquez, E. Hauben, S. NuytsAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): R. Dok, E.J. Van Limbergen, E. Hauben, S. NuytsWriting, review, and/or revision of the manuscript: R. Dok, I. V�azquez,E. Hauben, A. Sablina, S. NuytsAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): R. Dok, P. Kalev, L. Abbasi AsbaghStudy supervision: A. Sablina, S. Nuyts

Grant SupportThis work was supported by grants from the Flemish League Against Cancer

(VLK project and VLK van der Schueren grants) and FWO (grant numberG-083313).

The costs of publication of this article were defrayed in part by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received August 28, 2013; revised December 30, 2013; accepted January 14,2014; published OnlineFirst January 28, 2014.

References1. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics,

2009. CA Cancer J Clin 2009;59:225–49.2. Fakhry C, Gillison ML. Clinical implications of human papillomavirus in

head and neck cancers. J Clin Oncol 2006;24:2606–11.3. Lassen P. The role of human papillomavirus in head and neck cancer

and the impact on radiotherapy outcome. Radiother Oncol 2010;95:371–80.

4. Fakhry C, Westra WH, Li S, Cmelak A, Ridge JA, Pinto H, et al.Improvedsurvival of patientswith humanpapillomavirus-positive headand neck squamous cell carcinoma in a prospective clinical trial. J NatlCancer Inst 2008;100:261–9.

5. Marur S, Forastiere AA. Head and neck cancer: changing epidemiol-ogy, diagnosis, and treatment. Mayo Clin Proc 2008;83:489–501.

6. Sellers WR, Rodgers JW, Kaelin WG Jr. A potent transrepressiondomain in the retinoblastoma protein induces a cell-cycle arrest whenbound to E2F sites. Proc Natl Acad Sci U S A 1995;92:11544–8.

7. Bremner R, Cohen BL, Sopta M, Hamel PA, Ingles CJ, Gallie BL, et al.Direct transcriptional repression by pRB and its reversal by specificcyclins. Mol Cell Biol 1995;15:3256–65.

8. Weinberger PM, Yu Z, Haffty BG, Kowalski D, Harigopal M, BrandsmaJ, et al. Molecular classification identifies a subset of human papillo-mavirus–associated oropharyngeal cancers with favorable prognosis.J Clin Oncol 2006;24:736–47.

9. Dirix P, Nuyts S. Evidence-based organ-sparing radiotherapy in headand neck cancer. Lancet Oncol 2010;11:85–91.

10. Nuyts S, Dirix P, Clement PM, Poorten VV, Delaere P, Schoenaers J,et al. Impact of adding concomitant chemotherapy to hyperfractio-nated accelerated radiotherapy for advanced head-and-neck squa-mous cell carcinoma. Int J Radiat Oncol Biol Phys 2009;73:1088–95.

11. Rieckmann T, Tribius S, Grob TJ,Meyer F, BuschCJ, PetersenC, et al.HNSCC cell lines positive for HPV and p16 possess higher cellular

radiosensitivity due to an impaired DSB repair capacity. RadiotherOncol 2013;107:242–6.

12. Kimple RJ, Smith MA, Blitzer GC, Torres AD, Martin JA, Yang RZ, et al.Enhanced radiation sensitivity in HPV-positive head and neck cancer.Cancer Res 2013;73:4791–800.

13. SmeetsSJ,HesselinkAT,Speel EJ,HaesevoetsA,SnijdersPJ, PawlitaM, et al. A novel algorithm for reliable detection of human papilloma-virus in paraffin embedded head and neck cancer specimen. Int JCancer 2007;121:2465–72.

14. DeSchutterH,KimpeM, IsebaertS,NuytsS.Asystematic assessmentof radiationdose enhancement by5-Aza-20-deoxycytidine andhistonedeacetylase inhibitors in head-and-neck squamous cell carcinoma. IntJ Radiat Oncol Biol Phys 2009;73:904–12.

15. Moffat J, Grueneberg DA, Yang X, KimSY, Kloepfer AM,Hinkle G, et al.A lentiviral RNAi library for human and mouse genes applied to anarrayed viral high-content screen. Cell 2006;124:1283–98.

16. Xu B, Kim ST, Lim DS, Kastan MB. Two molecularly distinct G(2)/Mcheckpoints are induced by ionizing irradiation. Mol Cell Biol 2002;22:1049–59.

