Expression of nm23 antimetastatic gene product in head and neck squamous cell carcinoma

Abstract

The nm23 gene has been implicated as a suppressor gene involved in the control of the metastatic process of malignant cells. Reduced levels of nm23 gene product have been found in tumor cells with high metastatic potential such as several types of rodent tumors and human breast, colorectal, and lung carcinoma. This pilot study examines the expression of the nm23 gene product compared with nodal status in 70 consecutive patients with squamous cell carcinomas of the head and neck. Immunohistochemical staining was carried out on these tumor tissues with a monoclonal antibody to nm23-H1 peptide sequence. The tissues were scored from 0 to 2 by 2 independent observers. Reduced immunoreactivity, grades 0 and 1, was observed in 30 patients with positive nodal status (N = 34, 88%). Strong immunoreactivity, grade 2, was observed in 20 patients with negative nodal status (N = 36, 56%). Reduced expression of nm23 gene product was observed in patients with positive lymph node metastasis (P = 0.0006, χ 2 ). However, no significant differences in survival of these groups based on nm23 expression could be shown with the Kaplan-Meier analysis. This initial finding in the difference of nm23 gene product expression in patients with differing lymph node status is exciting but must be further validated with future studies. (Otolaryngol Head Neck Surg 2000;122:96–9.)

Molecular oncology, once believed to offer little benefit to the practicing physician, is now affecting and progressively altering everyday clinical practice. Cancer is caused by somatic mutations and does not arise from a single mutation but from an accumulation of several mutations. 1 A number of tumor-suppressor genes have been isolated and characterized to date, 2 , 3 with nm23 gene as a potential suppressor gene. The nm23 gene was originally identified by Steeg et al 4 in a rodent melanoma cell line. It has been shown to possess antimetastatic properties, and data suggest that normal expression of nm23 contributes to normal cell growth, whereas an altered or decreased level of nm23 expression leads to cell metastasis. The exact mechanism underlying the supposed nm23 antimetastatic suppression is not known. There has been considerable interest in this potential antimetastatic gene, and it is being studied in a variety of cancers. The most promising data to suggest its antimetastatic properties has been in adenocarcinoma of the breast. 5 – 5

The nm23 gene has been studied in laryngeal squamous cell carcinoma (SCC) with some suggestion that expression is altered compared with that in nonneoplastic laryngeal polyps. 8 This study looks at the nm23 gene expression in human SCCs of the head and neck region and compares the differences in its expression among patients with differing lymph node status. If proved to be reliable in predicting lymph node status, nm23 could aid in treatment planning and direct follow-up surveillance.

METHODS AND MATERIAL

Archival tissues of 70 consecutive patients treated for head and neck SCC between 1987 and 1993 at Madigan and Fitzsimmons Army Medical Centers were studied. nm23-H1 detection was accomplished by use of a version of an ABC technique. A mouse monoclonal nm23 antibody was applied, followed by a biotinylated antimouse IgG, and then by avidin/biotin/horseradish peroxidase complex. After incubation with an H₂O₂/diaminobenzidine substrate, an insoluble colored precipitate was produced where nm23 was found. The primary antibody and detection kit (UniTect) were from Oncogene Science Inc.

Paraffin-embedded tissue sections were mounted on positively charged slides. Sections were rehydrated with successive 10-minute changes of ClearRite-3 ×3, 100% dehydrant ×2, and H₂O ×2. Digestion was done with protease type XIV from Streptomyces griseus following the kit pronase procedure. The primary antibody was diluted to 5 μg/mL and incubated on the slides for 60 minutes. After the primary antibody, the sections were treated with a peroxidase suppressor for 30 minutes. An additional blocking step with 2% bovine serum albumin/phosphate-buffered saline solution was used after the peroxidase suppression. Dehydration involved the same reagents used to rehydrate.

Fig 1.

Normal skin epithelium.

Fig 2.

Grade 0: Very light to no staining for nm23 gene product.

The immunohistochemistry staining of the tissue sections was assessed according to the intensity and proportion of the cells. The tissue sections were compared with the positive and negative control sections run with each batch. The positive control used was adenocarcinoma of the breast proved positive to stain for nm23. The negative control was skin epithelium, which was derived from normal population groups (Fig 1). Variability among the normal population was minimal, with all scoring a grade 0. Therefore an attempt to stain the skin epithelium from each tumor patient was not pursued. An internal control to assess the integrity of the archival tissues was examined by staining for vimentin. The sections were scanned at low (×20) power to assess the heterogeneity of distribution. Multiple fields containing tumor were then examined at medium (×200) power for the purpose of grading. The tissues were graded 0 for very light to no staining, 1 for widespread faint or blush staining, and 2 for widespread strong or granular staining (Figs 2–4). The 3 grades had been used by previous immunohistochemistry studies of nm23 and therefore were used in our study for continuity. Each case was reviewed and graded by 2 independent blinded graders.

