Rapid detection of metastasis of gastric cancer using reverse transcription loop-mediated isothermal amplification

Abstract

Tailor-made surgeries for patients with solid malignancies have been under consideration on the basis of the development of new approaches for minor metastatic foci of malignant tumors. Accurate and reliable methods to detect metastases in biopsy specimens with certain rapidity are essential for the performance of these surgeries. The aim of this study was to develop a rapid and practical method to detect metastasis in specimens from patients with gastric carcinoma with the use of reverse transcription loop-mediated isothermal amplification (RT-LAMP) reaction, a novel technique for detecting mRNA expressions of targeted sequences with high sensitivity, specificity and rapidity under isothermal conditions. RT-LAMP primers to detect cytokeratin19 (CK19) mRNA were generated and 92 lymph nodes (LNs) obtained from 9 patients with gastric cancer were tested for tumor metastases with this technique. Among 92 LNs, 15 were metastasis-positive by routine histopathological examination. RT-LAMP reaction detected CK19 expression in all of the pathologically positive LNs and in 16 of 77 negative LNs. Nested RT-PCR assay for CK19 expression was also performed on 2 of the 9 cases including 32 LNs. The agreement rate of CK19 expression detection by RT-LAMP and RT-PCR analysis was 31/32 (97%). The RT-LAMP technique showed similar sensitivity to detect metastases as nested RT-PCR assay, with a rapidity comparable to that of intraoperative histopathological examination with frozen sectioning and hematoxylin and eosin staining. This method is expected to play an essential role in the performance of tailor-made surgeries in the near future. © 2006 Wiley-Liss, Inc.

The concept of sentinel lymph node (SLN), presumably the first site of metastasizing tumor cells via the lymphatic system, emerged in the middle of the last century, and so-called tailor-made surgery based on this concept has been conducted in the field of breast cancer and cutaneous malignant melanoma. Validity of the same concept has been tested in the field of gastrointestinal cancers as well, and favorable results are reported on the basis of conventional histopathological analysis and/or immunohistochemical (IHC) analysis of specimens.1, 2 Some of the major gastrointestinal surgeries are well known to cause unavoidable systemic disorders after surgery, and as a result the applicability of less invasive treatments according to individual tumor extent is now under investigation. The SLN concept could provide valuable information that might assist in making such tailor-made surgery a reality in the field of gastrointestinal cancers as well.

In the meantime, some authors have introduced assay techniques based on reverse transcriptase-polymerase chain reaction (RT-PCR) to detect metastasis in regional lymph nodes (LNs) of gastric cancers,3, 4, 5, 6, 7, 8 and they showed a higher incidence of revealing minor tumor deposits in LNs compared with the technique of routine hematoxylin and eosin (H&E) or IHC examination. The SLN concept has also been shown to be applicable to gastric cancer even when metastasis is analyzed by RT-PCR,9 suggesting that tailor-made gastric cancer surgeries should be performed with micrometastasis-detecting techniques of high sensitivity such as RT-PCR. However, the technique requires complex procedures and at least several hours to obtain a final result, and thus such surgeries have not yet been widely accepted, because local or LN recurrence of gastrointestinal tumors results in grim survival for patients. As a consequence, a metastasis-detecting technique with a similar sensitivity and swiftness to be performed as RT-PCR, is eagerly awaited.

Loop-mediated isothermal amplification (LAMP) is a novel DNA amplification technique first reported by Notomi and co workers.10, 11 It employs a strand displacement DNA polymerase and a set of specially designed primers recognizing different regions in the targeted DNA sequence, and this technique amplifies a target DNA with high specificity, efficiency and rapidity under isothermal conditions. This technique is also applicable to RNA upon use of reverse transcriptase together with DNA polymerase (reverse transcriptase-LAMP: RT-LAMP).12, 13 Amplification of a targeted gene is detectable in real-time fashion by an increase of turbidity of the solution derived from a side product of the reactions. Thus, the reaction could be completed in a single test tube and within 1 hr.

