A visualized automatic particle counting strategy for single‐cell level telomerase activity quantification

The accurate evaluation of telomerase activity, a typical cancer biomarker, is vital for early cancer screening. In this study, we developed a dark‐field microscopy (DFM) visual single‐particle detection scheme to detect telomerase activity based on automatic counting gold nanoparticles (AuNPs). This method started with attaching the telomerase substrate (TS) primer to the magnetic beads (MBs) through streptavidin‐biotin interaction. In the presence of telomerase and dNTPs, the TS primer was expanded with (TTAGGG)n repeat units to form the telomerase extension product (MBs‐telomerase extension product), which could be hybridized with the complementary DNA (cDNA) modified with AuNPs through Au‐S bonds (AuNPs‐SH‐cDNA). After magnetic separation and DNA double‐strand unwinding, AuNPs were collected from the supernatant, and the telomerase activity was quantitatively measured by visually counting bright spots based on DFM. This strategy achieved a limit of detection as low as 1 HeLa cell and distinguished telomerase activity among different cell lines, thus verifying its excellent sensitivity and specificity. Further, two common telomerase inhibitors (BIBR1532 and curcumin) were screened with the consistent IC50 values with other methods, respectively. It is worth mentioning that this strategy can clearly identify bladder cancer among various urinary diseases. Consequently, the visualized automatic particle counting strategy is potential as a powerful tool in early and noninvasive diagnosis of bladder cancer.


K E Y W O R D S
bladder cancer, dark-field microscopy, telomerase activity, visualized automatic particle counting INTRODUCTION The function of telomerase is compensating for its shortening by adding repeating sequences of TTAGGG only to the 3' end of the telomeres. 1 Most tumors (85%) highly express telomerase activity, while the telomerase activity of normal somatic cells is inhibited. 2,3 Therefore, it is important to detect telomerase activity with ultra-sensitivity, which acts as one of the main biological markers for screening and treatment of cancer in the early stages. [4][5][6] Since the discovery of telomerase in 1985, 7 a number of telomerase assays have emerged. Among them, the polymerase chain reaction (PCR)-based telomeric repeat amplification protocol (TRAP) invented by Kim et al. in 1994 is the most widely used. 8 Due to its disadvantages of time-consuming operation, and inability to analyze short telomerase products, 9 many modified TRAP assays have been developed. 10 However, due to PCR exponential amplification, telomere extension products may still suffer from the hidden danger of false-positive signals. 11 Subsequently, various telomerase activity assays, including fluorescent, [12][13][14] colorimetric, 15,16 electrochemical, 17,18 chemiluminescence, 19,20 nanoparticles-based assays, 21,22 have been investigated. It can thus be seen, convenient, ultra-sensitive, and specific systems for telomerase activity quantification are especially important.
Particularly, plasmonic nanoparticles combined with the spectrally resolved single-particle detection (SPD) method 23 have been popularly used to detect target molecules by virtue of their excellent optical properties, which is considered as one of the most sensitive technology in biochemical analysis. 24 With the advantages of easy synthesis and surface modification, 25 gold nanoparticles (AuNPs) have become mainstream, 26 which can be directly imaged and counted in dark-field microscopic (DFM) due to the local surface plasmon resonance (LSPR). At present, there are many detection methods combining DFM with automatic counting of AuNPs to establish a quantitative relationship between the counts of AuNPs and target molecules, 27 such as proteins, 28,29 DNA, 24,30 miRNA, 31 and enzyme. 32,33 For telomerase activity, although there were some reports utilizing DFM to quantify telomerase activity based on the peak shifts of LSPR Rayleigh scattering, [34][35][36] the visual automatic counting DFM strategy is also essential, with simplicity and intuitiveness. Therefore, it is of great significance to set up an easy-to-design biosensor for telomerase activity quantification using visual counting DFM.
Herein, we developed a convenient visualized automatic AuNPs counting strategy based on DFM for singlecell level telomerase activity detection. This strategy with excellent sensitivity and specificity was effective in inhibitor screening and different cells activity testing. It is worth mentioning that this strategy can clearly identify bladder cancer among various urinary diseases, suggesting its potential as a powerful tool to early and noninvasive diagnosis of bladder cancer.

