Downregulation of TESTIN and Its Association With Cancer History and a Tendency Toward Poor Survival in Head and Neck Squamous Cell Carcinoma FREE

Head and neck squamous cell carcinoma (HNSCC) is one of the most frequent cancers that lead to death, making it a major health problem in the world. It includes oral, oropharyngeal, nasopharyngeal, hypopharyngeal, and laryngeal cancers and accounts for more than 644 000 new cases worldwide each year, with a mortality rate of 53% and a male predominance of 3:1.¹ Despite advanced technology in its detection and treatment, HNSCC continues to pose a great threat to human life. Recent advances in the technology and molecular biology of human cancer, including head and neck carcinoma, have provided possible novel diagnostic and prognostic markers.

Tumor suppressor genes (TSGs) are defined as genetic elements whose loss or mutational inactivation allows cells to display 1 or more phenotypes of neoplastic growth. Loss of heterozygosity (LOH) is considered an indication of a TSG presence, whereby inactivation contributes to the development and/or progression of tumor.² Detailed LOH analysis of polymorphic loci distributed along a chromosome can reveal a common minimal deleted region where putative TSGs may reside.

Previous cytogenetic studies as well as microdeletion analysis have shown frequent abnormalities of chromosome 7 in various cancer types, including HNSCC.^3- 10 Moreover, microcell-mediated transfer of human chromosome 7 into a murine squamous cell carcinoma cell line was found to inhibit tumorigenicity of the cell line.¹¹ In another study, introduction of a single copy of human chromosome 7 into a highly aggressive human prostate carcinoma cell line increased tumor latency by at least 2-fold.¹² Similarly, insertion of an intact human chromosome 7 into an immortalized human fibroblast cell line with LOH in the 7q31-32 region suppressed immortality of the cells and restored their senescent ability.¹³ All these studies have provided strong evidence for the existence of a TSG(s) in chromosome 7q31 region.

Based on these studies, we previously examined the chromosome region 7q22-31 using a set of highly polymorphic microsatellite markers to evaluate allele loss ratios and candidate TSGs in HNSCC.¹⁴ The study indicated 2 differentially deleted chromosomal areas at 7q31 around the markers D7S486 and D7S643. The presence of microdeletion in both loci but retention of the chromosomal region between these 2 peaks in some cases suggested that chromosome 7q31 included 2 different TSGs, at least in HNSCC. We have already shown that ING3 located at one of these regions (D7S643) has been involved as a TSG in HNSCC.¹⁴ However, the role of the other possible candidate gene in HNSCC remained unknown. Therefore, in the current study, we focused on the other highly deleted chromosomal area around the marker D7S486, which is located in intron 6 of the TESTIN gene, confirming its high ratio of deletion. In fact, missense mutations and decreased expression of TESTIN were detected in various cancer cell lines, including breast cancer, pancreatic cancer, and hematologic malignant neoplasms.¹⁵ Moreover, adenoviral transfection of TESTIN into breast and uterine cancer cell lines reduced tumor growth in mice.¹⁶ A recent study also demonstrated tumor suppressor function of TESTIN in a knockout mouse model.¹⁷ Thus, we analyzed the expression level and mutation status of TESTIN to clarify its role in HNSCC. We also compared the messenger RNA (mRNA) status of TESTIN with clinicopathologic variables.

Paired normal and tumor samples were obtained from 38 patients with primary HNSCC between 1999 and 2005 at the Department of Otolaryngology–Head and Neck Surgery, Okayama University Hospital, Okayama, Japan, after informed consent was obtained from each patient. Normal control samples were obtained from grossly normal tissue as far as possible from the tumor tissue. Although surgical margins were examined during surgery, we also confirmed the histopathologic appearance of the normal tissues by hematoxylin-eosin staining. All tissues were frozen in liquid nitrogen immediately after surgery and stored at −80°C until the extraction of RNA. The study included 30 men and 8 women (mean age, 64 years; age range, 40-81 years). The histologic diagnosis in all cases was squamous cell carcinoma. None of the patients received preoperative chemotherapy or radiotherapy. The differentiation and diagnosis of the tumor was based on the surgical pathology reports of the hospital. All clinical information was obtained from the patient files, which included the initial diagnosis, treatment, and follow-up data. The bioethics committee of the institution approved the study. The clinicopathologic characteristics of the patients were shown in Table 1.

