Serum messenger RNA as a biomarker and its clinical usefulness in malignancies.

To detect a transcript of interest in cell-free serum, quantitative real-time RT-PCR was performed by using 1 μl of RNA extract and 2 μl of SYBR Green I (Roche, Basel, Switzerland) in a One Step RT-PCR kit (Qiagen, Tokyo, Japan) using LightCycler with reproducibility. After blood sampling, RNA was extracted with DNase treatment after three steps of centrifugation of serum as previously reported with a few modifications mainly including a precise gravity control of centrifugation to obtain cell-free serum and mRNA quantification using optimized primer set (INTEC Web and Genome Bioinformatics, Tokyo, Japan) ( 8 , 16 ). Other minor modifications are as follows. RNA from 200 μl of serum was dissolved in 200 μl of water. Quantitative RT-PCR was performed by using 1 μl of RNA extract and 2 μl of SYBR Green I (Roche, Basel, Switzerland) in a One Step RT-PCR kit (Qiagen, Tokyo, Japan). RNA in serum was extracted by using the same volume of serum and dried up to 20-fold concentration. RNAs from tissues were extracted using TRIzol Reagent according to manufacture’s instructions (Invitrogen Corp., Carlsbad, CA, U.S.A.). RT-PCR consisted of an initial incubation at 50 °C for 30 min followed by a 12-min incubation at 95 °C, then 50 cycles at 95 °C (0 s), 55 °C (10 s), and 72 °C (15 s), and a 20 second melting at 40 °C. The dynamic ranges of real-time PCR analysis for mRNA of interest were more than approximately 5 to 10 copies in this assay. We could therefore exclude the possibility of false negative results in serum samples from patients and controls. To examine significant clinicopathological findings affecting mRNA and other markers, the difference among diseases and stratified categories in each clinical parameter was statistically evaluated. Tumors, tumor size, number of tumors, histological findings including differentiation degree of tumors, clinical staging, and the presence of metastasis were analyzed. Correlations among biomarkers were calculated using Pearson’s relative test. To assess the accuracy of diagnostic tests, the matched data sets (patients with reference diseases or healthy individuals and those with malignancies) for bio-markers were analyzed using receiver operator characteristic (ROC) curve analysis. This assay demonstrated a strong linear relation between copy number and PCR cycles using RNA controls (r 2 > 0.99) ( Fig. 1A ). A correlation of mRNA between malignant tissue and serum was analyzed to examine whether serum mRNA is derived from the original tissues by both paired t test and Spearman’s test. Optimal predictive cut-off values and the sensitivity/specificity for mRNA expressions, and positive predictive value (PPV)/ negative predictive value (NPV) during carcinogenesis were calculated. All the primers were optimally designed (INTEC Web and Genome Informatics corp., Tokyo, Japan).

Since the discovery of circulating nucleic acids in plasma in 1948, many diagnostic applications have emerged. Small amounts of circulating nucleic acids (CNA) are present in the plasma of healthy individuals. Increased levels of plasma CNA have been reported in a number of clinical disorders such as cancer, stroke, trauma, myocardial infarction, autoimmune disorders and pregnancy-associated complications. CNA has received special attention because of its potential application as a non-invasive, rapid and sensitive tool for molecular diagnosis and monitoring of the diseases, and the prenatal diagnosis of fetal genetic diseases. A simple blood test for cancer detection has been the quest of many researchers in particular. Recently, CNA instead of a protein has been used in practical diagnosis. Cell-free circulating nucleic acids in plasma/serum derived from tumor tissues, have received much interest. Although it is well known that higher concentrations of DNA (deoxyribonucleic acid) are present in the plasma of cancer patients sharing some characteristics with DNA of tumor cells ( 1 , 2 , 3 , 4 ), it has been reported that mRNAs detected in blood reflects the early event in a clinical condition ( 5 ). Since RNA (ribonucleic acid) in plasma/serum may be a suitable source for the development of non-invasive diagnostic, prognostic and follow-up tests for cancer, this discovery has provided us with very promising assays useful for early detection of malignancies.

