PROGNOSTIC SIGNIFICANCE OF SERUM p53 PROTEINAND p53 ANTIBODY IN PATIENTS WITH SURGICALTREATMENT FOR HEAD AND NECK SQUAMOUSCELL CARCINOMA

Vivian Chow, M. Phil, 1 Anthony Po Wing Yuen, FRCS, 1 King Yin Lam, FRCPA, 2

Wai Kuen Ho, FRCS, 1 William Ignace Wei, FRCS 1

1 Department of Surgery, Queen Mary Hospital, Pokfulam Road, Hong Kong, SAR China. E-mail:pwyuen@hkucc.hku.hk2 Department of Pathology, The University of Hong Kong, Queen Mary Hospital, Hong Kong, SAR China

Accepted 25 September 2000

Abstract: Objectives. This study aims at investigating theprognostic values of serum p53 protein and anti-p53 antibody inpatients undergoing surgical treatment for head and neck squa-mous cell carcinoma (HNSCC).

Methods. Serum p53 protein and anti-p53 antibody concen-trations were determined by an enzyme-linked immunosorbentassay (ELISA) in 75 patients with HNSCC and 28 healthy con-trols. In 28 patients, formalin-fixed tumor tissues were also avail-able for immunohistochemical staining by an anti-p53 DO7monoclonal antibody. The results were correlated with the clini-copathologic parameters.

Results. The mean preoperative serum concentration of p53protein in patients with HNSCC was significantly higher thanhealthy controls (59.45 pg/mL vs 16.4 pg/mL, p = .007). Preop-erative serum p53 antibody was present in 23 (31%) patients andwas present in one healthy control. Eighteen (62%) tumor tissuesshowed p53 overexpression by immunohistochemistry. The pres-ence of serum anti-p53 antibody before operation was associatedwith a significantly higher incidence (65%) of nodal metastasis

compared with 27% nodal metastasis in patients with absence ofserum anti-p53 antibody (p = .002).

Conclusion. Preoperative serum p53 antibody was a signifi-cant prognostic factor for nodal metastasis of HNSCC. © 2001John Wiley & Sons, Inc. Head Neck 23: 286–291, 2001.

Keywords: ELISA; immunohistochemistry; cancer; metastasis

The p53 gene is a tumor suppressor gene locatedat chromosome 17p13.1. It encodes a 53 kDanuclear phosphoprotein (p53 protein), which con-trols the cellular proliferation and transforma-tion.1 Mutation of p53 gene is the most commongenetic abnormality of human cancer.2 Mutationsand overexpression of p53 are common in headand neck squamous cell carcinoma (HNSCC).3–7

Overexpression of p53 had been reported in 60%of laryngeal carcinomas,5 37% of hypopharyngealcarcinomas,6 and 52% of tongue carcinomas.7

The p53 protein in cancer cells may be re-leased into the extracellular compartment and isabsorbed into the circulation. The serum p53 pro-tein level can be measured quantitatively by us-

Correspondence to: A. P. W. YuenContract grant sponsor: Betty and Michael Kadoorie Cancer GeneticsResearch Foundation and The University of Hong Kong.

© 2001 John Wiley & Sons, Inc.

286 Serum p53 Protein and Antibody of Head and Neck Carcinoma HEAD & NECK April 2001

ing the enzyme-linked immunosorbent assay(ELISA). The serum p53 protein level in carci-noma of colon, lung, and pancreas8–10 was signifi-cantly raised compared with normal controls. Se-rum p53 protein level, however, has not beenreported in HNSCC.

The accumulation of mutant p53 protein alsoleads to the production of anti-p53 antibodies.These antibodies are mostly IgG, correspondingto a secondary immune response. p53 antibodieshave been reported in 17% patients withHNSCC.11 The presence of this p53 antibody wasfound to correlate with a poor prognostic outcomein HNSCC patients. The prognostic significanceand pattern of serial changes of serum p53 anti-body levels before and after surgical treatment ofhead and neck cancer are still unknown.

This study mainly aims at evaluation of theprognostic significance of preoperative serum p53protein and p53 antibody in patients who under-went surgical treatment for HNSCC.