Total RNA samples were prepared using a modified acid guanidinium phenol chloroform method (ISOGEN; Nippon Gene Co Ltd, Tokyo, Japan). Total RNA was reverse-transcribed with a preamplification system (ReverTra Ace Kit; Toyobo Co Ltd, Osaka, Japan) starting with 2 μg of total RNA from each sample, according to the procedures provided by the supplier. TESTIN (GenBank NM_015641) mRNA expression in paired tumor and normal tissues was examined by duplex reverse transcriptase–polymerase chain reaction (RT-PCR). One microliter of each RT reaction was amplified in 50 μL of mixture containing 1.2mM magnesium chloride, 1X PCR buffer, 200 μmol/L of each deoxynucleotide triphosphate, 20 pmol of each primer, and 1 U of recombinant Thermus thermophilus (rTth) DNA polymerase XL (Applied Biosystems, Foster City, California). Thirty PCR cycles of TESTIN primers, RT-S (5′- GAA TGA GAA GCT ATA CTG TGG C) and RT-AS (5′- ATG GCT CGA TAC TTC TGG GTG), and 25 cycles of β-actin primers, S1 (5′-GGC CAA CCG CGA GAA GAT GAC) and AS1 (5′-GCT CGT AGC TCT TCT CCA GGG), were used for amplification (the primers were designed with GENETYX-MAC 10.1; Software Development Co Ltd, Tokyo, Japan). An initial denaturation step at 94°C for 3 minutes was followed by 30 cycles of a denaturation step at 94°C for 30 seconds, an annealing step at 60°C for 1 minute, and an extension step at 72°C for 1 minute. A final extension step at 72°C for 7 minutes was added. β-Actin primers were added to each PCR tube at the end of the fifth cycle by holding the thermocycler at 94°C for awhile. Reproducibility was confirmed by processing all samples 2 times.

The PCR products were separated through 2% agarose gel and stained with ethidium bromide. As the sizes of the PCR products were 386 base pair (bp) for β-actin and 446 bp for TESTIN, they were readily distinguishable. The intensity of ethidium bromide staining of each band was measured by a CCD (charged-couple device) image sensor (Gel Print 2000/VGA; Toyobo Co Ltd) and analyzed by a computer program for band quantification (Quantity One; Toyobo Co Ltd). The value of tumor-specific TESTIN expression was determined by calculating the ratio of the expression levels in the tumor and in the matched normal sample, each of which was normalized for the corresponding β-actin expression level (T, TESTIN/β-actin expression ratio in tumor sample; N, TESTIN/β-actin expression ratio in matched normal sample; and T/N ratio, the relative TESTIN expression in the tumor sample compared with its matched normal sample after normalization). Decreased and increased expression levels were classified as L and H when this ratio was less than 0.6 and greater than 1.4, respectively, as reported previously.¹⁴ Class N (normal expression) represented the value of the ratio between 0.6 and 1.4.

Each of the coding regions of exon 3 through 5 of the TESTIN gene was amplified by PCR with intron-spanning primers designed using GENETYX-MAC 10.1 software (Software Development Co): for exon 3, EX3-S (5′- CGT GTT TTG TTT CCT TCT TGC) and EX3-AS (5′- AGT AAT GAG AAC CCC GGA AG); for exon 4, EX4-S (5′- ATT GGC CTC CTT GTG GTT CAG) and EX4-AS (5′- CAA ACA CGA TGA CCC TCT GTG); and for exon 5, EX5-S (5′- GAC TAG GTT GTT CTG GAT GGC TT) and EX5-AS (5′- TGT CAA CCC AAT TAA CAC AGA CAG). The PCR mixture contained 100 ng of genomic DNA, 1.2mM magnesium chloride, 1X PCR buffer, 200 mmol/L of each deoxynucleotide triphosphate, 20 pmol/L of each primer, and 1 U of rTth DNA polymerase XL (Applied Biosystems) in a 50-μL volume. Initial denaturation at 94°C for 3 minutes was followed by 30 cycles of a denaturation step at 94°C for 30 seconds, an annealing step at 56°C (exon 3) or 60°C (exons 4 and 5) for 1 minute, and an extension step at 72°C for 1 minute. A final extension step at 72°C for 7 minutes was added. The resultant PCR products were purified using ExoSAP-IT reagent (USB Corp, Cleveland, Ohio) before sequence-specific PCR amplification. Purified PCR products were reamplified with a cycle sequencing kit (BigDye Terminator v1.1 Cycle Sequencing Kit; Applied Biosystems); then, they were ethanol precipitated and direct sequenced on an automated capillary sequencer using the primers described above (ABI 3130xl; Applied Biosystems).