In the current study, this new technique was applied for the detection of minor metastases of gastric cancer by targeting cytokeratin19 (CK19) gene, which is expressed on cells of epithelial origin. The potential of this new technique was evaluated for its sensitivity to detect tumor metastases in LNs during gastric cancer surgeries.

AMV, avian myeloblastosis virus; CEA, carcinoembryonic antigen; CK18, cytokeratin18; CK19, cytokeratin19; CK20, cytokeratin20; GAPDH, glyceraldehyde-3 phosphate dehydrogenase; H&E, hematoxylin and eosin; histology⁺/histology⁻, metastasis positive/negative by histopathological examination; hTRT, human telomerase reverse transcriptase; IHC, immunohistochemistry; LNs, lymph nodes; RT-LAMP, reverse transcription loop-mediated isothermal amplification; RT-LAMP⁺/RT-LAMP⁻, CK19 mRNA expression positive/negative by RT-LAMP reaction; RT-PCR, reverse transcriptase-polymerase chain reaction; SLN, sentinel lymph node.

Material and methods

Clinical specimens

Study participants consisted of 9 patients with primary gastric adenocarcinoma who underwent gastrectomy with curative intent and agreed with the written protocol of genetic analysis of clinical specimens, at the Department of Surgery, Chiba University Hospital, between October 2003 and July 2005. Five men and 4 women, with a mean age of 60 years (range, 43–76 years), were included in this study. Clinicopathological features of the 9 patients are shown in Table I. A total of 92 LNs of more than 2 mm in diameter were collected from these 9 patients. Pathological diagnosis and classification were determined based on the Japanese Classification of Gastric Carcinoma.14 The nodes were bisected, and halves were fixed in 10% buffered formalin and embedded in paraffin for H&E staining. The opposite halves of each node were stored at −80°C until genetic analysis. Among these 92 LNs, 15 LNs were revealed as metastatic-positive (histology⁺), and 77 LNs were metastatic-negative (histology⁻) after routine histopathological analysis. A total of 18 mesenteric LNs were obtained from 4 patients undergoing surgery for nonmalignant conditions (2 LNs from a patient with Crohn's disease undergoing ileocecal resection, 2 LNs from a patient with volvulus of the sigmoid colon undergoing sigmoid colon resection, 7 LNs from a patient with diverticulitis of sigmoid colon undergoing anterior resection of rectosigmoid, and 7 LNs from a patient with Crohn's disease undergoing ileocecal resection and resection of transverse colon) as negative controls.

Table I. Clinicopathological Features and Metastatic Lymph Node Distributions of the Patients Studied

Case no.	Sex	Age	pT	Histological type	No. of lymph nodes
					HE-/LAMP^−	HE-/LAMP^+	HE+/LAMP^+	Total
pT: depth of tumor invasion according to UICC-TNM classification, por: poorly differentiated adenocarcinoma, tub: tubular adenocarcinoma, sig: signet-ring cell carcinoma, muc: mucinous adenocarcinoma, HE^−: metastasis negative by hematoxylin-eosin staining, HE^+: metastasis positive by hematoxylin-eosin staining, LAMP^−: CK19 mRNA expression negative by RT-LAMP reaction, LAMP^+: CK19 mRNA expression positive by RT-LAMP reaction.
1	F	45	1	por	6	3	1	10
2	F	55	1	tub	7	3	0	10
3	M	43	1	sig	10	1	0	11
4	M	64	3	tub	0	1	1	2
5	M	64	2	tub	17	3	0	20
6	M	65	2	por	3	0	0	3
7	F	60	1	por	4	0	0	4
8	F	76	2	tub	10	5	2	17
9	M	68	3	muc	4	0	11	15

Carcinoma cell lines

Five human gastric carcinoma cell lines (MKN1, MKN28, MKN45, MKN74, KATO-III), 2 esophageal carcinoma cell lines (T.Tn, TE2), 1 colon carcinoma cell line (RKO) and 2 leukemia cell lines (K562, Jurkat) were tested for characterization of RT-LAMP reaction targeting CK19.