PATIENTS AND METHODS

Seventy-five Chinese patients with HNSCC whounderwent primary surgical treatment in QueenMary Hospital, Hong Kong, between 1993 and1998 were selected for the study. All patients hadno prior radiotherapy or chemotherapy beforesurgical treatment. Blood samples and tumor tis-sues were obtained before operation. In 16 of the75 patients, blood samples after resection of tu-mor were also available. We also obtained bloodsamples from 28 healthy volunteers as normalcontrols. The study has been approved by the eth-ics committee of the Faculty of Medicine of TheUniversity of Hong Kong.

There were 56 men and 19 women withHNSCC. The median age was 62 years (range,32–86 years). The tumor sites were 49 oral cavity,4 oropharynx, 11 hypopharynx, and 11 larynx.The clinical stages before operation were 19 stageI, 19 stage II, 14 stage III, 23 stage IV; and allpatients were M0. All patients had primary sur-gical treatment, and 35 patients were also givenpostoperative adjuvant radiotherapy. The medianpostoperative follow-up duration of patients alivewithout tumor was 22 months.

Blood was collected with EDTA-coated bottleand was centrifuged at 800 g for 10 minutes. Theserum and white cells were separately collectedand were stored at −80°C until analysis. Tumortissues were taken from the resected specimens.They were fixed in 10% buffered formalin.

The tumors were classified into grades I, II,

and III in cellular differentiation (equivalent towell, moderately, and poorly differentiated tu-mors) according to WHO classification.12 Theclinical staging of patients was defined accordingto the criteria of International Union AgainstCancer in 1997.13

P53 Protein ELISA. The ELISA was done accord-ing to the manufacturer’s protocol (All reagents ofthe ELISA were supplied by Boehringer Mann-heim, Germany). Sera and standard sampleswere diluted 50 times and were put into the wellsof the ELISA plate; 100 mL anti-p53 protein an-tibody (anti-p53-POD, prediluted, BoehringerMannheim, Germany) were added. The plate wasincubated for 2 hours at room temperature. Afterthe incubation, incubation buffer was removedthoroughly by tapping off. The wells were rinsedfive times with washing buffer; 200 mL substratesolution for color development was added intoeach well, and the plate was again incubated atroom temperature for 20 minutes in dark. Thereaction was stopped with the stop solution pro-vided by the manufacturer. The color was thenquantified by examining its absorbance at 450 nmusing a spectrophotometer (with reference wave-length: 690 nm) against blank solution (distilledwater). The concentration of p53 protein in serumsamples were determined by comparison againsta standard curve generated from standardsamples with known amount of p53 protein (0,56.78, 162.45, 370.15, 568.76, and 1147.06 pg/mL).

Anti-p53 ELISA. Two distinct wells (one coatedwith p53 protein and another one coated with con-trol protein) were used for the test of each sample(All reagents of the ELISA were supplied by Phar-maCell, Paris, France). Two positive controls eachat different concentrations and one negative con-trol provided by the manufacturer were also used.The serum samples and the positive controls werediluted 100 times. All samples were tested twicein two separate plates; 100 mL each of the pa-tients’ serum samples, positive controls, andnegative control were added to the wells and wereincubated for 60 minutes at room temperature.After the incubation, washing was done withwashing buffer for four times. After the finalwashing, residual fluid was removed by tappingthe plate on a paper towel. One hundred microli-ters conjugate solution (diluted 1000 times) wasadded to each well. The reaction mixtures werethen incubated for 60 minutes at room tempera-

Serum p53 Protein and Antibody of Head and Neck Carcinoma HEAD & NECK April 2001 287

ture. After washing, 100 mL substrate was addedin each well and was incubated for 10 minutes atroom temperature in dark; 100 mL stop solutionwas added to stop the enzymatic reaction. Thecolor of the reaction mixture changed from blue toyellow. The reaction products were read with aspectrophotometer at reference wavelength (450nm) against blank solution (distilled water). TheELISA absorption data of our study were calcu-lated according to the manufacturer’s formula.