Pearson χ², Fisher exact, and t tests were used to evaluate the correlation between TESTIN mRNA expression and the clinicopathologic characteristics of the patients. Survival curves were calculated according to Kaplan-Meier. The log-rank test was used to compare survival between low and normal-high expression of TESTIN. Overall survival in months was calculated for 5 years starting from the day after surgery. The duration of disease-free survival was determined from the day after surgery to the initial recurrence of the surgically resected cancer, evaluated by clinical examination. All statistical analyses were performed with SPSS version 10 (SPSS Inc, Chicago, Illinois), and P<.05 was considered significant.

The LOH at the chromosome 7q31 region has been reported for various cancers, including head and neck carcinoma. We recently redefined the map of the 7q31 region according to the contiguous sequences and detected 2 preferentially deleted areas around the markers D7S486 and D7S643.¹⁴ In that study, we documented ING3 as a TSG at one of these frequently deleted loci, D7S643. The other marker, D7S486, is located in intron 6 of the TESTIN gene. Figure 1 shows their relative locations at the chromosome 7q31 region.

Because our previous data demonstrated frequent allelic loss at the TESTIN gene area (D7S486), we examined the mRNA expression status of TESTIN to see whether it was inactivated with other mechanisms in HNSCC. In 38 available matched RNA samples, the expression levels of TESTIN mRNA were compared by semiquantitative RT-PCR using β-actin mRNA as a control. Seventeen of the 38 tumor tissue samples (45.0%) showed decreased expression of TESTIN mRNA compared with each paired normal tissue sample, while 12 samples (31.0%) showed a similar level of expression in normal and tumor tissue samples (Figure 2 and Table 2). Increased expression of TESTIN was detected in 9 tumor tissue samples (24.0%).

A previous study demonstrated missense mutations in exons 3, 4, and 5 of TESTIN in 3 cell lines, including CEM (leukemic cell line), MDAMB453 (breast cancer cell line), and CAOV3 (ovarian cancer cell line).¹⁵ Based on this information, we also examined the mutation status of exons 3 through 5 of TESTIN in our 38 head and neck cancer samples. The PCR amplification and subsequent direct sequencing of each sample demonstrated no substitution in exons 3 and 5. However, we detected the same nucleotide change (codon 221 of TESTIN transcript 1 in exon 4, from GCC to GTC, with an amino acid substitution from alanine to valine) as reported in the MDAMB453 breast cancer cell line.¹⁵ Six (cases 5, 6, 13, 20, and 24 [heteromutant] and case 7 [homomutant]) of 38 samples (16.0%) showed this nucleotide and amino acid change in our samples (Figure 3). Five of the 6 nucleotide changes were heterozygous, while only 1 sample displayed a homozygous substitution. To clarify whether this change was a simple polymorphism or a tumor-specific mutation, we analyzed corresponding normal DNA by direct sequencing. Our analysis detected the same substitutions in the normal tissues, suggesting that the nucleotide and amino acid change in codon 221 of exon 4 was a polymorphism.

The relationship between the expression status of TESTIN and the clinicopathologic characteristics of the patients is shown in Table 2. We compared these results between the samples with and without decreased mRNA expression of TESTIN. Seventeen of the 38 tumor samples (45.0%) showed decreased TESTIN mRNA expression in tumor samples compared with their normal counterparts. The mean age of the patients with and without decreased TESTIN expression was similar (62.5 years vs 65.5 years, respectively). A significant difference was detected between TESTIN expression and sex. All tumor samples but 1 (94.1%) with low TESTIN expression were from male patients, while 14 of 21 samples (66.7%) with normal or high TESTIN expression were from male patients (Table 3 [P < .05]). With regard to alcohol consumption and smoking, there was no significant difference between the decreased TESTIN expression group and the normal-high TESTIN expression group. However, the cases with decreased TESTIN expression were associated with a higher rate of smoking compared with the cases with normal-high TESTIN expression (70.6% vs 42.1%, respectively; Table 3). When the expression status of TESTIN was compared with the T stage, N stage, overall (TNM) stage, and histologic differentiation, no significant alterations were detected (Table 3).