Preparation of total RNA or lysate from tumor cell lines and clinical samples

Each carcinoma cell line and frozen clinical specimen were ground down with a Multi Beads Shocker (Yasui Kikai, Osaka, Japan). Total RNA was extracted from these samples by RNeasy mini kit (Qiagen, Hilden, Germany) in accordance with the manufacturer's protocol. All clinical samples from case no. 8 indicated in Table I were divided into 2 sets, and total RNA was prepared from 1 set, and lysate from the other. For preparation of lysate, each sample was dissolved in 0.6 ml of lysis buffer [1% Nonidet-P40 (Nacalai Tesque (Kyoto, Japan)/100 mM Tris-HCl (pH 8.0)/0.5 unit/μl RNaseOUT Ribonuclease Inhibitor (Invitrogen, Carlsbad, CA))]. The solutions were shaken for 1 min and centrifuged at 15,000 rpm for 10 min. Then, the supernatant was collected and diluted at 5:1 for use.

RT-LAMP reaction

LAMP reactions were conducted according to the published procedures.11 This reaction relies on auto-cycling strand displacement DNA synthesis that is performed with a DNA polymerase with high-strand displacement activity and 4 specific primers termed inner primers (called forward inner primer (FIP) and backward inner primer (BIP)) and outer primers (called B3 and F3) recognizing 6 independent sequences, and specifically synthesizes a large amount of amplification products, a mixture of stem–loop DNAs with several inverted repeats of the target DNA and cauliflower-like structures with multiple loops. Loop primers hybridize to the stem-loops, except for the loops that are hybridized by the inner primer, and prime strand displacement DNA synthesis. When the target DNA is amplified by LAMP reaction, a white precipitate derived from magnesium pyrophosphate (byproduct of LAMP reaction) is observed.15 So, the LAMP method does not require special reagents or electrophoresis to detect the amplified DNA. Further, this reaction is accelerated by the use of 2 additional loop primers (called Loops F and B).16 In addition, the LAMP method is also applicable for RNA upon use of reverse transcriptase together with DNA polymerase.12, 13

LAMP primers targeting CK19 gene were designed using PrimerExplorer version 3 (Fujitsu System Solutions, Tokyo, Japan). The primer sequences are shown in Figure 1a. To quantify and prove the integrity of isolated RNA, RT-LAMP for glyceraldehyde-3 phosphate dehydrogenase (GAPDH) was also carried out. The primer targeting GAPDH was also designed using the same software, and its sequences are shown in Figure 1b.

Figure 1. Primer design for detection of CK19 and GAPDH mRNA by RT-LAMP. (a) Upper panel: sequence alignment of CK19 cDNA. Underlined portions represent the primer location of RT-LAMP for the detection of CK19 mRNA. Square (genome positions 555–559) shows the location of Hinf I site. Lower panel: sequences of the primers for the detection of CK19 mRNA. (b) Upper panel: sequence alignment of GAPDH cDNA. Underlined portions represent the primer location of RT-LAMP for the detection of GAPDH mRNA. Lower panel: sequences of the primers for the detection of GAPDH mRNA.

RT-LAMP method was carried out with a 25-μl total reaction mixture volume with a Loopamp RNA amplification kit (Eiken Chemical Co., Tokyo, Japan) containing 40 pmol each of inner primers FIP and BIP, 5 pmol each of outer primers F3 and B3, 20 pmol each of loop primers loop F and loop B, 35 pmol dNTPs, 20 μM Betamine, 0.5 μM Tris-HCL (pH 8.8), 0.25 μM KCL, 0.25 μM (NH₄)SO₄, 0.2 μM MgSO₄, 0.2% Tween20, 1.0 μl Enzyme Mix (mixture of BstDNA polymerase and avian myeloblastosis virus (AMV) reverse transcriptase) and 5 μl RNA at a constant temperature of 63.5°C for 60 min.

The DNA fragments synthesized by LAMP reaction were detected based on the production of a white precipitate of magnesium pyrophosphate. For real-time turbidity monitoring, absorbance of the reaction mixture at 650 nm was measured. Temperature control for the LAMP reaction and turbidity measurement were performed using a turbidimeter (LA-200; Teramecs Co., Kyoto, Japan) especially developed for DNA analysis by LAMP reaction.