Fig 3.

Grade 1: Widespread faint to blush staining for nm23 gene product.

Fig 4.

Grade 2: Widespread strong or granular staining. Compare with grades 0 and 1 (Figs 2 and 3).

RESULTS

nm23-H1 expression was analyzed in 70 cases of SCC of the head and neck. There were 40 laryngeal, 19 tongue, 4 floor of the mouth, 4 pharynx, and 3 tonsil SCC cases. No attempts were made to stain the actual nodes of the tumor cases. Only the actual tumor site was histologically examined. Twelve cases showed weak to no staining (grade 0), 34 showed light staining (grade 1), and 24 showed heavy staining (grade 2). nm23-H1 expression was significantly reduced (grades 0 and 1) in SCC cases with lymph node metastasis (P = 0.0006, χ 2 contingency test; Table 1). Fifteen of 40 laryngeal cases had metastatic lymph nodes. Of these 15 metastatic lymph node cases, 12 showed reduced immunoreactivity (grades 0 and 1), whereas 15 of 24 nonmetastatic cases showed strong immunoreactivity (grade 2; P = 0.02). There were 13 metastatic lymph node tongue carcinoma cases, 12 of which showed reduced immunoreactivity (P = 0.01). There were 4 cases each of floor of the mouth and pharynx carcinoma and 3 cases of tonsil carcinoma. No statistically significant conclusion could be derived from these last 2 small groups (floor of the mouth/pharynx and tonsil). However, 3 of 4 lymph node–positive floor-of-mouth cases showed decreased immunoreactivity, whereas all 4 of the pharyngeal lymph node–positive cases had decreased immunoreactivity. All 3 cases of the tonsil carcinoma cases showed decreased immunoreactivity regardless of lymph node status.

Fig 5.

Patient outcome related to immunohistochemistry result.

Table 1.

Summary of immunoreactivity

Location	Grade 0	Grade 1	Grade 2
Larynx
– Nodes	2	7	16
+ Nodes	4	8	3
Tongue
– Nodes	2	1	3
+ Nodes	1	11	1
Floor of mouth
– Nodes	0	3	1
+ Nodes	0	0	0
Pharynx
– Nodes	0	0	0
+ Nodes	2	2	0
Tonsil
– Nodes	0	1	0
+ Nodes	1	1	0

– Nodes, Negative nodal status; + Nodes, positive nodal status.

Table 2.

Immunoreactivity with patient outcome

Status	Grade 0	Grade 1	Grade 2
NED	5	23	16
DWD	7	7	4
LT F	0	4	4

NED, No evidence of disease; DWD, dead with disease; LTF, lost to follow-up.

The mean follow-up of the disease-free group was 40 months. The mean survival for the subjects who died was 30 months. Of the 18 dead subjects, 7 were in the grade 0, 7 in the grade 1, and 4 in the grade 2 groups (Table 2). The mean mortality rate and survival length for those who died of their disease in each group were as follows—grade 0: 50%, 22.2 months; grade 1: 21%, 37.7 months; and grade 2: 19%, 16.5 months.

Kaplan-Meier statistical analysis was used to estimate the survival of these groups, but no apparent differences were statistically significant (P = 0.77). We also looked at a multivariate model using SPSS regression to determine whether nm23 gene product expression is independent of other variables. These variables were the nodal status, total number of nodes, tumor size, follow-up months, and TNM staging. None of these variables was able to predict the nm23 gene product score. Therefore nm23 gene product expression appears to be independent of these other prognostic factors (P = 0.0527, SPSS regression).

