Nerve Growth Factor Links Oral Cancer Progression, Pain...

Therapeutic Discovery

Nerve Growth Factor Links Oral Cancer Progression,Pain, and Cachexia

Yi Ye1, Dongmin Dang5, Jianan Zhang1, Chi T. Viet2, David K. Lam5, John C. Dolan1,3,Jennifer L. Gibbs4, and Brian L. Schmidt1,2

AbstractCancers often cause excruciating pain and rapid weight loss, severely reducing quality of life in cancer

patients. Cancer-induced pain and cachexia are often studied and treated independently, although both

symptoms are strongly linked with chronic inflammation and sustained production of proinflammatory

cytokines. Because nerve growth factor (NGF) plays a cardinal role in inflammation and pain, and because it

interacts with multiple proinflammatory cytokines, we hypothesized that NGF acts as a key endogenous

molecule involved in the orchestration of cancer-related inflammation. NGF might be a molecule common to

the mechanisms responsible for clinically distinctive cancer symptoms such as pain and cachexia as well as

cancer progression. Here we reported that NGF was highly elevated in human oral squamous cell carcinoma

tumors and cell cultures. Using two validated mouse cancer models, we further showed that NGF blockade

decreased tumor proliferation, nociception, and weight loss by orchestrating proinflammatory cytokines and

leptin production. NGF blockade also decreased expression levels of nociceptive receptors TRPV1, TRPA1,

and PAR-2. Together, these results identified NGF as a common link among proliferation, pain, and cachexia

in oral cancer. Anti-NGF could be an important mechanism-based therapy for oral cancer and its related

symptoms. Mol Cancer Ther; 10(9); 1667–76. �2011 AACR.

Introduction

Pain and cachexia significantly impair function anddegrade quality of life in patients suffering from cancer(1–7). Pain control andweight maintenance are especiallychallenging in patients with head and neck (e.g., oral)cancer. Many oral cancer patients suffer from symptomsthat are more severe than symptoms produced by othercancers (4–7). Oral cancer patients often have troubleeating, drinking, swallowing, and speaking. Despiterecent advances in treatment, pain control and weightmaintenance persist as 2 important clinical challenges.Cancer pain and cachexia are usually assessed and trea-ted as separate, unrelated entities (1). However, thesesymptoms may share the same underlying mechanism,because both are linked to chronic inflammation andshare common proinflammatory mediators (1, 8).

Currently, such a mechanism has not been elucidatedfor cancer pain and cachexia.

Nerve growth factor (NGF) is a key modulator of theneuroendocrine immune axis that serves diversebiological functions (9–13). In a variety of rodent cancermodels, NGF has been shown to play a direct rolein tumor proliferation (13–15) and in perineural invasion.Perineural invasion is a predictor of disease progressionin oral cancer (15). NGF has also been proposed as animportant mediator of tumor-induced bone pain second-ary to prostate cancer (16–18). However, the effect of NGFon cancer proliferation and pain is dependent on tumortype. For example, anti-NGF treatment has been shown toreduce disease progression in breast cancer but has noeffect in bone cancer (14, 16, 17). Bone cancer pain,however, is attenuated by anti-NGF treatment in mice(16–18). Oral cancer pain is usually caused by canceroriginating from soft tissue and clinically distinct frombone cancer pain. Oral cancer pain is exacerbated byfunction, generally does not arise spontaneously, andis not correlated with tumor size. Even the smallest oralcancers can produce severe pain (4, 5). In contrast, bonecancer pain is commonly spontaneous and incessant, andtypically increases with disease progression (16, 17). Theeffect of anti-NGF on oral cancer pain or proliferation isunknown.

The role of NGF in the regulation of body weightalso requires further study. NGF has been shown toparticipate in glucose and lipid metabolism as well as

Authors' Affiliations: 1Bluestone Center for Clinical Research, 2Depart-ment of Oral Maxillofacial Surgery, 3Department of Orthodontics, 4Depart-ment of Endodontics, New York University, New York, New York; and5Department of Oral and Maxillofacial Surgery, University of California SanFrancisco, San Francisco, California

Corresponding Author: Brian L. Schmidt, Bluestone Center for ClinicalResearch, New York University College of Dentistry, 421 First Avenue,233W, New York, NY 10010. Phone: 212-998-9543; Fax: 212-995-4843;E-mail: [email protected]

doi: 10.1158/1535-7163.MCT-11-0123

�2011 American Association for Cancer Research.

MolecularCancer

Therapeutics

www.aacrjournals.org 1667

on August 10, 2019. © 2011 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst July 12, 2011; DOI: 10.1158/1535-7163.MCT-11-0123

http://mct.aacrjournals.org/

feeding behavior (19). Intraperitoneal injection of NGFstimulates the hypothalamic—pituitary–adrenal axis andcauses weight loss in rats (20). Patients with obesity havealtered levels of NGF (9, 19, 21). NGF might contribute toinflammation and metabolic disorders associated withbody weight changes (19, 21). Similarly, cancer cachexiais a complex wasting syndrome comprising inflamma-tory and metabolic disturbances (2). Proinflammatorycytokines including TNF-a and interleukin-6 (IL-6) havebeen shown to induce cachexia by altering metabolism oflipids and muscle proteins (2, 22, 23). NGFmodulates theexpression and release of these proinflammatory cyto-kines (10, 24). Accordingly, NGFmay play a role in cancercachexia.

The varied proinflammatory and nociceptive effects ofNGF suggest a key role for NGF in cancer symptomatol-ogy. We hypothesize that NGF acts as a mediator ofproliferation, pain, and cachexia associated with oralcancer. In this study, we first quantified NGF releaseand expression in patients with oral squamous cell car-cinoma (SCC). We then used 2 separate oral cancermodels in mice to show that anti-NGF reduces prolifera-tion, pain, and weight loss. Interactions of NGF withproinflammatory cytokines and receptors involved innociception were also evaluated.

Materials and Methods

Cancer patientsImmunohistochemistry for NGF in human tumors.