The maximum turbidity value of the negative control materials including cell line cells of nonepithelial origin and LN samples from patients undergoing surgeries for benign disease was 0.0034 within 60-min reaction time. None of these samples showed increased turbidity after 60 min. Thus, the cut-off value was set at 0.1, the minimum scale indicated in the real-time monitoring window of the turbidimeter used in this study.

Immunohistochemical staining

The protocol of IHC analysis was described elsewhere.1 Briefly, tissue sections were deparaffinized, dehydrated and subjected to heat antigen retrieval in 10 mM citrate buffer, pH 6.0. Then, the sections were stained with pan-cytokeratin monoclonal antibodies AE1/AE3 (Dako Cytomation, Glostrup, Denmark) and secondary antibody reagent (Envision/AP; Dako Cytomation) on the TechMate Horizon automated staining system (Dako Cytomation).

Nested RT-PCR assay

The protocol of nested RT-PCR assay was designed as described previously.17 The primer sequences were as follows: for CK19 (outer), 5′-AAGCTAACCATGCAGAACCTCAACGACCGC-3′ and 5′-TTATTGGCAGGTCAGGAGAAGAGCC-3′ for CK19 (inner), 5′-TGCCCGGAGCGGAATCCACCTC-3′ and 5′-CCTGCGCAGGGTGCTGGATGAGCTGAC-3′ and for β-actin, 5′-CGGAATTCATCATGTTTGAGACCTTC-3′ and 5′-AGAAGCTTCATCTCTTGCTCGAAGTCCAGGGCG-3′. Total RNA (0.6 μg) was reverse-transcribed in a 20-μl reaction mixture containing 10 mM Tris-HCl (pH 8.3), 10 mM KCl, 2 mM dithiothreitol, 2 mM MgCl₂, 0.1 mM spermidine, 0.4 mM dNTP, 25 μM oligo(dT)₁₈, 10 U of RNase inhibitor and 5 U of AMV reverse transcriptase (Promega, Madison, MI) for 1 hr at 42°C, and the reaction was terminated by heating at 99°C for 5 min. First-round PCR was performed using 2 μl of cDNA from the reverse transcription. The PCR mixture contained 20 pmol of each CK19 outer primer and each β-actin primer, together with 10 mM Tris-HCl (pH 8.8), 50 mM KCl, 1.5 mM MgCl₂, 0.1% Triton X-100, 10 mM dNTP and 0.5 U of DyNazyme™ II DNA polymerase (Finnzymes Oy, Espoo, Finland) in a total volume of 50 μl. The PCR profile included 1 cycle at 94°C for 5 min, 40 cycles at 94°C for 40 sec, 60°C for 1 min and 72°C for 50 sec, and a final extension at 72°C for 10 min. The second round of PCR was carried out under the same conditions, except only 30 cycles were performed. The PCR products were analyzed by electrophoresis on 2% agarose and visualized under UV light after ethidium bromide staining. Negative controls consisting of reactions with no added RNA or cDNA and omission of reverse transcriptase were processed in parallel with patient samples in every RT-PCR run.

Results

Expression of CK19 mRNA in various carcinoma cell lines

RT-LAMP reaction targeting CK19 was performed using mRNAs that were extracted from various carcinoma cell lines. In the present study, 5 gastric cancer cell lines (MKN1, MKN28, MKN45, MKN74, KATO- III), 2 esophageal cancer cell lines (T.Tn and TE2), 1 colon cancer cell line (RKO), and 2 hematopoietic cancer cell lines (K562 and Jurkat) were used. CK19 mRNA expression was detected exclusively in gastrointestinal carcinoma cell lines, and not in hematopoietic carcinoma cell lines (Fig. 2a). The RT-LAMP reaction could amplify the 256-bp target sequence within the CK19 cDNA. Amplification was observed as a ladder-like pattern on the gel due to the formation of a mixture of stem–loop DNAs with various stem lengths and cauliflower-like structures of the product (Fig. 2b). Restriction enzyme digestion of the product at a Hinf I site, a unique site within the targeted sequence, revealed the expected size of the fragment on the gel.