Ovarian cancer is the fifth most common cancer in women. Despite the fact that it is highly curable if diagnosed early, ovarian cancer kills more women each year than all other gynecologic malignancies ( 37 ). There are no proven methods of prevention, and it is often a rapidly progressive and fatal disease. The only validated marker for ovarian cancer is CA125, which is detectable in the serum of more than 80% of women with ovarian cancer. In contrast, cervical cancer is the third leading cause of cancer death in women. Over half a million new cases are diagnosed every year worldwide. The most common histological type of cervical cancer is squamous cell carcinoma which accounts for more than 80% of all cervical cancers. An increasing number of reports indicate that other factors are involved along with human papilloma virus (HPV) to induce cervical carcinogenesis. A routinely used biomarker for advanced cervical cancer is SCC, which is detectable in the serum of less than 50% of women with cervical cancer. CA125 and SCC are reliable only in monitoring the response to treatment or recurrence, but not as a diagnostic or prognostic marker ( 38 ). Thus, there is considerable interest in identifying molecular diagnostic and prognostic indicators to guide treatment decisions. It is well known that one of the carcinogenic biomarkers, human telomerase reverse transcriptase (hTERT) ( 39 , 40 ), is not only expressed in mild dysplastic lesions in cervices ( 41 ), but is also often expressed in gynecological malignancies. A total of 176 consecutive patients (47 patients with ovarian cancer, 63 with uterine cancer, 2 with other gynecologic cancer, 2 with border lesions, and 62 with benign diseases) that were admitted to Tottori University Hospital between December 2003 and January 2005, were enrolled in this study. Of the patients with benign disease, 41 had benign ovarian diseases including 11 with an ovarian dermoid cyst, 4 had endometriosis in the ovary, 13 had an ovarian cyst and 21 had benign uterine diseases including 1 patient with invasive hydatidiform mole, 4 with adenomyosis of the uterus, 14 with uterine myoma and 2 with uterine prolapse. Twenty healthy patients served as controls. The mean age of the patients was 55 years (range 18 to 85 years). CA125 (ChemilumiACS-CA125II, Bayer Japan, Tokyo) and SCC (SCC RIAbeads, SRL, Tokyo, Japan) in the serum were measured in routine laboratory tests. Human papilloma virus (HPV) was not examined for uterine lesions. The patients were diagnosed by chief complaints, ultra-sonography, computed tomography (CT), CA125 or SCC, cytology or biopsy under internal examination, histological examination after laparoscopy or surgical therapy. The clinicopathological findings, gender, age, etiology, histological findings, CA125 for ovarian disease, SCC for uterine disease, tumor size, clinical staging, and presence of recurrence were evaluated. According to this quantitative assay, the copy numbers of hTERT mRNA were significantly higher in the gynecologic cancer patients than in the healthy individuals (p < 0.01, each). Clinicopathological findings showed significant relations to hTERT mRNA in the serum ( Table 5 ). By multivariate analysis, the hTERT mRNA level was significantly correlated with age, the presence of cancer, ovarian disease in organs, ovarian malignancies, tumor size, and CA125 (p < 0.001, p = 0.004, p = 0.045, p = 0.004, p = 0.044, p = 0.035 and p < 0.001, respectively). On the other hand, EGFR mRNA did not show any significant correlation with any parameters or with other markers. SCC was significantly associated with uterine malignancies (p = 0.021). With the Friedrich test, hTERT mRNA, SCC, and CA125 were