Immunohistchemical Staining. Immunohisto-chemical staining was carried out with avidin-biotin complex immunoperoxidase technique. Theembedded specimens were cut at 4 mm, dewaxedin xylene for 15 minutes, and rehydrated withethanol. The slides were treated with 3% hydro-gen peroxide for 30 minutes at room temperatureand proteinase K at 37°C for 25 minutes. Afterthree washes in phosphate-buffered saline (PBS),the antigen retrieval was performed by micro-wave in citrate buffer (pH 6) for 4 minutes. Weused the monoclonal p53 antibody at a dilution of1:50 (NCL-p53-DO7, Novocastra LaboratoriesLtd., Newcastle Upon Tyne, UK). The incubationwas carried out overnight at 4°C. The sectionswere incubated with secondary antibody for 30minutes. Staining was performed by using avidin-biotin reagents (Dakopatts, Gloustrup, Denmark)3,38-diaminobenzidine, and hydrogen peroxide.Afterwards, the sections were counterstained byMayer’s hematoxylin. A semiquantitative scoringsystem was used according to the characteristicstaining pattern (nuclear staining) by a 4-pointscale (0, 0%–10%; +, 11%–25%; ++, 26%–50%,+++, >50%). A HNSCC with known p53 overex-pression was used as positive and negative con-trol at each IHC procedure. The negative controlwas done by omission of the primary antibodyduring the IHC procedure. Overexpression of p53was defined for specimens with score +, ++, or+++. The slides were reviewed with consensusagreement by both Yuen and Lam, and the clini-cal data were not known at the time of scoring ofslides.

RESULTS

Of the 28 sera from healthy blood donors, 5 (18%)had detectable serum p53 protein. All 5 healthydonors were nonsmokers and nondrinkers with-out clinical evidence of cancer. Of the 75 patientswith HNSCC, 68 (91%) had detectable serum p53protein in the preoperative blood. The mean se-rum p53 protein concentration of the HNSCC pa-

tients was 59.4 pg/mL, which was significantlyhigher than the mean concentration of 16.4 pg/mL of the healthy donors (t test, p 4 .007).

Twenty-three (31%) HNSCC patients and one(4%) healthy control had detectable serum p53antibody (chi-square, p 4 .001). The one healthycontrol with detectable serum p53 antibody was anonsmoker and nondrinker without clinical evi-dence of cancer. The mean serum p53 proteinlevel of patients with positive serum p53 antibodywas 36.5 pg/mL compared with 69.6 pg/mL of pa-tients with negative serum p53 antibody (t test,p 4 .08).

Of the 29 patients who had suitable tumor tis-sues available for immunohistochemical staining,18 (62%) specimens had overexpression of p53(semiquantitative score: 11+, 6++, 1 +++). Themean serum level of p53 protein was 37.3 pg/mLfor patients with p53 overexpression comparedwith 58.9 pg/mL for patients with no overexpres-sion of p53 in the tumor tissues (t test, p 4 .08).The serum p53 antibody was positive in 5 of 18(28%) patients with overexpression of p53 com-pared with 4 of 11 (36%) patients with no overex-pression of p53 in their tumor tissues (Fisher’stest, p 4 .69).

The preoperative serum p53 protein and p53antibody were not significantly correlated withsites of tumor, sex, age, differentiation, clinicalstage, smoking, and drinking (all with respectivep value > .05). Nodal metastasis (the presenceis evidenced either by clinically positive metasta-sis node, nodal recurrence, or pathologic evidenceof nodal metastasis in the neck dissection speci-mens) was present in 29 (38.7%) patients. Preop-erative serum p53 protein level was not associ-ated with nodal metastasis (t test, p 4 .265).Preoperative serum p53 antibody was found to besignificantly associated with nodal metastasis;nodal metastasis was present in 15 of 23 (65%)patients with positive serum p53 antibodiescompared with 14 of 52 (27%) patients with un-detectable serum p53 antibody (chi-square test,p 4 .002).

There were 27 patients who had recurrenceafter operation at a mean postoperative durationof 8 months (range, 2– 20 months). The first sitesof clinical recurrence were 7 (9.3%) local, 12(16.1%) nodal, 6 (8%) distant, 2 (2.7%) both localand distant. The preoperative presence of serump53 protein and p53 antibodies had no significantcorrelation with tumor recurrence (all with p >.05). The overall 5-year survival rate of the 75patients was 75%. Both preoperative serum p53

288 Serum p53 Protein and Antibody of Head and Neck Carcinoma HEAD & NECK April 2001

protein and p53 antibodies were not significantlyassociated with tumor-related death and survival(all with p > .05).