Oral SCC on the affected side and normal epitheliumfrom an anatomically matched area on the contralateralside of 14 oral cancer patients treated at the Universityof California San Francisco (UCSF) Department of Oral& Maxillofacial Surgery were obtained. Tissues werefixed with 4% paraformaldahyde (PFA), dehydrated,embedded in paraffin, and cut into 8-mm sections. Micro-wave antigen unmasking was done by using Dakoantigen retrieval solution (Dako). Sections were thenincubated with rabbit polyclonal antibodies againstNGF (1:100; Serotec, Inc.) for 2 hours. The specificity ofNGF antibody was tested and validated in our previousstudies (15). Human submandibular gland was used as apositive control as it is known to produce an abundanceof NGF. Normal rabbit serum containing mixed immu-noglobulins at the concentration of the primary antibodywas used as a negative control on the salivary glandtissue specimens. Immunoreactions were visualized withdiaminobenzidine chromogen (Vector Laboratories) andcounterstained with Mayer’s hematoxylin. This researchprotocol complies with the Committee on HumanResearch at the UCSF.

Reverse transcription PCR to quantify NGF mRNA inhuman tumors. Oral SCC and anatomically matched,contralateral normal oral epithelium from 11 oral cancerpatients were surgically removed and immediately snapfrozen in liquid nitrogen and stored at�80�C. Total RNAisolation of each sample was conducted with a Qiagen

DNA/RNA Kit (Qiagen Inc.) and 1 mL samples werealiquoted for RNA quantitative analysis. Reverse tran-scription was carried out with a High Capacity cDNAReverse Transcription Kit (Applied Biosystems Inc.) onthe biometra thermocyler (template 10 mL volumes perreaction). Quantitative real-time PCR assays were carriedout in triplicate with a Taqman Gene Expression AssayKit (Applied Biosystems Inc.). The housekeeping geneb-gus was chosen as the internal control. Controls con-sisted of total human brain RNA (�12 ng/mL; Ambion)and were negative in all runs. Relative quantificationanalysis of gene expression data was conducted accord-ing to the 2�DDCt method.

Cell culture of human oral SCC and control keratino-cytes. Human oral cancer cells (HSC-3) were cultivatedfor inoculation into mouse models of cancer pain andproliferation. The HSC-3 SCC cell line was obtained fromthe Japanese Collection of Research Bioresources (JCRB).The JCRB Cell Bank authenticated the cell line by usingshort tandem repeat analysis of loci with the PCR-basedPowerPlex1.2 system. Once we received the cells fromJCRB, the cells were expanded and frozen stocks wereprepared. For experiments, cells were used at low pas-sage for not more than 6 months before replenishing withfresh samples from the frozen stocks. Primary normaloral keratinocytes (NOK) harvested from normal oraltissues were cultured as previously described (25).

ELISA quantification of NGF in human oral cancercells. HSC-3 cells and NOK were grown to confluenceand then washed to remove unattached cells. The mediafor both HSC-3 and NOK were replaced with DK-SFMand incubated for an additional 72 hours. The condi-tioned media was then removed, centrifuged to removecell debris, aliquoted, and stored at �20�C. Cells werelysed in cold radioimmunoprecipitation assay buffercontaining protease inhibitor cocktails (Sigma-Aldrich).The lysatewas centrifuged (13,000 rpm for 5minutes) andthe supernatant was removed, aliquoted, and stored at�20�C. NGF concentration was measured by using theNGF Emax Immunoassay ELISA Kit (Promega).

Proliferation assay with anti-NGF in human oralcells. To quantify the effect of NGF on cancer cellproliferation, 3 � 103 HSC-3 cells were seeded in indivi-dual wells. Control groups received PBS (10 mL) andexperimental groups receivedNGFneutralizing antibody(R&D Systems) in PBS (10 mL of 0.025 mg/mL). At pre-determined time points (0, 24, 48, and 72 hours), 20 mL ofMTS (Promega BioSciences) was added to each well for 1hour and incubated at 37�C. The samples were thenquantified by MTS colorimetric assay at 490 nm. Theexperiment was repeated in triplicate.

Mouse cancer modelsBehavioral mouse models of human oral cancer pain.

Six- to 8-week-old female athymic, immunocompromisedmice (BALB/c) were purchased from Charles RiverLaboratories. They were housed in a temperature-controlled room on a 12:12 light/dark cycle (6 AM to 4

Ye et al.

Mol Cancer Ther; 10(9) September 2011 Molecular Cancer Therapeutics1668




PM light), with ad libitum access to food and water. TheUCSF Committee on Animal Research approved allprocedures and researchers were trained under the Ani-mal Welfare Assurance Program.Paw model. The paw-withdrawal cancer pain mouse

model was produced as previously described (26). Adultfemale nudemice were inoculatedwith 106 HSC-3 cells in50mL ofDulbecco’sModified EagleMediumandMatrigelinto the plantar surface of the right hind paw.Tongue model. To create a mouse model that is more

biologically homologous with human oral cancer, micewere inoculated with 50 mL of 106 HSC-3 cells intothe floor of the mouth as previously described (27).The anatomic and functional features of this mousecancer model parallel those found in human patientswith oral cancer (27).Anti-NGF treatment and control groups.Paw. In the mouse paw tumor model, anti-NGF anti-

body (Mab 256; R&D Systems; 12.5 mg in 20 mL PBS) orvehicle control (20 mLPBS)was injected into the right hindpaw of mice starting on postinoculation day, 4 followingthe pain behavior measurement and twice a week there-after until postinoculation day 21. Dosage of anti-NGFused was based on a study by Adriaenssens and collea-gues (14). Mice were randomly placed into 4 treatmentgroups: group 1 received an injection of HSC-3 cells andanti-NGF treatment (tumor þ anti-NGF, n ¼ 7); group 2received an injection of HSC-3 cells and PBS (vehiclecontrol, tumorþ PBS, n¼ 7); group 3 received an injectionof HSC-3 in the right paw and anti-NGF in the contra-lateral (CL) paw to see whether anti-NGF has a systemiceffect (tumor þ CL-anti-NGF, n ¼ 5); group 4 was treatedwith anti-NGF to determine whether anti-NGF is hypo-analgesic in naive mice (naive þ anti-NGF, n ¼ 5). Allgroupsofmicewerebriefly anesthetizedwith inhalationalisoflurane (Summit Medical Equipment Company)during HSC-3 inoculation and drug treatments.Tongue. In the mouse tongue cancer model, 2 groups