significantly correlated with the clinical stage (p < 0.001, p = 0.033, and p = 0.028, respectively). Pearson’s relative test between clinical parameters revealed that the hTERT mRNA level was significantly associated with the presence of cancer (p = 0.009) and, in the ovary, with CA125 and age (p = 0.049 and p = 0.045, respectively). To examine the sensitivity and specificity of tumor markers for diagnosis of gynecologic malignancies, ROC curve analysis showed the sensitivity/specificity of hTERT mRNA for gynecologic malignancies to be 74.4%/74.1% (data not shown). For ovarian malignancies, area under the curve (AUC) of hTERT mRNA and CA125 were 90.9% and 83.3%, respectively, and the sensitivity/specificity of hTERT mRNA in ovarian cancer is 95.0%/90.0% ( Fig. 3C ). The sensitivity/specificity of hTERT mRNA and SCC in uterine cancer is 70.1%/81.5% and 50.0%/67.5% respectively. On the other hand, the sensitivity/specificity of hTERT mRNA and CA125 in ovarian cancer is 100%/76.5% and 100%/75.5% respectively. For uterine malignancies, AUC of hTERT mRNA and SCC were 72.2% and 30.6% ( Fig. 3D ), respectively. The optimal cut-off values for hTERT mRNA and EGFR mRNA were calculated as 10 4.1 copies/0.2ml and 10 2.99 copies/0.2 ml, respectively. All patients were stratified into three categories based on the timing of blood sampling (before, during, and after therapy) and the therapeutic effect such as anti-tumor agents or surgical treatment was estimated by a t test. Although there was no correlation between before and during blood sampling, hTERT mRNA significantly decreased after therapy, compared with before and during the therapy (p < 0.05, each) ( Fig. 4B ). This suggests that the measurement of hTERT may be useful for the evaluation of a therapeutic effect. In uterine malignancies, hTERT mRNA and SCC were significantly useful biomarkers to evaluate therapeutic effect (p = 0.001 and p = 0.026, respectively). In ovarian malignancies, hTERT mRNA and CA125 were significantly useful for the diagnosis of cancer (p = 0.001 and p = 0.043, respectively). The copy number of hTERT mRNA in serum was significantly correlated with that in cancer tissue (p = 0.028 in Wilcoxon’s test, p = 0.035 in the paired t test). We attempted to elucidate the diagnostic evaluation of serum hTERT mRNA for gynecologic malignancies with our method. In 174 female patients with gynecological lesions (47 with ovarian lesions, 63 with uterine lesions, 2 with malignancies in other gynecological lesions, and 62 benign lesions) and 20 healthy individuals, we measured serum hTERT mRNA and EGFR mRNA by our real-time quantitative RT-PCR in the same way as lung cancer. We examined their sensitivity and specificity in cancer diagnosis, clinical significance in comparison with conventional tumor markers, and their correlations with the clinical parameters by using multivariate analyses. Serum hTERT mRNA showed higher values in patients with gynecologic cancers than in those with benign diseases and healthy individuals. The hTERT mRNA level independently correlated with the presence of cancers (p = 0.004 for both ovarian and uterine cancer) and clinical stage (p < 0.001). The sensitivity and specificity of hTERT mRNA in cancer diagnosis was 74.4% and 74.1%, respectively. The hTERT mRNA level showed a significant correlation with CA125 by Pearson’s relative test (p = 0.035) and with histological findings in ovarian cancer by the Friedrich test (p < 0.004). EGFR mRNA never displayed any differences between the diseases. hTERT mRNA is useful for diagnosing gynecologic cancer and is superior to conventional tumor markers. Therefore, serum hTERT mRNA is a novel and available biomarker for gynecologic malignancies.