The postoperative sera were available forstudy in 16 patients, 13 were taken at a timewithout clinical evidence of recurrence, and an-other 3 were taken at the time of tumor recur-rence. The results of serial analysis of serum p53protein and antibody levels in relation to tumorrecurrence are shown in Table 1.

DISCUSSION

The prognostic significance of serum p53 proteinlevel has not been reported in HNSCC patients.In this study, the preoperative serum p53 protein

was detectable in 91% patients. Serum p53 pro-tein was more commonly found in head and neckcancer patients, and the concentrations were at asignificantly higher level than normal control.Similar findings of higher serum p53 protein levelcompared with normal control had also been re-ported in lung and pancreatic carcinomas.9–10

The higher level of serum p53 protein in patientswith head and neck cancer is compatible with theknown high incidence of p53 mutation and over-expression in these patients. The p53 protein isalso expressed in normal cells; although withshorter half-life, it is not unexpected to be able todetect a small quantity of p53 protein in the se-rum of normal individuals even though at lower

Table 1. Serial analysis of serum p53 protein and P53-Ab.

Patient

Postoperativeblood takentime (mo)

Serum p53 Ab(+/−)

Serum P53 protein(pg/mL)

RecurrencePreop Postop Preop Postop

1 5 + + 20.65 18.87 No recurrence2 4 + + 84.71 32.84 No recurrence3 3 + + 0 0 No recurrence4 20 + + 21.83 0 No recurrence5 7 + − 24.2 0 No recurrence6 3 + − 0 5.23 No recurrence7 33 + − 59.14 0 No recurrence8 3 − − 20.94 51.45 No recurrence9 6 − − 0 — No recurrence

10 6 − − 45.12 21.83 No recurrence11 3 + + 11.28 3.22 Nodal recurrence at 7 months12 3 + − 0 0 Distal metastasis at 9 months13 3 − − 7.758 16.83 Nodal recurrence at 7 months14 6* + − 38.52 801.52 Nodal recurrence at 6 months15 9* − − 76.82 0 Local recurrence at 9 months16 5* − + 68.29 0 Local and distant metastasis at 5 months

*Blood was taken at time of recurrence for the last three patients.

Table 2. Literature review of prognostic significance of serum p53 antibody in HNSCC.

AuthorsPatient

no.Seropositive

p53 antibody (%)Significantcorrelation

Insignificantcorrelation

Bourhis J16 80 15 (19) Nodal metastasisRecurrenceSurvival

—

Lavielle JP17 74 32 (44) Chemosensitivity Site of tumorWollenberg B18 97 37 (38) — Stage, differentiationMaass JD19 82 18 (22) — Stage, nodal metastasis, differentiationWerner JA20 143 39 (27) Recurrence

Secondary primaryStage, age, sex

Ralhan R25 70 24 (34) DifferentiationStage

—

Present study 75 23 (31) Nodal metastasis Tumor site, sex, age, differentiation, stage, smoking, drinkingTotal 621 188 (30) Nodal metastasis,

RecurrenceTumor site, sex, age, stage, differentiation, smoking, drinking

Serum p53 Protein and Antibody of Head and Neck Carcinoma HEAD & NECK April 2001 289

incidence and lower level. Because serum p53protein can be found in small percentage of nor-mal controls, its presence is not diagnostic of can-cer. Although serum p53 protein is not diagnosticmarker of cancer, whether its serum titer can beused as a screening marker or predictive factorfor the development of cancer particularly inhigh-risk groups (eg, chronic smokers and drink-ers) is worth further prospective study.

The preoperative serum p53 protein level wasalso found to be unrelated to clinicopathologic pa-rameters including sex, age, differentiation, tu-mor stage, smoking, and drinking habits. Thepreoperative serum p53 protein level was alsofound to be not significant in the prediction ofnodal metastasis, recurrence, and survival.