Figure 2. (a) Detection of CK19 mRNA expression in various carcinoma cell lines by RT-LAMP method. (b) Electrophoresis of RT-LAMP products from mRNAs extracted from various carcinoma cell lines after digestion with Hinf I. 1: 100-bp DNA ladder, 2: MKN-1, 3: MKN-1/Hinf I, 4: MKN-28, 5: MKN28/Hinf I, 6: MKN-45, 7: MKN-45/Hinf I, 8: MKN-74, 9: MKN-74/Hinf I, 10: KATO-III, 11: KATO-III/Hinf I, 12: T.Tn, 13: T.Tn/Hinf I, 14: TE2, 15: TE2/Hinf I, 16: RKO, 17: RKO/Hinf I, 18: K562, 19: Jurkat, 20: negative control.

Sensitivity for identifying gastric carcinoma cell line cells by RT-LAMP technique

Sensitivity assay for detecting a gastric carcinoma cell line (MKN1) was performed. Although a 5–6 min detection-time delay was observed in the analysis with cell lysate, as low as 10¹ MKN1 cells in 10⁶ normal lymphocytes from healthy volunteers were detectable with the RT-LAMP procedures targeting CK19 mRNA using either extracted mRNA or lysate of cell mixtures (shown in Figs. 3a and 3b). RT-LAMP products were also confirmed through electrophoresis after digestion with Hinf I (data not shown).

Figure 3. Analysis of detection sensitivity for gastric carcinoma cell line by RT-LAMP method. From 10⁶ to 10⁰ diluted MKN1 cells were added to normal white blood cells obtained from a healthy volunteer. (a) Extracted mRNAs from cell mixtures were used as templates. (b) Lysates from cell mixtures were used as templates.

Detection of metastasis in LNs with RT-LAMP reaction targeting CK19

A total of 92 regional LNs were obtained from 9 patients with gastric carcinoma, with 15 LNs being histology⁺ and 77 LNs histology⁻. All of these LNs were tested by RT-LAMP reaction targeting CK19 mRNA expression after preparation of mRNA. With this procedure, CK19 mRNA expression was detected in all 15 histology⁺ LNs and in 16 of 77 histology⁻ LNs (histology⁻/RT-LAMP⁺). No CK19 mRNA expression was detectable in the remaining histology⁻ LNs (histology⁻/RT-LAMP⁻). The results of RT-LAMP reaction of 17 LN samples from 1 patient (case no. 8 in Table I) are shown in Figure 4. All of the histology⁻/RT-LAMP⁺ LNs were revealed in the perigastric LN regions, not in the extraperigastric areas. The mean duration to detect CK19 mRNA expression in RT-LAMP⁺ LNs was 24.3 min (range, 2.0–55.0 min). In addition, lysates from LN samples from the same patients as indicated in Figure 4 were used as templates for the same RT-LAMP reaction, and the results are shown in Figure 5. The reaction targeting CK19 mRNA revealed that 7 of 17 LNs tested positive, identical to the results shown in Figure 4.

Figure 4. Detection of (a) CK19 and (b) GAPDH mRNA expression in all 17 LN samples from case no. 8 indicated in Table I by RT-LAMP method. Extracted mRNA was used as template.

Figure 5. Detection of (a) CK19 and (b) GAPDH mRNA expression in crude samples from clinical specimens. Lysates prepared from the same LN samples as tested in Figure 4 were used as templates.

CK19 mRNA expression was not detected in 18 mesenteric LNs obtained from 4 patients undergoing surgeries for benign diseases. GAPDH mRNA expression was detected in all LN samples examined (data not shown).

IHC analysis of LNs with pan-cytokeratin antibodies

For IHC analysis, additional single sections were newly prepared from all paraffin blocks of the 16 histology⁻/RT-LAMP⁺ LNs. No tumor deposits were revealed among these histology⁻/RT-LAMP⁺ LN sections stained with pan-cytokeratin antibodies AE1/AE3.