of mice were used. The control group (n ¼ 10) receivedisotype immunoglobulin G (50 mg in 50 mL PBS; R&DSystems). The anti-NGF treatment group (n ¼ 10)received 50 mg of the anti-NGF antibody in 50 mL PBS.All injections were intraperitoneal and administeredtwice per week starting at postinoculation day 13, whenall mice exhibited visible tumor masses and increasedgnaw time. We were concerned that repeated local injec-tion of anti-NGF into the tonguewould affect the rodent’seating and gnawing behavior so we chose a systemicroute of injection (intraperitoneal). Higher doses of sys-temic anti-NGFwere used in the tonguemodel comparedwith the dose given in the paw model to ensure enoughantibodies reached the tongue tumor.Behavioral measurement. Testing was done by an

observer blinded to the experimental groups as pre-viously described (25). The paw withdrawal thresholdwas measured by an electronic von Frey anesthesiometer(IITC Life Sciences). Paw withdrawal threshold wasdefined as the force in grams (mean of 8 trials) sufficient

to elicit a distinct paw withdrawal flinch upon applica-tion of a rigid probe tip.

Dolognawmeter. The dolognawmeter is a validateddevice/assay invented to measure oral function andnociception in mice (27). Mice with tongue tumors wereevaluated twice per week with a dolognawmeter aspreviously described (27). In brief, each mouse wasplaced into a confinement tube with 2 obstructing dowelsin series. The mouse voluntarily gnaws through the 2dowels to escape from confinement within the tube. Eachobstructing dowel is connected to a digital timer. Whenthe dowel is severed by the gnawing of the mouse, thetimer is automatically stopped and records the durationof time to sever each of the 2 dowels. To acclimatize themice and improve consistency in gnawing duration, allmice were trained for 10 sessions in the dolognawmeter.Training involved placing the animals in the device andallowing them to gnaw through the obstructing dowels inexactly the same manner that they do so during thesubsequent experimental gnawing trials. A baselinegnaw-time value to sever the second dowel was estab-lished for each mouse as the mean of the final 3 trainingsessions. After baseline gnaw times were established foreach mouse, the mice were inoculated with cancer cells.

Tumor size and body measurement. Mouse hind-pawvolume was measured by using a plethysmometer (IITCLife Science). Tongue tumor volumewas calculated at theend of the experiment by multiplying tumor length bywidth by thickness. Body weight was recorded beforeeach behavioral test. Mice did not show any significantweight changes during baseline training trials.

Tissue and blood processing. At the end of the experi-ment, mice with either cancer model were sacrificed withisoflurane. Both tumor and normal paws were dissectedand stored at �80�C in preparation for NGF proteinquantification. The paw samples were homogenized,lysed, centrifuged, and the supernatant was removed.Total protein concentration in each sample was deter-mined by using a BCA protein assay (Thermo Scientific).NGF concentration was measured by using the samemethod as previously described. In mice with tonguecancer, blood was rapidly collected from the heart intoEDTA-coated tubes. Plasma was separated with a cen-trifuge and stored at�20�C. Because adipose tissue is themain source of leptin, abdominal fat was also rapidlydissected out and stored at �80�C. Following thesemanipulations, mice were perfused transcardially with0.1 mol/L PBS followed by 4% PFA. Trigeminal gangliawere harvested in preparation for sensory receptorand ion channel evaluation. Both trigeminal gangliaand tongues were removed, postfixed in 4% PFA, andcryoprotected in sucrose gradient (20%–50%, 4�C). Serialsections of frozen trigeminal ganglia (10 mm) and tongue(20 mm) were cut on a cryostat and thaw mounted ongelatin-coated slides for processing.

Immunohistochemistry for sensory receptors and pro-liferation. After sectioning, trigeminal ganglia andtongue sections were briefly rinsed in PBS, incubated

Anti-NGF as a Therapy in Head and Neck Cancer

www.aacrjournals.org Mol Cancer Ther; 10(9) September 2011 1669




in goat serum (5% in PBS with 0.1% Triton X-100) for 1hour, and then incubated overnight in the primary anti-body. Trigeminal ganglia were stained for PAR-2 (goatanti-PAR2, 1:200; Santa Cruz Biotechnology), TRPV1(rabbit anti-TRPV1, 1:400; Fisher Scientific), and TRPA1(rabbit anti-TRPA1, 1:200; Abcam). Tongue sections werestained for Ki67 (rabbit anti-Ki67, 1:400; Abcam), anuclear protein used to evaluate proliferation. Afterincubation in primary antibody, sections were rinsedin PBS 3 times for 10 minutes each and then incubatedin the FITC-AffiniPure goat anti-rabbit secondary anti-body (1:400; Jackson ImmunoResearch Laboratories) for 1hour at room temperature. Image analysis was carriedout by NIH Image J software. The area of staining wasoutlined and pixel density within the selected area wasthen measured and divided by the total area. Data werecollected from 4 randomly selected sections from aminimum of 5 animals per treatment group.

ELISA measurement for cytokines. Plasma IL-6 andTNF-a were measured by using an ELISA kit fromeBioscience, Inc. Plasma and fat leptin were measuredby using the mouse Leptin Quantikine ELISA Kit (R&DSystems). The optical density of the standards and sam-ples was read at 450-nm wavelength by using a Model680 Microplate Reader (Bio-Rad Laboratories, Inc.). Allsamples were run in triplicate.