The serum p53 protein level was not corre-lated with the immunohistochemical finding ofp53 expression in the primary tumor tissue. Thecellular cytosol p53 protein had been reported tobe correlated with the immunohistochemicalstaining intensity.13 The serum p53 protein levelmay theoretically reflect the total amount of sta-bilized p53 protein in the body. However, the totalamount of stabilized protein in the serum may notbe simply directly related to the total cytosol p53protein or its semiquantitative immunohisto-chemical staining level of expression in biopsy tis-sue of the carcinoma. The other factors, includingtumor volume, the total number of cancer cellswith overexpression of p53, the degree of overex-pression in individual cells, the degree of cellularnecrosis,14 the vascularity of tumor and the rateof absorption of p53 protein into the circulation,13

and the rate of degradation or absorption in se-rum, may theoretically all affect the final concen-tration of serum p53 protein. It is therefore notunexpected to find that there was no direct corre-lation between immunohistochemical stainingscore and serum protein level.

The preoperative serum p53 antibody wasfound in 31% patients with HNSCC. Because p53antibody can also be found in normal controls,although at lower incidence compared with can-cer in this study and in other studies, its presenceis not diagnostic of cancer.15 Similar to serum p53protein, we are not sure whether serial analysis ofserum p53 antibody can be used as a tumormarker for screening high-risk patients for thedevelopment of cancer. It has been reported thatserum p53 antibody can be found in the sera ofworkers occupationally exposed to vinyl chlo-ride.15 Further large scale prospective study isworthwhile to evaluate the value of serum p53

antibody for the screening of cancer in the popu-lation. In the literature, the reported incidence ofpositive serum p53 antibody ranged from 19% to44%, with an overall average of 30% as shown inTable 2.

The prognostic significance of serum p53 anti-body is controversial as shown in Table 2. Wefound that preoperative serum p53 antibody wassignificantly associated with nodal metastasis; asimilar finding was also reported by Bourhis etal.16 The association of the high incidence (65%) ofnodal metastasis with the presence of serum p53antibody may reflect the lymphocyte’s reaction tothe p53 protein in nodal metastatic cancer cells.16

Nodal recurrence is common after treatment ofclinically node-negative head and neck can-cer.21,22 It is still controversial whether prophy-lactic treatment with neck dissection or radio-therapy of the clinically node-negative neck maybe considered in the treatment of these patientsto reduce the nodal failure rate. We are stillsearching for the best possible panel of prognosticmarkers of HNSCC. Serum p53 antibody may beone of the useful prognostic markers in combina-tion with other panels of prognostic markers forthe management of clinically node-negative headand neck carcinoma. Further study is, however,necessary.

The postoperative serum p53 antibody hasbeen studied in hepatocellular carcinoma23 andovarian carcinoma.24 It was found to have no sig-nificant prognostic value for recurrence and sur-vival in both cancers. There is no report of theprognostic significance of postoperative serumlevel of p53 protein and antibody in HNSCC. Inthis study, a small number of patients were inagreement for the study in the postoperative pe-riod. Despite the small sample size, the resultsmight suggest that postoperative serial analysisof p53 protein and antibody were also not usefultumor markers for monitoring tumor recurrencecomparable with findings of hepatocellular carci-noma and ovarian carcinoma. Larger scale studyis, however, necessary to verify our preliminaryresults.

In conclusion, preoperative serum p53 anti-body was a significant prognostic factor of nodalmetastasis HNSCC.

REFERENCE1. Vogelstein B, Kinzler KW. p53 Function and dysfunction.

Cell 1992;70:523–526.2. Hollstein M, Sidransky D, Vogelstein B, Harris CC. p53

mutation in human cancers. Science 1991;253:49–53.

290 Serum p53 Protein and Antibody of Head and Neck Carcinoma HEAD & NECK April 2001

3. Ahomadegbe JC, Barrios M, Fogel S, et al. High incidenceof p53 alterations (mutation, deletion, over-expression) inhead and neck primary tumors and metastases; absenceof correlation with clinical outcome. Frequent proteinoverexpression in normal epithelium and in early non-invasive lesions. Oncogene 1995;10:1217–1227.

4. Somers KD, Merrick MA, Lopez ME, et al. Frequent p53mutations in head and neck cancer. Cancer Res 1992;52:5997–6000.

5. Maestro R, Dolcetti R, Gasparotto D, et al. High frequencyof p53 gene alterations associated with protein over-expression in human squamous cell carcinoma of larynx.Oncogene 1992;7:1159–1166.