Sensitivity to detect metastasis by RT-LAMP technique versus RT-PCR assay

Both RT-LAMP and RT-PCR analyses targeting CK19 mRNA expression were performed on specimens from 2 cases (cases no. 8 and no. 9 in Table I), and the results of CK19 mRNA detection sensitivity by these 2 methods were compared. All of 13 histology⁺ LNs were RT-LAMP⁺ and 5 of 14 histology⁻ LNs were RT-LAMP⁺. CK19 mRNA expression was detected by RT-PCR analysis in all RT-LAMP⁺ LNs and in one of the histology⁻/RT-LAMP⁻ LNs of case no. 8. Thus, the agreement rate of the results was 31/32 (97%).

Discussion

Introduction of the SLN concept was one of the greatest advances in the field of surgery in the last century. SLN navigation surgery, so-called tailor-made surgery based on the tumor status of the SLN, has been conducted in breast cancer and malignant melanoma surgery. The same concept has been validated in the field of gastrointestinal cancer, and favorable results have been reported based on the results of histopathological analyses of surgical specimens.1, 2 If the metastatic status of gastric cancers in regional LNs can be precisely predicted during operations, patients free of LN metastases can undergo local resection of tumors and benefit greatly by avoiding any number of postgastrectomy disorders including reflux esophagitis, dumping syndrome, gastritis of the remnant stomach, anemia and so on. Routine histopathological examinations performed during operations have been reported to have limitations regarding the detection of minor metastatic lesions. To perform SLN navigation surgery for gastric cancer without sacrificing curability of tumors, techniques for meticulous analyses of metastasis that can be conducted during the operation have been eagerly awaited, as LN recurrence results in devastating outcomes in gastric cancers.

For routine histopathological diagnosis of materials, a 4-μm section is made from each of 2-mm slices of frozen LN materials in our hospital. Theoretically, it is impossible by this method to cover the entire LNs for metastasis because minor metastasis, as single or small clusters of tumor cells could lie between slices. In fact, Isozaki et al. reported that serial sectioning resulted in more accurate evaluation of the extent of LN metastasis than the routine one-section method in the analysis of gastric cancer specimens,18 and Mesker et al. also reported that detectability of metastasis in LN of colorectal cancer decreased as the distance between sequential sections of LN samples was increased.19 Thus, serial-step sectioning could be a strategy that achieves higher sensitivity in the detection of metastasis, but the flipside is that the more step sectioning is conducted, the greater will be the cost and the time required.

IHC analysis has been applied to identify minor tumor deposits in LNs,20, 21, 22, 23 and it has also been indicated that it could be done during surgery. But theoretically, this technique could only contribute to avoiding “overlooking” minor tumor deposits that are not easily distinguishable on H&E sections. As a matter of fact, an elaborate examination of H&E stained sections demonstrated that the difference in positive LN ratios between IHC and H&E analyses was negligible.1

To overcome the aforementioned obstacles, genetic analysis such as RT-PCR was introduced to achieve higher sensitivity for detecting metastasis, and favorable results have been reported.3, 4, 5, 6, 7, 8 Various kinds of target molecules such as carcinoembryonic antigen (CEA), human telomerase reverse transcriptase (hTRT), cytokeratin18 (CK18), cytokeratin20 (CK20), MAGE-3 and MUC-1 were used for RT-PCR analysis in these studies. However, most of them were conducted using a combination of 3 or 4 of these target molecules, because none of these individual markers were expressed universally on gastric cancer cells.

CK19 has been shown to be widely expressed by cancer cells of epithelial origin but not by lymphoid or hematopoietic cells.17 Thus, this molecule may be a suitable general marker of occult metastasis in LNs of patients with those cancers. However, it has rarely been used for detection of gastric cancer metastasis because of the existence of pseudogene interference of CK19 gene. Nested RT-PCR is required to overcome this problem.17

Besides the matter of target molecule, RT-PCR analysis requires complex procedures including extraction of RNAs, reverse transcription reaction from RNAs, PCR amplification of cDNA and agarose gel electrophoresis, in total taking almost a whole day to obtain the final results. The use of real-time assays such as TaqMan RT-PCR could significantly reduce PCR reaction time and make gel electrophoresis unnecessary, but several hours will still be required, not to mention the high-precision instruments for amplification. Thus, its application during surgery would hardly be feasible, nor would it be widely available in general hospitals.