DISCUSSION

Survival after treatment for SCC of the head and neck has improved marginally during the last 20 years. 9 Thirty percent to 50% of these patients have a local or regional recurrence after surgical resection, 20% to 30% generate distant metastasis, and 10% to 40% produce a new second primary SCC. 10 The nm23 gene has been implicated as a suppressor gene observed in the metastatic process of malignant cells. Levels of nm23 have been shown to be reduced in tumor cells of high metastatic potential in several types of rodent tumors and human breast, colorectal, and lung carcinomas. The nm23 gene was first isolated on the basis of its enhanced expression in the nonmetastatic murine melanoma cell line compared with its low-level expression in a highly metastatic matched cell line. 4 In human breast carcinoma, low nm23 mRNA levels were found in patients with lymph node metastasis. 5 , 6 These studies suggested that nm23 may have a negative influence on the process of tumor metastasis. 7 Other varieties of carcinomas have shown a correlation between nm23 expression and tumor recurrence and metastasis. Overexpression of nm23 mRNA in small solitary hepatocellular carcinomas has been associated with a lower recurrence rate after surgical resection. 11 , 12 The nm23 gene has also been theorized to protect against death from colorectal cancer, and this protective effect has been found to increase with each increment of nm23 positivity. 12 Recently, nm23 gene expression has been evaluated in laryngeal carcinoma and compared with expression in laryngeal polyps. There was reduced expression of the nm23-H1 gene in the SCC cases versus the nonneoplastic laryngeal polyp cases. 8 Although the number of cases of laryngeal carcinoma in this previous study was too small to establish statistical significance, comparison of these patients showed no difference in the nm23-H1 expression between those with positive and those with negative lymph node metastasis. 8

This pilot study of 70 head and neck SCC cases shows a statistically significant decrease in expression of nm23-H1 (P = 0.0006) in those patients with known lymph node metastasis. Expression of nm23-H1 is inversely proportional to lymph node status. Patient survival appears to be decreased in groups with grades 0 and 1 that exhibit decreased expression of the nm23 gene, although this was not statistically significant (P = 0.77, Kaplan-Meier). Interestingly, group 2, which had the strongest expression of nm23 and the lowest mortality rate (19%), exhibited a mean survival rate of 16.5 months (Fig 5). This could be attributed to the patient's comorbid factors such as cardiovascular and pulmonary disease processes.

Immunohistochemistry is only a qualitative way to assess the expression of this particular gene. Because this pilot study demonstrates that there is indeed a difference in the qualitative expression of the proposed antimetastatic nm23 gene, there is exciting data to suggest its possible role in SCC of the head and neck and lymph node metastasis. This initial finding must be further validated by quantitative analysis to reach a full conclusion of the role of nm23 in the metastasis of head and neck SCC. The next study should also include examination of nm23 gene expression in the nodes themselves. Our involvement and understanding of molecular oncology can only contribute to new diagnostic methods and treatment plans.

We thank the Department of Clinical Investigations at Madigan Army Medical Center for their support and Troy H. Patience and Elizabeth Pulos, medical statisticians, for their assistance.

References

Hall

Herring

Hall

Molecular oncology and the surgeon. Am Surg 1995;61:156–60.

Marshall

Tumor suppressor genes. Cell 1991;64:313–26.

Weinberg

Tumor suppressor genes. Science 1991;254:1138–46.

Steeg

Bevilacqua

Kopper

Evidence for a novel gene associated with low tumor metastatic potential. J Natl Cancer Inst 1988;80:200–4.

Bevilacqua

Sobel

Liotta

Association of low nm23 RNA levels in human infiltrating ductal breast cancer with lymph node involvement and other histopathologic indicators of high metastatic potential. Cancer Res 1989;49:5185–90.

Hennessy

Henry

May

FEB

Expression of the anti-metastatic gene nm23 in human breast cancer: an association with good prognosis. J Natl Cancer Inst 1991;83:281–5.

Hirayama

Sawai

Takagi

Positive relationship between expression of anti-metastatic factor (nm23 gene product of nucleoside diphosphate kinase) and good prognosis in human breast cancer. J Natl Cancer Inst 1991;83:1249–50.

Lee

Redshaw

Boag

nm23-H1 protein immunoreactivity in laryngeal carcinoma. Cancer 1996;77:2246–50.

Snow

Clark

Andersen

Multimodality therapy for head and neck cancer. Stuttgart: Georg Thieme Verlag; 1992. p. 28–36.

10.

Hong

Lippman

Itri

Prevention of second primary tumors with isotretinoin in squamous cell carcinoma of the head and neck. N Engl J Med 1990;323:795–801.

11.

Boix

Bruix

Campo

nm23-H1 expression and disease recurrence after surgical resection of small hepatocellular carcinoma. Gastroenterology 1994;107:486–91.

12.

Yamaguchi

Urano

Goi

Expression of human nm23-H1 and nm23-H2 proteins in hepatocellular carcinoma. Cancer 1994;73:2280–4.