17. Mladenov E, Tsaneva I, Anachkova B. Cell-cycle–dependent asso-ciation of Rad51 with the nuclear matrix. DNA Cell Biol 2007;26:36–43.

18. Kalev P, Simicek M, Vazquez I, Munck S, Chen L, Soin T, et al. Loss ofPPP2R2A inhibits homologous recombinationDNA repair andpredictstumor sensitivity to PARP inhibition. Cancer Res 2012;72:6414–24.

19. Seo HR, Lee DH, Lee HJ, Baek M, Bae S, Soh JW, et al. Cyclin G1overcomes radiation-induced G2 arrest and increases cell deaththrough transcriptional activation of cyclin B1. Cell Death Differ 2006;13:1475–84.

20. Lassen P, Eriksen JG, Hamilton-Dutoit S, Tramm T, Alsner J, Over-gaard J. Effect of HPV-associated p16(INK4A) expression on response

Dok et al.





to radiotherapy and survival in squamous cell carcinoma of the headand neck. J Clin Oncol 2009;27:1992–8.

21. Cerqueira A, Santamaria D, Martinez-Pastor B, Cuadrado M, Fernan-dez-Capetillo O, Barbacid M. Overall Cdk activity modulates the DNAdamage response in mammalian cells. J Cell Biol 2009;187:773–80.

22. Johnson N, Shapiro GI. Cyclin-dependent kinases (cdks) and the DNAdamage response: rationale for cdk inhibitor-chemotherapy combina-tions as an anticancer strategy for solid tumors. Expert Opin TherTargets 2010;14:1199–212.

23. Alvarez-Fernandez M, Halim VA, Krenning L, Aprelia M, Mohammed S,HeckAJ, et al. Recovery fromaDNAdamage-inducedG2arrest requiresCdk-dependent activation of FoxM1. EMBO Rep 2010;11:452–8.

24. Fry DW, Harvey PJ, Keller PR, Elliott WL, Meade M, Trachet E, et al.Specific inhibition of cyclin-dependent kinase 4/6 by PD 0332991 andassociated antitumor activity in human tumor xenografts. Mol CancerTher 2004;3:1427–38.

25. ToogoodPL, Harvey PJ, Repine JT, SheehanDJ, VanderWel SN, ZhouH, et al. Discovery of a potent and selective inhibitor of cyclin-depen-dent kinase 4/6. J Med Chem 2005;48:2388–406.

26. WiedemeyerWR,Dunn IF,Quayle SN, Zhang J,ChhedaMG,DunnGP,et al. Pattern of retinoblastoma pathway inactivation dictates responseto CDK4/6 inhibition in GBM. Proc Natl Acad Sci U S A 2010;107:11501–6.

27. Sherr CJ, Roberts JM. CDK inhibitors: positive and negative regulatorsof G1-phase progression. Genes Dev 1999;13:1501–12.

28. Kaminagakura E, Werneck da Cunha I, Soares FA, Nishimoto IN,Kowalski LP. CCND1 amplification and protein overexpression in oralsquamous cell carcinoma of young patients. Head Neck 2011;33:1413–9.

29. Shimura T, KakudaS,Ochiai Y, NakagawaH,KuwaharaY, Takai Y, et al.Acquired radioresistance of human tumor cells by DNA-PK/AKT/GSK3beta-mediated cyclin D1 overexpression. Oncogene 2010;29:4826–37.

30. Milas L, Akimoto T, Hunter NR,MasonKA, Buchmiller L, YamakawaM,et al. Relationship between cyclin D1 expression and poor radio-response of murine carcinomas. Int J Radiat Oncol Biol Phys 2002;52:514–21.

31. ChangAR,WuHG,ParkCI, JeonYK,KimCW.Expressionof epidermalgrowth factor receptor and cyclin D1 in pretreatment biopsies as apredictive factor of radiotherapy efficacy in early glottic cancer. HeadNeck 2008;30:852–7.

32. Lai JP, Tong CL, Hong C, Xiao JY, Tao ZD, Zhang Z, et al. Associationbetween high initial tissue levels of cyclin D1 and recurrence ofnasopharyngeal carcinoma. Laryngoscope 2002;112:402–8.