Ishii et al. reported a TRC amplification technique, a novel genetic approach for detecting minor cancer metastasis. This method allows the detection of the mRNA expression level of cancer-specific genes without the reverse transcription process of mRNA, and can be performed within ∼1 hr. However, it still requires the extraction process of mRNA, which consists of complex procedures and is time-consuming.

LAMP reaction is a novel approach to the DNA amplification of target sequences, providing high sensitivity, specificity, and rapidity under isothermal conditions, and it could be conducted simultaneously with reverse transcription from mRNA.10, 11, 12, 13 The RT-LAMP technique requires only simple reaction procedures, the compact and inexpensive incubator/turbidimeter equipment costs less than $5,000, and less than 1 hr is needed to obtain the final results.15, 16 In this study, RNA extraction using spin columns was performed on the lysate of LN samples prior to RT-LAMP reaction, but even crude samples without RNA extraction were applicable to RT-LAMP reaction, with similar results being obtained. This technique might be one of the most promising candidates for analyzing the genetic features of biopsy samples obtained during surgery.

In this study, LAMP primers were generated to detect the CK19 sequence, and the performance of RT-LAMP reaction to detect gastric cancer cells was tested. As for cell line analysis, only those of epithelial origin were detected by this technique, with sensitivity as low as 10 cells per 10⁶ normal white blood cells. Analysis of clinical specimens showed similar sensitivity for detecting CK19 expression between RT-LAMP and nested RT-PCR. No CK19 expression in cells or cell lines of hematopoietic origin, nor CK19 pseudogene interference, was observed in RT-LAMP reaction, presumably due to the high specificity of the reaction that requires the recognition of 6 independent sequences within the target molecule.

Regarding the analysis of clinical specimens, RT-LAMP reaction detected CK19 expression in 16 of 77 histology⁻ LNs (histology⁻/RT-LAMP⁺) as well as in all histology⁺ LNs studied. This may indicate a higher specificity of RT-LAMP reaction in the detection of metastases as compared with histopathological analysis. However, CK19 has also been shown to be expressed in peritoneal mesothelial cells,8 and therefore contamination of these cells cannot be excluded. Thus, the combination with other target molecules such as CEA, hTRT, CK18, CK20, MAGE-3 and MUC-1 would be ideal for obtaining higher specificity, and comparative analysis and sensitivity analysis of combinations with such markers would certainly be preferable in the future. However, all of 18 mesenteric LN samples from patients with benign disease showed RT-LAMP⁻, indicating that the amount of contaminated mesothelial cells or the CK19 expression level on such cells could be generally below the threshold of the sensitivity of this new technique.

In addition, only a single section of each LN was subjected to histopathological analysis, while the current referred golden standard for frozen sections is at intervals of 2 mm, suggesting a lower sensitivity of histopathological analysis in this study for detecting metastasis compared with other reports. Furthermore, genetic and histopathological analyses were performed on the discrete halves of each LN studied, but it cannot be ruled out that small clusters of cancer cells might exist only in one half of the LN examined, not in the other. Thus, the underlying reason for the discrepancies between the results of RT-LAMP and histopathological analysis is still unclear. Given that the much higher sensitivity of RT-LAMP accounted for these discrepancies, further evaluation is required to elucidate the minimum size of metastatic lesion that can be detected by this technique, and the biological activities of such minor populations of tumor cells, because the clinical impact of single tumor cells or small clusters of tumor cells is still under debate. At least, accumulation of long-term follow-up data of patients with LNs of histology⁻/RT-LAMP⁺versus histology⁻/RT-LAMP⁻ should be essential for the estimation of this new technique.

In summary, the development of a novel rapid detection technique of metastasis for gastric cancer with the use of the RT-LAMP reaction was herein described. The technique showed a similar sensitivity to nested RT-PCR analysis, and could be performed as an alternative to an intraoperative pathological examination of frozen sections. It could provide the opportunity to perform reliable tailor-made surgery for gastric cancers as a common procedure in general hospitals.