Statistical analysisThe statistics software SigmaPlot for Windows (version

11.0) was used to carry out all data analysis. Student’s t testor Mann–Whitney U test was used to compare mean ormedian for the2groups.Repeated-measuresANOVAwithone within-subject factor (time) and one between-subjectfactor (treatment) followed by Holm–Sidak post hoc testswere used to compare the effect of different treatmentsovertime. Simple linear regressionwasused to examine the

correlation of cytokines with changes in body weight,gnawing time, and tumor size. In the tonguemodel, 3 micein the control group were euthanized at day 22 due toadvanced cancer and severe cachexia. To best model thetrend over time, missing values due to death were treatedby using the last observation carried forward method fordata analysis and figure presentation. To make sure thatour results are not biased by thismethod,we also analyzedour data by including the missing values and found thegeneral conclusions and statistical significance were notaffected. P < 0.05 was considered statistically significant.Results are presented as mean � SEM.

Results

NGF mRNA and protein levels were significantlyelevated in oral cancer

Tissue biopsies from 14 oral SCC patients showedstrong NGF immunoreactivity (Fig. 1A) whereas thenormal oral epithelium from the same patients showedextremely low NGF labeling (Fig. 1B). NGF mRNA inSCC tumorswas approximately 8.9 times higher than thatin normal oral tissues (Fig. 1C). NGF protein concentra-tion was also compared between HSC-3 and NOK cells.Both cell lysate and supernatant of HSC-3 cells containedmuch higher NGF levels than those of NOKs (89.2 � 3.9vs. 21.3� 6.6 pg/mL in lysate; 56.6 � 3.7 vs. 20.2 � 5.9pg/mL in supernatant, respectively; Fig. 1D). Evengreater elevation of NGF protein was found in cancertissues collected from the mouse paw cancer model(30.3 � 5.1 vs. 4.7 � 2.0 pg/mg, respectively; Fig. 1E).

Anti-NGF reduced proliferation in culture and invivo in the tongue

In HSC-3 cell culture, anti-NGF treatment significantlydecreased cell proliferation at 24 (P < 0.01), 48 (P < 0.05),

Figure 1. NGF is elevated inhuman cancer and mouse modelsof oral cancer. A, immunohis-tochemistry shows strongNGF immunoreactivity in arepresentative tissue biopsy froman oral cancer patient. B, incontrast, normal oral epitheliumshows faint NGF immuno-reactivity. C, real-time PCRquantification of NGF mRNA inoral SCC biopsies shows a largeincrease over normal tissue. D,NGF protein concentrations inHSC-3 cell culture lysate andsupernatant are much higher thanNOKs. E, NGF protein is muchhigher in the tumor-bearing paw.Scale bar, 100 mm. *, P < 0.05;**, P < 0.01; ***, P < 0.001 versuscontrol. Two-tailed Student's ttest was used for data analysis.

Ye et al.





and 72 hours (P < 0.01) after treatment when comparedwith the corresponding control group (Fig. 2A). In thetongue tumormousemodel, anti-NGF treatment reducedtumor volume to approximately half of that in the controlgroup (P < 0.05; Fig. 2C). We confirmed that the decreasein tumor volume correlated with a decrease in cancer cellproliferation through quantification of Ki-67–positivecells in tongue tumors (P < 0.05; Fig. 2D and E). In thepaw cancer model, however, all tumor groups hadsignificantly larger right paw size than the group withouttumor (P < 0.001). In the anti-NGF–treated group, astatistically insignificant trend of tumor volume reduc-tion was found (Fig. 2B).

Anti-NGF attenuated nociception in the cancermodel animalsPaw model. Injection of anti-NGF into the mouse

paw tumor reversed tactile allodynia by an average of25% relative to that seen in both tumor þ PBS (P < 0.001)and tumor þ CL-anti-NGF groups (P < 0.001) throughoutthe observation period. Starting on postinoculation day 4,implantation of SCC cells into the right hind pawsproduced significant mechanical allodynia (Fig. 3A)consistent with our previous report (28). In naive mice,anti-NGF treatment did not cause a hypoalgesiceffect (P ¼ 0.53; Fig. 3A). Both tumor þ PBS and tumor

þ CL-anti-NGF mice showed a decrease of the samemagnitude in paw withdrawal threshold compared withnaive þ anti-NGF mice.

Tongue model. In the tongue cancer model, all micedeveloped visible tumor masses by postinoculation day7. Anti-NGF treatment completely abolished the progres-sive increase in gnaw time caused by cancer at day 21(P < 0.05), day 23 (P < 0.01), and day 27 (P < 0.01; Fig. 3B).Tumor size did not correlate positively with gnaw time(P ¼ 0.35).

Anti-NGF reduced TRPV1, TRPA1, and PAR-2labeling

Anti-NGF treatment led to a 26% reduction in TRPV1labeling (P < 0.05), a 52% reduction in TRPA1 labeling(P < 0.001), and a 15% reduction in PAR-2 labeling intrigeminal ganglion cells (P < 0.05; Fig. 3C and D).

Anti-NGF prevented cancer-induced weight lossPaw tumor groups treated with anti-NGF did not

develop significant weight loss, whereas both tumor þPBS mice lost more than 5% of their baseline body massstarting at the beginning of the third week (Fig. 4A).Unexpectedly, naive mice that were treated only withanti-NGF also lost more than 5% body mass relative totheir baseline at the beginning of the third week (Fig. 4A).

Figure 2. Anti-NGF therapy decreased cancer proliferation in human cancer cells and mouse models. A, anti-NGF decreased cell proliferation rate ofcultured HSC-3 cells at 24, 48, and 72 hours. B, anti-NGF treatment did not affect tumor size in the mouse paw model. All the tumor groups had significantlylarger paw volumes compared with naive mice (P < 0.001). C, anti-NGF treatment greatly reduced tongue tumor size. D, Ki-67 immunointensity wasdecreased in a representative tongue tumor following anti-NGF treatment. E, quantification of Ki-67–positive cells in total 40,6-diamidino-2-phenylindole(DAPI)-positive cells showed significantly less Ki-67 activity following anti-NGF treatment. Horizontal scale bar, 100 mm. *, P < 0.05; **, P < 0.01; ***, P < 0.001.A, C, and E, 2-tailed Student's t test or Mann–Whitney U test was used; B, 2-way ANOVA multiple comparisons were used.