Additional diagnostic methods (e.g. imaging), or combination of other biomarkers would be still needed to determine the type and location of the tumor. In our investigations, circulating mRNA is a very sensitive molecule to detect diseases in organs with rich blood flow or systemic inflammatory disease. The presence of acute phase diseases, including cancer progression, could be detected if mRNA is tested for within one day of blood sampling. However, this assay has a limitation. It can be influenced by organs with rich blood flow such as the liver. The disease of interest may be masked, for example in patients suffering from progressive liver disease. Based on these actual circumstances, it is likely that further developments over the next few years in the field of circulating RNA will provide us with new diagnostic and monitoring possibilities.

In a subsequent quantitative study, we have improved the sensitivity to detect the instable nucleotides in blood by removing cellular proteins and minimizing the contamination of cellular nucleic acids in serum and a primer set which can amplify hTERT mRNA efficiently ( 35 ). Furthermore, the correlation between tumor tissue and serum in terms of hTERT mRNA was demonstrated in Figure 1B , suggesting that hTERT mRNA detected in serum is derived from tumor cells. AFP is widely used as a reliable marker of HCC, not in earlier stage but in advanced stage ( 42 ). Since HCC recurs repeatedly and polyclonally due to biological characteristics even after any treatments, the monitoring of serum hTERT mRNA might make it possible to diagnose the recurrence earlier. In this respect, we prospectively have to conduct a follow-up study after the treatment of HCC (manuscript in preparation). hTERT mRNA expression was found to be closely associated with a well to moderate degree of differentiation of HCC. Nakashio et al. previously reported a significant correlation between HCC differentiation and telomerase expression ( 43 ). The results in the present study confirmed their findings. hTERT mRNA showed more sensitivity and specificity compared with AFP mRNA in HCC diagnosis. AFP mRNA did more sensitivity and specificity compared with AFP level ( 5 ). The higher specificity of hTERT mRNA may be related to the fact that AFP mRNA is produced in HCC cells and injured hepatocytes. However, hTERT is produced mainly in HCC cells. Waguri et al proved that hTERT mRNA in circulating cancer cells, derived from HCC tissues, can be detectable by using a cell-sorting system that they developed. The authors indicated that HCC cells released from the original HCC were within 10 mm in size ( 44 ) and suggested that the value is consistent with the result of the mRNA detection method. Our studies suggest that quantification of hTERT mRNAs in serum has diagnostic implications for NSCLC and ovarian cancer as well as HCC. We will evaluate the correlation between prognosis and hTERT mRNA ( 45 ), and assess the availability of hTERT mRNA in other hypervascular cancers by comparing hTERT mRNA with other tumor markers. We are performing a large-scale study with more than 500 patients to confirm our results for the monitoring and detection of HCC. Since our assay is difficult to perform manually, we are developing an automatic, concise and rapid diagnostic apparatus with a reagents kit to be introduced into clinics, even at the primary care level. In the diagnostic field of medicine, current patent laws involved in intellectual property rights are limiting diagnostic progression. This method will be suitable for the diagnosis of a number of hTERT-positive malignancies and for cancer-specific RNA genes or other genes showing weak protein levels. The systemic introduction of this assay into the primary care level can be expected in the near future. In pulmonary and gynecological malignancies, we suggest that the correlation with response/survival may be more evident if this more sensitive mRNA-based detection assay is used. The induction of a combination of hTERT mRNA and EGFR mRNA into the early diagnosis of lung cancer may improve the follow-up of patients. Since the data shows that hTERT and EGFR mRNA levels are associated with previously detected lung cancers, it does not demonstrate whether the method is sufficiently robust for early detection. However, the evaluation of hTERT mRNA combined with EGFR mRNA may be a useful biomarker to diagnose and assess the clinical stage and effect of treatments in patients with lung cancer. Further case-control studies should be planned to evaluate both long-term and former smokers who do not have lung cancer. hTERT mRNA is useful in the diagnosis of gynecologic cancers as well and is superior to conventional tumor markers. Therefore, serum hTERT mRNA is a novel and available biomarker for gynecologic malignancies.

A recent development in this new field is the identification of tumor-related RNA in the plasma/serum of cancer patients ( 6 ). These include tyrosine kinase mRNA ( 7 ), telomerase components ( 8 , 9 ), the mRNAs that are encoded by different tumor-related genes ( 10 , 11 , 12 , 13 , 14 ), and viral mRNA ( 14 ). As more RNA than DNA markers can be detected in the circulation of cancer patients, an assay using RNA markers produced higher sensitivity than other conventional assays. In one study, two telomerase markers of breast cancer yielded 44% positive rates ( 8 ). However, telomerase RNA seems to be a promising marker as it can be detected even in the serum of patients with small, undifferentiated breast cancers without any metastatic lesions. Dasi et al. showed that circulating telomerase RNA is a sensitive marker, using real-time reverse transcription-polymerase chain reaction (real time RT-PCR) ( 9 ). In their study, 8 of 9 plasma samples from colorectal cancer patients and 9 of 9 plasma samples from patients with lymphoma tested positive for human telomerase reverse transcriptase. The plasma samples of all 10 healthy individuals were negative.

In other malignancies such as thyroid cancer and pancreatic cancer (in preparation), this assay has been applied (manuscript under preparation for publication). We are investigating other disease-specific mRNAs as biomarkers to be applied in the clinic or in primary care. The field of medical care we are engaged in is the diagnosis of intractable disease ( Table 6 ). It is necessary to find improved mRNA such as hTERT, compared to Positron Emission Tomography (PET), as a diagnostic vehicle. We are studying inflammatory diseases such as fulminant hepatitis ( 46 ) or acute respiratory distress syndrome (in preparation), and lifestyle-related diseases, autoimmune diseases, and disease from gene dysfunction which are difficult to diagnose or estimate the clinical condition.