6. Frank JL, Bur ME, Garb JL, et al. P53 tumor suppressoroncogene expression in squamous cell carcinoma of hypo-pharynx. Cancer 1994;73:181–186.

7. Leedy DA, Trune DR, Kronz JD, et al. Tumor angiogene-sis, the p53 antigen, and cervical metastasis in squamouscell carcinoma of the tongue. Otolaryngol Head Neck Surg1994;111:417–422.

8. Luo JC, Neugut AI, Garbowski G, et al. Levels of p53antigen in the plasma of patients with adenomas and car-cinomas of the colon. Cancer Lett 1995;91:235–240.

9. Luo JC, Zehab R, Anttila S, et al. Detection of serum p53protein in lung cancer patients. J Occup Med 1994;36(2):155–160.

10. Suwa H, Ohshio G, Okada N, et al. Clinical significance ofserum p53 antigen in patients with pancreatic carcinoma.Gut 1997;40:647–653.

11. Cough MJ, Koch M, Brenna JA, Sidransky D. The hu-moral response to oncoprotein p53 in head and neck can-cer. Head Neck. 1994;16:495.

12. Pindborg JJ, Reichart PA, Smith CJ, et al. World HealthOrganization international histological classification oftumors: Histological typing of cancer and precancer of theoral mucosa. International Histological Classification.Berlin: Springer; 1971. pp 11–13.

13. Vojtesek B, Fisher CJ, Barnes DM, Lane DP. Comparisonbetween p53 staining in tissue sections and p53 proteinslevel measured by an ELISA technique. Br J Cancer 1992;67:1254–1258.

14. Labreque S, Naor N, Thomson D, Matlashewski G. Analy-sis of the Anti-p53 response in cancer patients. CancerRes 1993;53:3468–3471.

15. Trivers GE, Cawley HL, Virna MG, et al. Anti-p53 anti-bodies in sera of workers occupationally exposed to vinylchloride. J Natl Cancer Inst 1995;87:1400–1407.

16. Bourhis J, Lubin R, Roche B, et al. Analysis of p53 serumantibodies in patients with head and neck squamous cellcarcinoma. J Natl Cancer Inst 1996;88:1228–1233.

17. Lavieille JP, Righini C, Reyt E, et al. Implications of p53alternations and anti-p53 antibody response in head andneck squamous cell carcinomas. Oral Oncol 1998;34:84–92.

18. Wollenberg B, Jan NV, Pitzke P, et al. Anti-p53 antibodiesin serum of smokers and head and neck cancer patients.Anticancer Res 1997;17:413–418.

19. Maass JD, Gottschlich S, Goeroegh T, et al. Head andneck cancer and p53 immunogenicity. Anticancer Res1997;17:2873–2874.

20. Werner JA, Gottschlich S, Folz BJ, et al. p53 serum anti-bodies as prognostic indicator in head and neck cancer.Cancer Immunol Immunother 1997;14:112–116.

21. Yuen APW, Lam KY, Chan ACL, et al. Clinicopathologicalanalysis of elective neck dissection for N0 neck of earlyoral tongue carcinoma. Am J Surg 1999;177:90–92.

22. Yuen APW, Wei WI, Wong YM, Kwong CT. Elective neckdissection versus observation in the treatment of earlyoral tongue carcinoma. Head Neck 1997;19:583–588.

23. Saffroy R, Lelong JC, Azoulay D, et al. Clinical signifi-cance of circulating anti-p53 antibodies in European pa-tients with hepatocellular carcinoma. Br J Cancer 1999;79:604–610.

24. Gadducci A, Ferdehini M, Buttitta F, et al. Serum anti-p53 antibodies in the follow-up of patients with advancedovarian carcinoma. Anticancer Res 1998;18:3763–3765.

25. Ralhan R, Nath N, Agarwal S, et al. Circulating p53 an-tibodies as early markers of oral cancer: correlation withp53 alternations. Clinical Cancer Res 1998;4:2147–2152.

Serum p53 Protein and Antibody of Head and Neck Carcinoma HEAD & NECK April 2001 291