At days 18 and 21, anti-NGF–treated tumor micewere significantly heavier than tumor þ PBS and naiveþ anti-NGF mice. In the tongue cancer model, by the endof the experiment, the anti-NGF–treated mice retainedtheir original body mass, whereas the control mice lost10% of their body mass. Significant differences in percentweight change were found at days 17 (P < 0.05), 21(P < 0.001), 23 (P < 0.01), and 27 (P < 0.01; Fig. 4B). Thetransient body mass loss immediately following cancercell inoculation (Fig. 4B) can be attributed to attenuatedfeeding secondary the transient tongue trauma.

Anti-NGF modulated cytokine levelsMice treated with anti-NGF exhibited 50% lower

plasma TNF-a and IL-6 (Fig. 5A and B) and 3 to 4 timeshigher leptin levels in plasma as well as in adipose tissue(the main source of leptin; Fig. 5C and D).

Cytokines correlated positively with tumor size,nociception, and weight loss

In the tongue tumor mousemodels, IL-6 was positivelycorrelated with tumor size (R ¼ 0.5, P < 0.05), gnaw time(R ¼ 0.5, P < 0.05), and weight loss (R ¼ 0.6, P < 0.05;

Fig. 6A–C). TNF-a was positively correlated with weightloss (R ¼ 0.8, P < 0.001; Fig. 6D) but not with tumor size(P ¼ 0.16) or gnaw time (P ¼ 0.10). Plasma and adiposeleptin concentrations were both inversely correlated withbody mass loss (R ¼ �0.6, P < 0.05; R ¼ �0.5, P < 0.05,respectively; Fig. 6E and F) but not significantly corre-lated with tumor size (P ¼ 0.1 and P ¼ 0.86, respectively)or gnaw time (P ¼ 0.26 and P ¼ 0.12).

Discussion

Our results show that NGF affects progression of oralSCC as well as pain and cachexia associated with thiscancer in patients as well as in mouse models. It does so,in part by increasing TNF-a and IL-6, by decreasingleptin, and by increasing expression levels of nociceptivesensory receptors. We showed that NGF productionincreased in the orthotopic model of human oral SCCinmice as well as in oral SCC cell culture. Moreover, NGFblockade with antibodies reduced tumor proliferation,nociception, and loss of body mass. In summary, NGFblockade (i) yielded lower levels of TNF-a and IL-6,(ii) upregulated leptin, and (iii) downregulated the

Figure 3. Anti-NGF therapy decreased cancer-induced nociception. A, anti-NGF treatment greatly attenuated mechanical allodynia in cancer-bearingpaws of mice, but had no effect on basal mechanical sensitivity. Two-way repeated-measures ANOVA showed a significant treatment effect(F ¼ 511.677, P < 0.001) and a time � treatment interaction effect (F ¼ 1.661, P < 0.05); the anti-NGF group was significantly different from all other groups(P < 0.001). B, in the tongue cancer model, 2-way repeated-measures ANOVA showed a significant main effect of treatment (F ¼ 10.16, P < 0.001) with asignificant treatment� time interaction (F¼ 2.29, P < 0.05). Mice treated with anti-NGF did not show the escalation in gnaw time but was shown by the controlmice starting postinoculation day 21. C, immunohistochemistry of a representative trigeminal ganglion from an anti-NGF–treated tongue cancer mouseshowed lower expression of TRPV1, TRPA1, and PAR-2 receptors. Scale bar, 100 mm. D, anti-NGF treatment markedly reduced TRPV1, TRPA1, and PAR-2immunointensity in trigeminal ganglia of mice with tongue cancer. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Ye et al.





nociceptive receptors TRPV1, TRPA1, and PAR-2 intrigeminal ganglia.The effect of anti-NGF on tumor proliferation varies

with tumor type, tumor location, and route of adminis-tration. In this study, we evaluated the effect of anti-NGFon tumor growth in both a paw and a tongue model. Forthe pawmodel, we administered the anti-NGF locally in amanner similar to the approach used by Adriaenssensand colleagues in their study of the effect of anti-NGF in a

breast cancer model (14). For the tongue model, weadministered anti-NGF systemically to preclude thepossibility that repeated injection into the tongue wouldaffect feeding behavior and maintenance of body mass.

NGF is tumor promoting for a variety of cancers andanti-NGF has been shown to reduce tumor proliferation.Andriaenssens and colleagues showed that locallyinjected anti-NGF (12.5 mg) decreased tumor size in abreast cancer model (14). However, we found that localinjection of anti-NGF (12.5 mg) in a paw model did nothave a significant effect on tumor growth. In our tonguecancer model, anti-NGF significantly decreased tumorsize in animals given a larger systemic dose. The sup-pressive effect of anti-NGF on cancer proliferation in vivois confirmed by an anti-Ki-67 assay in the tonguecancer tissue sections and corroborated with an in vitroproliferation assay.

NGF promotes oral cancer progression in part througha mechanism that involves IL-6. We investigated whetherNGF could exert its effect by modifying cytokine levelsand showed that anti-NGF reduces IL-6. We also foundthat IL-6 was positively correlated with tumor size. Thiseffect is supported by the clinical finding that patientswith advanced oral cancer exhibit increased IL-6. IL-6expression also correlates with poor prognosis (29, 30).Furthermore, IL-6 activates the transcription factorNF-kB and the STAT3 signal transduction pathway,which in turn regulate the expression of genes controllingcell proliferation and apoptosis in oral cancer (29). Addi-tional mechanisms are possible. For example, NGF couldalso promote progression by interactingwith its receptorsTrkA and p75 in both an autocrine and paracrine manner