An important and potentially significant finding was revealed when the expression status of TESTIN was compared with cancer history: 10 of 16 cases (62.5%) with low TESTIN mRNA expression had no cancer history, while all but 1 of the cases (94.7%) with normal-high TESTIN expression had no cancer history (P < .03, Table 3). In other words, 6 of the 16 cases (37.5%) with decreased TESTIN expression were associated with a cancer history, while only 1 of the 19 cases (5.3%) with normal-high TESTIN expression had a cancer history. One case (No. 12) with high TESTIN mRNA expression had a history of mucoepidermoid carcinoma, while 6 cases (Nos. 3, 4, 6, 20, 29, and 34) with low TESTIN mRNA expression had a history of chondrosarcoma, gastric cancer, tongue cancer, non-Hodgkin lymphoma, tongue cancer, and colon carcinoma, respectively. On the other hand, although a tendency of a higher ratio of family cancer history was detected in the cases with decreased TESTIN expression, no statistically significant change was shown (Table 3).

When we examined the relationship between survival and the expression status of TESTIN, no statistically significant difference was detected in either overall or disease-free survival (Figure 4A and B). In terms of disease-free survival, the rates among all cases were similar, although the cases with low TESTIN expression had a lower ratio of survival (Figure 4A, P = .28). As shown in Figure 4A, the mean (SD) disease-free survival in the low expression group was 31 (7) months (95% confidence interval [CI], 19-44 months), while that in the normal-high expression group was 40 (5) months (95% CI, 30-50 months), with no significant difference between the groups (P = .28, log-rank test). However, interestingly, a shorter survival was shown in the patients with decreased TESTIN mRNA expression compared with the patients without such alteration in terms of overall survival, almost reaching statistical significance (Figure 4B, P = .09). As shown in Figure 4B, the overall survival in the low expression group was 42 (6) months (95% CI, 31-53 months) and that in the normal-high expression group was 52 (4) months (95% CI, 44-59 months), and the log-rank test showed that patients with low TESTIN expression had an almost significantly shorter overall survival than those with normal-high expression (P = .09). Overall, while the 5-year survival ratio of the cases with decreased TESTIN mRNA expression was approximately 50%, it was nearly 80% among the cases without decreased TESTIN mRNA expression (Figure 4B).

Our previous study demonstrated frequent allelic loss at the marker D7S486 location, which is an intragenic marker and confirms the deletion of the TESTIN gene.¹⁴ The appearance of 2 different peaks of allelic loss suggested that TESTIN is likely to be another TSG involved in HNSCC other than ING3, which was identified by our group as a TSG from the 7q31 area. In fact, an increasing number of studies support a tumor suppression role for TESTIN. The detection of tumor-specific mutations as well as decreased mRNA expression in human tumor samples and in vitro functional analysis all suggested that TESTIN is a TSG in various cancers.^15- 19

Knudson’s²⁰ definition of a classic TSG requires inactivation of both alleles of a candidate gene in tumors. Inactivation of these classic TSGs occurs through the deletion of one of its alleles and mutation in its other allele. We previously found that one of the TESTIN alleles (D7S486) was deleted in our cases.¹⁴ Therefore, in the present study, we examined the inactivation status of TESTIN's second allele in HNSCC. Mutation analysis revealed nucleotide and amino acid substitutions in exon 4 in a considerable number of the cancer cases. Existence of this possible mutation was previously reported in a breast cancer cell line.¹⁵ However, analysis of corresponding normal samples for the tumor cases with this substitution showed the same alteration, suggesting that it was not a tumor-specific somatic change but a polymorphism, which can still influence TESTIN protein function and be important during carcinogenesis. To clarify the importance of this polymorphic change, in vitro functional studies of the preparation of the expression vectors for each variant should be performed.