33. Wang MB, Yip HT, Srivatsan ES. Antisense cyclin D1 enhancessensitivity of head and neck cancer cells to cisplatin. Laryngoscope2001;111:982–8.

34. Jirawatnotai S, HuY,MichowskiW, Elias JE,Becks L,BienvenuF, et al.A function for cyclin D1 in DNA repair uncovered by protein inter-actome analyses in human cancers. Nature 2011;474:230–4.

35. Li Z, Jiao X, Wang C, Shirley LA, Elsaleh H, Dahl O, et al. Alternativecyclin D1 splice forms differentially regulate the DNA damageresponse. Cancer Res 2010;70:8802–11.

36. Oliveira NG, Castro M, Rodrigues AS, Goncalves IC, Gil OM, Fer-nandes AP, et al. Wortmannin enhances the induction of micronucleiby low and high LET radiation. Mutagenesis 2003;18:37–44.

37. FenechM.Micronuclei and their associationwith spermabnormalities,infertility, pregnancy loss, preeclampsia and intrauterine growthrestriction in humans. Mutagenesis 2011;26:63–7.

38. Burd CE, Sorrentino JA, Clark KS, Darr DB, Krishnamurthy J, Deal AM,et al. Monitoring tumorigenesis and senescence in vivo with a p16(INK4a)-luciferase model. Cell 2013;152:340–51.

39. Anders L, Ke N, Hydbring P, Choi YJ, Widlund HR, Chick JM, et al. Asystematic screen for CDK4/6 substrates links FOXM1 phosphoryla-tion to senescence suppression in cancer cells. Cancer Cell 2011;20:620–34.

40. Lewis JS Jr. p16 Immunohistochemistry as a standalone test for riskstratification in oropharyngeal squamous cell carcinoma. Head NeckPathol 2012;6:S75–82.

41. Lewis JS, ThorstadWL, Chernock RD, Haughey BH, Yip JH, Zhang Q,et al. p16-positive oropharyngeal squamous cell carcinoma: an entitywith a favorable prognosis regardless of tumor HPV status. Am J SurgPathol 2010;34:1088–96.

42. Zhao N, Ang MK, Yin XY, Patel MR, Fritchie K, Thorne L, et al.Different cellular p16(INK4a) localisation may signal different sur-vival outcomes in head and neck cancer. Br J Cancer 2012;107:482–90.

43. Hoffmann M, Tribius S, Quabius ES, Henry H, Pfannenschmidt S,Burkhardt C, et al. HPV DNA, E6�I-mRNA expression and p16INK4Aimmunohistochemistry in head and neck cancer—how valid isp16INK4A as surrogate marker? Cancer Lett 2012;323:88–96.

44. Burkhart DL, Sage J. Cellular mechanisms of tumour suppression bythe retinoblastoma gene. Nat Rev Cancer 2008;8:671–82.

45. Shintani S, Mihara M, Ueyama Y, Matsumura T, Wong DT. Cyclin D1overexpression associates with radiosensitivity in oral squamous cellcarcinoma. Int J Cancer 2001;96:159–65.

46. Yoo SS, Carter D, Turner BC, Sasaki CT, Son YH, Wilson LD, et al.Prognostic significance of cyclin D1 protein levels in early-stagelarynx cancer treated with primary radiation. Int J Cancer 2000;90:22–8.

47. Machtay M, Moughan J, Trotti A, Garden AS, Weber RS, Cooper JS,et al. Factors associated with severe late toxicity after concurrentchemoradiation for locally advanced head and neck cancer: an RTOGanalysis. J Clin Oncol 2008;26:3582–9.

48. Ang KK, Harris J, Wheeler R, Weber R, Rosenthal DI, Nguyen-Tan PF,et al. Human papillomavirus and survival of patients with oropharyn-geal cancer. N Engl J Med 2010;363:24–35.

49. Rouleau M, Patel A, Hendzel MJ, Kaufmann SH, Poirier GG. PARPinhibition: PARP1 and beyond. Nat Rev Cancer 2010;10:293–301.






Documents

p16INK4a Impairs Homologous Recombination−Mediated DNA ...scholar.harvard.edu/.../p16ink4a_impairs_homologous_recombinatio… · impaired the recruitment of RAD51 to damaged DNA