Figure 4. Anti-NGF differentially modulates body weight in normal andcancer-bearing mice. A, 2-way repeated-measures ANOVA showed asignificant main effect of treatment (F ¼ 10.16, P < 0.001) and time(F ¼ 7.63, P < 0.001). The anti-NGF–treated group was significantlydifferent from tumor þ PBS and naive þ anti-NGF (P < 0.001), but was notdifferent from tumor þ CL-anti-NGF group (P ¼ 0.09). Vehicle-controlledtumor mice had significant weight loss from their baseline at day 18 (a). Incontrast, no weight loss is shown for anti-NGF–treated mice. Note thatnaive mice treated with anti-NGF had significant weight loss at day 18 (b).At days 18 and 21, tumor þ anti-NGF group had significantly larger bodymass compared with tumorþ PBS or naiveþ anti-NGF groups. B, in micewith tongue tumors, 2-way repeated-measures ANOVA showed asignificant main effect of treatment (F¼ 11.53, P < 0.001) with a significanttreatment � time interaction effect (F ¼ 2.8, P < 0.01). Anti-NGF treatmentprevented weight loss starting at day 17. *, P < 0.05; **, P < 0.01;***, P < 0.001; #, P < 0.05 tumor þ anti-NGF versus tumor þ PBS;##, P < 0.01 tumor þ anti-NGF versus tumor þ PBS; þþ, P < 0.01tumor þ anti-NGF versus naive þanti-NGF; þþþ, P < 0.001 tumorþ anti-NGF versus naive þ anti-NGF.

Figure 5. Anti-NGF in tongue cancer mouse models decreased plasmaTNF-a (A) and IL-6 (B), and increased plasma and fat leptin (C and D).*, P < 0.05; **, P < 0.01; ***, P < 0.001 versus control. Two-tailedStudent's t test was used for data analysis.






(31). NGF may also exert its effect on tumor progressionby promoting angiogenesis. NGF has been shown to playa role in angiogenesis in ovarian and breast cancer (14, 32,33), and NGF inhibition strongly reduces angiogenesisand tumor development in mice (14).

Previous studies showed that treatment with anti-NGFhas a strong antinociceptive effect in animal models ofbone cancer (16, 17). Here we present evidence that anti-NGF also exhibits a strong antinociceptive effect in micewith soft-tissue cancers. One of the pathways by whichNGF induces pain and hyperalgesia entails modulationof expression and function of nociceptive receptors andsensory ion channels. For example, the TRPV1 channel isknown to play a role in cancer pain (3). TRPA1 sensorychannels have also been recently reported to play impor-tant roles in orofacial pain (34) and PAR-2 is involved inanimal models of oral cancer pain (25). NGF upregulatesTRPV1 and TRPA1 in several painful states (11, 34, 35),but a link between NGF and PAR-2 is less certain. In thisstudy we found that NGF not only modulates TRPV1 andTRPA1 expression, but also upregulates PAR-2 expres-sion. Most importantly, anti-NGF reduced TRPA1 byalmost 50%.We infer from this finding that TRPA1mightplay an important role in mediating oral cancer pain.

In agreement with previous findings (16, 17), we con-clude that the antinociceptive effect of anti-NGF does notstem from a reduction in tumor proliferation alone. First,in the mouse pawmodel, we did not observe a significantreduction in tumor size following anti-NGF administra-tion but a significant antinociceptive effect was found.

Second, in the tongue cancer model, no correlationwas found between cancer size and gnaw time. Third,anti-NGF reduced expression of nociceptive receptorsTRPV1, TRPA1, and PAR-2 in trigeminal ganglion cells.This finding shows that anti-NGF reduces pain at least inpart by decreasing nociception transduction and signal-ing via these receptors.

Perhaps the most novel finding of our study is thatNGF is associated with cancer-induced cachexia. Theunexpected finding that anti-NGF given to naive miceled to reduced body mass leaves open the possibility thatthat a basal level of NGF is necessary for maintenance ofnormal body mass. If basal NGF is too low, body mass islost. However, when NGF is abnormally elevated,cachexia results. For example, in animal models of arthri-tis, NGF levels are elevated and anti-NGF blocks loss ofbody mass. These findings support the conclusion thatelevated levels of NGF contribute to arthritis-inducedcachexia (36). The manner by which NGF exerts its effecton body mass regulation is not entirely clear. A variety ofcytokines linked to inflammation, including TNF-a, IL-6,and leptin, have been proposed to play a role inbody mass regulation and cachexia (2, 22). TNF-a andIL-6 have been proposed to induce cancer cachexiathrough inflammation, altered metabolism, appetite sup-pression, and enhanced lipolysis and proteolysis (23).Elevated IL-6 and TNF-a levels have been found incachectic cancer patients (22). Both of these cytokinescorrelate inversely with body mass index (BMI) inpatients with gastrointestinal cancer (37, 38). Leptin is

Figure 6. Correlation of cytokines with cancer progression, nociception, and weight change in the tongue cancer model. A–C, plasma IL-6 is positivelycorrelated with tumor size, gnaw time, and weight loss. D, TNF-a is positively correlated with weight loss. E and F, plasma and fat leptin are inversely correlatedwith weight loss. Open circles, control group; closed circles, anti-NGF–treated group. Simple linear regression was used to test correlation.

Ye et al.





a cytokine that is produced mainly by adipocytes (23, 39).This cytokine regulates fat mass by decreasing the level ofneuropeptide Y in the hypothalamus and increasingresting energy expenditure (23). These changes resultin reduced food intake. The relationship between leptinand cancer cachexia is equivocal. Serum leptin is reducedin cachectic patients with cancers of the digestive organs(37) and ovaries (40). However, leptin levels are higherthan normal in breast cancer and prostate cancer patients(41, 42). In patients with oral cancer, reduced leptin levelsare accompanied by decreased body mass (38, 43, 44).Our results suggest that TNF-a, IL-6, and leptin all con-tribute to oral cancer–induced loss of body mass. Circu-lating levels of these molecules can be influenced bytargeting and manipulating NGF levels. Plasma levelsof leptin correlate with levels in adipose tissue. Becauseplasma leptin is proportional to levels in fat stores, it isnot surprising that untreated cancer mice have decreasedplasma leptin because of loss of fat mass. However, theability of adipocytes to produce leptin might have alsobeen reduced in untreated cancer mice. NGF is known tobe secreted from both murine and human adipocytes incell culture (45, 46), and fat cells express the high and lowaffinity NGF receptors TrkA and P75 (45). Therefore,NGF might act directly on adipocytes to modulate leptinrelease. Further studies are needed to elucidate themechanism through which NGF regulates body massand cachexia.Patients with advanced cancer often suffer from pain

and loss of body mass. Despite the general consensusthat tumor progression, pain, cachexia, and other

cancer-related symptoms result from production anddysregulation of proinflammatory cytokines (1, 8), nostudies have been conducted to investigate cancer-related symptoms in parallel. Such an approach mightallow us to identify cellular and molecular mediatorscommon to cancer-related symptoms for a specifictumor type.

