Clinical samples (142 cervical tissues, including 44 HPV16-positive, 18 HPV18-positive and 80 HPV-negative samples) were collected from Ubon Ratchathani Cancer Hospital (Ubon Ratchathani, Thailand) during the year 2015. The human research ethics committee of Ubon Ratchathani Cancer Hospital approved the protocol (EC04/2015). DNA was extracted from all clinical samples using the ExiPrep Dx Viral DNA Kit (Bioneer Corporation, Daejeon, Korea) according to the manufacturer's protocol, and the initial detection of DNA samples for high-risk HPV genotyping was performed using nested PCR according to Sotlar et al (12). All DNA samples were stored at −20°C until further use.

The LAMP primer sets were obtained from a previous study (8). These primer sets (listed in Table I) comprised a 5′biotin-labelled forward inner primer (FIP; termed here as BIP), outer primers (F3, B3) and loop primers (LF, LB), synthesized at Pacific Science Co, Ltd. (Bangkok, Thailand). The optimal conditions for the LAMP method were determined after assessing varying reaction temperatures and times. The 25 µl reaction mixture contained 1.6 µM of each inner primer, 0.2 µM of each outer primer, 0.8 µM of each loop primer, 1.4 mM dNTPs (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA), 0.8 mM betaine (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany), 6 mM MgSO₄, 8 units of Bst 2.0 DNA polymerase (New England BioLabs, Inc., Ipswich, MA, USA), 1x Bst buffer (New England BioLabs, Inc.) and 3 µl of HPV16 and HPV18 plasmid DNA at 10⁴ copies (provided by Professor Ethel-Michele de Villiers, The German Cancer Research Centre, Heidelberg, Germany) (13). The reaction was conducted at different temperatures (60, 63 and 65°C) for 10, 20, 30, 40, 50 and 60 min.

The HPV16 and HPV18 LAMP products were hybridized with the appropriate DNA probe (Pacific Science Co, Ltd.) labelled with fluorescein isothiocyanate (FITC) at the 5′-end, designed according to the HPV16 and HPV18 sequences between the LB and B3 primer targets (Table I). The LAMP-LFD conditions were optimized after assessing various hybridization times and DNA probe concentrations. The reaction mixtures containing LAMP product and DNA probe (200 and 20 pmol, respectively) were incubated at 65°C for 5, 10 or 15 min. Subsequently, 8 µl of hybridized product was added to 150 µl of assay buffer in a new tube (14). An LFD strip (Milenia HybriDetect; Milenia Biotec GmbH, Giessen, Germany) was dipped into the reaction mixture for 5 min. A red-purple line was observed at the control line for all strips, which confirmed that the test was correctly operated.

To detect sensitivity limits, the HPV16 and HPV18-containing plasmid DNA, varying from 10⁵ to 10⁰ copies, was used as a template for the LAMP reaction. The plasmids pBR322-HPV16 and pBR322-HPV18 were kindly provided by Professor Ethel-Michele de Villiers (13). The LAMP products were detected using LAMP-turbidity (15) compared with LAMP-LFD.

The specificity of the DNA probe was examined using 10⁴ copies of HPV16 and HPV18 plasmid DNA for the LAMP reaction, and subsequently, the LAMP products were detected using the LFD assay. The negative control was performed using the same test, but water was added instead of DNA.

All 142 clinical samples were examined using LAMP-turbidity (15) and LAMP-LFD assays. The sensitivity and specificity of these assays were calculated using standard formulae based on the results of nested PCR.

All samples with inconsistent results were further analysed using qPCR. The set of primers and probes for E2/E6 of HPV16 and HPV18 were designed according to previous studies (16,17). qPCR was conducted using an Exicycler 96 system (Bioneer Corporation), as previously described (16,17), and 10 µl of the clinical samples were used as the DNA templates. The standard curves were obtained from amplification of a dilution series using 10⁸, 10⁷, 10⁶, 10⁵, 10⁴, 10³, 10² and 10¹ copies of HPV16 and HPV18 plasmid DNA.

The LAMP-LFD assay was designed to detect biotin-labelled LAMP products hybridized to a FITC-labelled specific DNA probe (design explained in the schematic of Fig. 1). Subsequently, the FITC-labelled specific DNA probe was recognized by a gold-labelled anti-FITC antibody. This triple-labelled complex was then trapped at a test line using avidin, generating a red-purple band (positive result). By contrast, non-LAMP products hybridized with the FITC-labelled specific probe and bound the gold-labelled anti-FITC antibody, but did not bind avidin, due to lack of biotin; therefore, this complex moved past the test line and was trapped at the control line (18).

The LAMP conditions for HPV16 and HPV18 detection were optimized by assaying variable temperatures and reaction duration times, and the turbidity of the amplified product was observed using the naked eye. The results demonstrated that all three temperatures tested (60, 63 and 65°C) generated a slightly different turbidity; however, the highest turbidity was observed at 65°C (data not shown). In addition, the turbidity was initiated at 20 min, but it was difficult to observe under the naked eye. Therefore, the best set of conditions, 65°C for 30 min, was selected for the present study.

The specific FITC-labelled DNA probes were hybridized with the HPV16 and HPV18 LAMP products. It was determined that the best conditions for LAMP product hybridization were as follows: 20 pmol of DNA probe as hybridization input and incubation at 65°C for 15 min (data not shown). Subsequently, 8 µl of hybridization product was added to 150 µl of assay buffer in a new tube. Next, the LFD strip was dipped into this reaction mixture for 5 min, and positive result was denoted by the presence of 2 red-purple lines on LFD strips using the naked eye.

The limits of LAMP-turbidity detection for HPV16 and HPV18 were 10³ and 10¹ plasmid copies, respectively (Figs. 2A and 3A, respectively). The LAMP-LFD assay showed a limit of detection of 10¹ and 10⁰ copies of HPV16 and HPV18, respectively (Figs. 2B and 3B, respectively). Therefore, the detection limits of LAMP-LFD were ~100 and 10-fold more sensitive for the detection of HPV16 and HPV18, respectively, compared with the LAMP-turbidity method.

The specificity of the DNA probe was examined using 10⁴ copies of HPV16 and HPV18 as templates for HPV18 and HPV16 LAMP, respectively. The results revealed no cross-reactivity between HPV16 and HPV18 using the LAMP-LFD assay (Fig. 4).

All 142 clinical samples, including HPV16-positive (n=44), HPV18-positive (n=18) and HPV-negative (n=80) samples, were examined by the LAMP turbidity assay compared with the novel LAMP-LFD assay. The results from the LAMP-turbidity assay revealed that 44 and 18 samples were positive for HPV16 and HPV18, respectively, while the results from the LAMP-LFD assay revealed that 57 samples and 27 samples were positive for HPV16 and HPV18, respectively (Table II). Therefore, a higher number of samples gave positive readings with the novel LAMP-LFD assay developed in the present study than with the nested PCR, gold standard method.

The samples with inconsistent results between LAMP-LFD assay and the gold standard nested PCR method were further analysed using qPCR. The limits of qPCR detection for HPV16 and 18 were 10³ copies (data not shown). When the 22 samples that exhibited inconsistent results between the LAMP-LFD assay and the gold standard method were repeated using qPCR, the results demonstrated that 13 of 13 samples and 7 of 9 samples were positive by qPCR for HPV16 and HPV18, respectively, while 2 samples generated negative results by qPCR (Tables III and IV).