Recently, a new class of TSG with haploid insufficiency, in which one of alleles is lost and the other allele is haploinsufficient, has been described. These hemizygous TSGs showed a tumor-prone phenotype when challenged with carcinogens.^21- 22 Moreover, promoter methylation has also been attributed to the inactivation mechanism of TSG.²³ Whether haploinsufficiency or promoter methylation occurs, the result is a decrease in mRNA expression and then possibly a decrease in the protein product and deficient function of TSG, resulting in cancer development. To identify the possibility of such a result, we analyzed the mRNA expression status of TESTIN in matched normal-tumor cases and showed downregulation of TESTIN in approximately 50% the head and cancer cases compared with their corresponding normal controls. This finding suggested that inactivation of TESTIN as a class 2 TSG is likely to occur, especially in head and neck cancer.

Comparison of the patients' clinicopathologic variables with their TESTIN mRNA expression level demonstrated interesting findings, the most remarkable of which was a significant association with a history of cancer. Clearly, the patients with a low TESTIN expression level had a higher rate of cancer history. Although it was not statistically significant (P = .43), a higher rate of smoking was also shown in the patients with low TESTIN expression level. Therefore, a higher rate of cancer history could be related with early smoking in those cases. However, a tendency of a higher rate of family cancer history was also shown in the patients with low TESTIN expression, although it was far from reaching statistical significance. Nevertheless, the higher ratio of a personal as well as a family cancer history and the finding of a polymorphism with unknown significance in a considerable number of the patients suggest that TESTIN inactivation or a decrease in its activity could be strongly related to the development of head and neck cancer. On the other hand, survival analysis demonstrated a possible prognostic value of TESTIN in HNSCC. An almost significant difference was revealed between low and normal-high TESTIN expression in terms of overall survival. Although TESTIN expression alone is not enough evidence for a strong prognostic association, evaluation of the expression status of other TSGs in the human genome together with TESTIN expression may provide important information, which could be used as a molecular prognostic factor.

In conclusion, downregulation of TESTIN mRNA suggests that it may function as a tumor suppressor in HNSCC and that its inactivation leads to cancer development. Further studies in various human cancer tissues using a larger sample size and in vitro functional studies as well as clinical comparison research studies would give us a better evaluation of TESTIN's role along with its possible future applications in molecular diagnosis and treatment of different cancer types, including HNSCC.

Correspondence: Mehmet Gunduz, MD, PhD, Department of Oral Pathology and Medicine, Graduate School of Medicine, Dentistry, and Pharmaceutical Sciences, Okayama University, 2-5-1 Shikatacho, Okayama 700-8558, Japan (mgunduz@md.okayama-u.ac.jp).

Submitted for Publication: March 19, 2008; final revision received May 20, 2008; accepted July 6, 2008.

Author Contributions: Dr M. Gunduz had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: M. Gunduz, Shimizu, and Nagai. Acquisition of data: E. Gunduz and Beder. Analysis and interpretation of data: E. Gunduz, M. Gunduz, Nagatsuka, Fukushima, Sutcu, Delibas, Yamanaka, and Shimizu. Drafting of the manuscript: E. Gunduz. Critical revision of the manuscript for important intellectual content: M. Gunduz, Beder, Nagatsuka, Fukushima, Sutcu, Delibas, Yamanaka, Shimizu, and Nagai. Statistical analysis: Beder. Obtained funding: M. Gunduz. Administrative, technical, and material support: M. Gunduz, Nagatsuka, Sutcu, Delibas, Yamanaka, Shimizu, and Nagai. Study supervision: M. Gunduz and Fukushima.

Financial Disclosure: None reported.

Funding/Support: This work was supported in part by grants-in-aid for scientific research from the Ministry of Education, Culture, Sports, Science, and Technology (18-06262 [Dr E. Gunduz], 19592109 [Dr Nagatsuka], and 17406027 [Dr Nagai]); by “innovative seed” research from Japan Science and Technology Agency (Dr M. Gunduz); by Sumitomo Trust Haraguchi Memorial Cancer Research Promotion (Dr M. Gunduz); and by a research grant from AstraZeneca (Dr M. Gunduz).

Additional Contributions: Surgeons from the Department of Otolaryngology–Head and Neck Surgery, Okayama University, generously provided fresh samples for this experiment.