We have identified NGF as a key endogenous factor inthe panoply of molecules involved in the orchestration ofcancer-related inflammation. NGF might be a moleculecommon to the mechanisms responsible for clinicallydistinctive cancer symptoms such as pain and cachexia.By suppressing proinflammatory cytokines and promot-ing anti-inflammatory cytokines, anti-NGF couldpotentially correct the altered cytokine balance observedin those who suffer from cancer. Inhibiting chronicinflammation by neutralizing NGF might be a noveland effective approach to treat pain and cachexiaassociated with certain cancers.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Grant Support

This work was funded by NIH/NIDCR R21 DE018561.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received February 25, 2011; revised June 22, 2011; accepted July 1, 2011;published OnlineFirst July 12, 2011.

References1. Reyes-Gibby CC, Wu X, Spitz M, Kurzrock R, Fisch M, Bruera E, et al.

Molecular epidemiology, cancer-related symptoms, and cytokinespathway. Lancet Oncol 2008;9:777–85.

2. Murphy KT, Lynch GS. Update on emerging drugs for cancercachexia. Expert Opin Emerg Drugs 2009;14:619–32.

3. Mantyh PW, Clohisy DR, KoltzenburgM, Hunt SP. Molecular mechan-isms of cancer pain. Nature Rev Cancer 2002;2:201–9.

4. Epstein JB, Elad S, Eliav E, Jurevic R, Benoliel R. Orofacial pain incancer: part II–clinical perspectives and management. J Dent Res2007;86:506–18.

5. Connelly ST, Schmidt BL. Evaluation of pain in patients with oralsquamous cell carcinoma. J Pain 2004;5:505–10.

6. Chasen MR, Bhargava R. A descriptive review of the factors con-tributing to nutritional compromise in patients with head and neckcancer. Support Care Cancer 2009;17:1345–51.

7. Couch M, Lai V, Cannon T, Guttridge D, Zanation A, George J, et al.Cancer cachexia syndrome in head and neck cancer patients: part I.Diagnosis, impact on quality of life and survival, and treatment. HeadNeck 2007;29:401–11.

8. Seruga B, Zhang H, Bernstein LJ, Tannock IF. Cytokines and theirrelationship to the symptoms and outcome of cancer. Nat Rev Cancer2008;8:887–99.

9. Hristova M, Aloe L. Metabolic syndrome–neurotrophic hypothesis.Med Hypotheses 2006;66:545–9.

10. Otten U, Marz P, Heese K, Hock C, Kunz D, Rose-John S. Cytokinesand neurotrophins interact in normal and diseased states. Ann N YAcad Sci 2000;917:322–30.

11. Nicol GD, Vasko MR. Unraveling the story of NGF-mediated sensi-tization of nociceptive sensory neurons: ON or OFF the Trks? MolInterv 2007;7:26–41.

12. Lewin GR, Rueff A, Mendell LM. Peripheral and central mecha-nisms of NGF-induced hyperalgesia. Eur J Neurosci 1994;6:1903–12.

13. Kr€uttgen A, Schneider I, Weis J. The dark side of the NGF family:neurotrophins in neoplasias. Brain Pathol 2006;16:304–10.

14. Adriaenssens E, Vanhecke E, Saule P, Mougel A, Page A, Romon R,et al. Nerve growth factor is a potential therapeutic target in breastcancer. Cancer Res 2008;68:346–51.

15. Kolokythas A, Cox DP, Dekker N, Schmidt BL. Nerve growth factorand tyrosine kinase A receptor in oral squamous cell carcinoma: isthere an association with perineural invasion?J Oral Maxillofac Surg2010;68:1290–5.

16. Halvorson KG, Kubota K, Sevcik MA, Lindsay TH, Sotillo JE, GhilardiJR, et al. A blocking antibody to nerve growth factor attenuatesskeletal pain induced by prostate tumor cells growing in bone. CancerRes 2005;65:9426–35.

17. Sevcik MA, Ghilardi JR, Peters CM, Lindsay TH, Halvorson KG, JonasBM, et al. Anti-NGF therapy profoundly reduces bone cancer pain andthe accompanying increase in markers of peripheral and centralsensitization. Pain 2005;115:128–41.

18. Mantyh WG, Jimenez-Andrade JM, Stake JI, Bloom AP, KaczmarskaMJ, Taylor RN, et al. Blockade of nerve sprouting and neuromaformation markedly attenuates the development of late stage cancerpain. Neuroscience 2010;171:588–98.






19. Chaldakov GN, Fiore M, HristovaMG, Aloe L. Metabotrophic potentialof neurotrophins:implication in obesity and related diseases? Med SciMonit 2003;9:HY19–21.

20. Taglialatela G, Foreman PJ, Perez-Polo JR. Effect of a long-term nervegrowth factor treatment on body weight, blood pressure, and serumcorticosterone in rats. Int J Dev Neurosci 1997;15:703–10.

21. Bullo M, Peeraully MR, Trayhurn P, Folch J, Salas-Salvado J. Circu-lating nerve growth factor levels in relation to obesity and the meta-bolic syndrome in women. Eur J Endocrinol 2007;157:303–10.

22. George J, Cannon T, Lai V, Richey L, Zanation A, Hayes DN, et al.Cancer cachexia syndrome in head and neck cancer patients: part II.Pathophysiology.Head Neck 2007;29:497–507.

23. Barton BE. IL-6-like cytokines and cancer cachexia: consequences ofchronic inflammation. Immunol Res 2001;23:41–58.

24. Takei Y, Laskey R. Interpreting crosstalk between TNF-alpha andNGF: potential implications for disease. Trends Mol Med 2008;14:381–8.

25. LamDK, Schmidt BL. Serine proteases and protease-activated recep-tor 2-dependent allodynia: a novel cancer pain pathway. Pain2010;149:263–72.

26. Schmidt BL, Pickering V, Liu S, Quang P, Dolan J, Connelly ST, et al.Peripheral endothelin A receptor antagonism attenuates carcinoma-induced pain. Eur J Pain 2007;11:406–14.

27. Dolan JC, Lam DK, Achdjian SH, Schmidt BL. The dolognawmeter: anovel instrument and assay to quantify nociception in rodent modelsof orofacial pain. J Neurosci Methods 2010;187:207–15.

28. Quang PN, Schmidt BL. Endothelin-A receptor antagonism attenu-ates carcinoma-induced pain through opioids in mice. J Pain2010;11:663–71.

29. Wang F, Arun P, Friedman J, Chen Z, Van Waes C. Current andpotential inflammation targeted therapies in head and neck cancer.Curr Opin Pharmacol 2009;9:389–95.

30. Pries R, Wollenberg B. Cytokines in head and neck cancer. CytokineGrowth Factor Rev 2006;17:141–6.

31. Papatsoris AG, Liolitsa D, Deliveliotis C. Manipulation of the nervegrowth factor network in prostate cancer. Expert Opin Investig Drugs2007;16:303–9.

32. Romon R, Adriaenssens E, Lagadec C, Germain E, Hondermarck H,Le Bourhis X. Nerve growth factor promotes breast cancer angiogen-esis by activating multiple pathways. Mol Cancer 9:157.

33. Davidson B, Reich R, Lazarovici P, Nesland JM, Skrede M, Risberg B,et al. Expression and activation of the nerve growth factor receptorTrkA in serous ovarian carcinoma. Clin Cancer Res 2003;9:2248–59.

34. Diogenes A, Akopian AN, Hargreaves KM. NGF up-regulates TRPA1:implications for orofacial pain. J Dent Res 2007;86:550–5.

35. Hefti FF, Rosenthal A, Walicke PA, Wyatt S, Vergara G, Shelton DL,et al. Novel class of pain drugs based on antagonism of NGF. TrendsPharmacol Sci 2006;27:85–91.

36. Shelton DL, Zeller J, HoWH, Pons J, Rosenthal A. Nerve growth factormediates hyperalgesia and cachexia in auto-immune arthritis. Pain2005;116:8–16.

37. Takahashi M, Terashima M, Takagane A, Oyama K, Fujiwara H,Wakabayashi G. Ghrelin and leptin levels in cachectic patients withcancer of the digestive organs. Int J Clin Oncol 2009;14:315–20.

38. Gharote HP, Mody RN. Estimation of serum leptin in oral squamouscell carcinoma. J Oral Pathol Med 39:69–73.

39. Tilg H, Moschen AR. Adipocytokines: mediators linking adiposetissue, inflammation and immunity. Nat Rev Immunol 2006;6:772–83.

40. Mor G, Visintin I, Lai Y, Zhao H, Schwartz P, Rutherford T, et al. Serumprotein markers for early detection of ovarian cancer. Proc Natl AcadSci U S A 2005;102:7677–82.

41. Ozet A, Arpaci F, YilmazMI, Ayta H, Ozturk B, Komurcu S, et al. Effectsof tamoxifen on the serum leptin level in patients with breast cancer.Jpn J Clin Oncol 2001;31:424–7.

42. Hsing AW, Deng J, Sesterhenn IA, Mostofi FK, Stanczyk FZ, BenichouJ, et al. Body size and prostate cancer: a population-based case-control study in China. Cancer Epidemiol Biomarkers Prev 2000;9:1335–41.

43. Mantovani G, Proto E, Massa E, Mulas C, Madeddu C, Mura L, et al.Induction chemotherapy followed by concomitant chemoradiationtherapy in advanced head and neck cancer: a phase II study fororgan-sparing purposes evaluating feasibility, effectiveness and toxi-city. Int J Oncol 2002;20:419–27.

44. Mantovani G, Massa E, Astara G, Murgia V, Gramignano G, FerreliL, et al. Six-week induction chemotherapy followed by concomi-tant chemoradiation therapy in stage IV head and neck cancer: aphase II study with organ-sparing purposes. Oncol Rep 2003;10:759–66.

45. Peeraully MR, Jenkins JR, Trayhurn P. NGF gene expression andsecretion in white adipose tissue: regulation in 3T3-L1 adipocytes byhormones and inflammatory cytokines. Am J Physiol EndocrinolMetab 2004;287:E331–9.

46. Wang B, Jenkins JR, Trayhurn P. Expression and secretion of inflam-mation-related adipokines by human adipocytes differentiated inculture: integrated response to TNF-alpha. Am J Physiol EndocrinolMetab 2005;288:E731–40.

Ye et al.





2011;10:1667-1676. Published OnlineFirst July 12, 2011.Mol Cancer Ther Yi Ye, Dongmin Dang, Jianan Zhang, et al. CachexiaNerve Growth Factor Links Oral Cancer Progression, Pain, and

Updated version

10.1158/1535-7163.MCT-11-0123doi:

Access the most recent version of this article at:

Cited articles

http://mct.aacrjournals.org/content/10/9/1667.full#ref-list-1

This article cites 44 articles, 6 of which you can access for free at:

Citing articles

http://mct.aacrjournals.org/content/10/9/1667.full#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles 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

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mct.aacrjournals.org/content/10/9/1667To request permission to re-use all or part of this article, use this link



http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-11-0123

http://mct.aacrjournals.org/content/10/9/1667.full#ref-list-1

http://mct.aacrjournals.org/content/10/9/1667.full#related-urls

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

mailto:[email protected]

http://mct.aacrjournals.org/content/10/9/1667


Nerve Growth Factor Links Oral Cancer Progression, Pain...

Documents

Transcript of Nerve Growth Factor Links Oral Cancer Progression, Pain...