Apparatus
The specific model of the DFM was Olympus BX53. The color charge-coupled device model was Olympus DP73 with 1600 × 1200 pixels. Olympus Cellsence software (Olympus, Japan) was used to analyze the DFM images, which were obtained with a 40× objective and 200 ms exposure. Matlab software was used to count the AuNPs with the bright spots. Ultrapure water was prepared with Milli-Q water purification system (≥18.2 MΩ, Millipore, USA).

Cell culture
HeLa cells (human cervical cancer cells), A431 cells (human skin squamous cells), MCF-7 cells (human breast cancer cells), and HepG-2 cells (human liver cancer cells) were cultured in DMEM medium. A549 cells (human lung cancer cells) and HL-7702 cells (normal human liver cells) were cultivated in F-12K medium or 1640 medium, respectively. These cells were cultured in the air at 37 • C containing 5% CO 2 . Specially, MDA-MB-468 cells (human breast cancer cells) were cultured in L-15 medium without CO 2 . All cells were collected for telomerase extracts during the exponential growth phase.

Extraction of telomerase
We first quantified the number of Hela cells using cell counter (JIMBIO, Jiangsu, China) ( Figure S1). Note that 1 × 10 6 cancer or normal cells were collected after cell digestion and washed with 1× phosphate buffer saline (PBS) twice (pH 7.2-7.4, 8.24 mM Na 2 HPO 4 , 2.67 mM KCl, 1.76 mM NaH 2 PO 4 , 136.89 mM NaCl). One milliliter CHAPS lysis buffer (10 mM Tris-HCl, pH 7.5, 1 mM MgCl 2 , 1 mM EGTA, 5 mM C 2 H 6 OS, 0.1 mM PMSF, 10% v/v glycerol, 0.5% m/v CHAPS) was added and incubated on ice for over 30 min to fully lyse the cells to obtain telomerase. The supernatant containing telomerase was transferred to a new RNA-free EP tube and stored at -80 • C for later use after centrifugation for 20 min at 12,000 rpm at 4 • C.
The collection of telomerase in urine sample was based on the previous works. [37][38][39] Two hundred milliliter fresh morning urine samples were collected from various patients by centrifugation at 1000 rpm at 4 • C for 10 min. After washed by 1 × PBS and centrifuged at 1800 rpm at 4 • C for 5 min, the urine precipitate was rinsed in 2 mL CHAPS lysis buffer at 4 • C for 30 min. The following processes were the same as above method.

Modification of the glass slide
The process of slide modification was finished referring to our earlier works. 33 The prewashed slides were placed in piranha solution (H 2 SO 4 (98.3%) : After rinsing with ultrapure water, the glass slides were completely immersed in 10% (v/v) APTES ethanol solution for another 6 h. Finally, the glass slides were washed with ultrapure water and dried in a vacuum oven for later use.

Preparation of AuNPs-SH-cDNA complexes
The process of AuNPs' modification of DNA referred to our previous works. 32,40 In short, TCEP buffer was firstly mixed with cDNA in dark for 1 h to fully activate the thiol-modified DNA. Then a certain amount of AuNPs and excessive activated cDNA were mixed for 16 h at 4 • C. 100 μL of 1 M NaCl solution was continuously added to the reaction buffer. Then, the AuNPs-SH-cDNA solution was enriched at 7000 rpm centrifugation and resuspended in PBS three times. Finally, the solution was stored at 4 • C for later use.

Preparation of MBs-TS complexes
In brief, streptavidin-coated MBs were rinsed thrice with 1 × B&W buffer (

Visualized detection of telomerase activity
The telomerase elongation reaction was completed first. To achieve a clearer background of AuNPs, it is necessary to separate MBs and AuNPs-SH-cDNA. The probe was rinsed once with ultrapure water including 0.02% (v/v) Tween-20 and resuspended twice in ultrapure water, then resuspended in 20 μL NaOH (0.15 M, pH = 12.1) for 30 min to obtain DNA double-stranded unwinding. 41,42 After magnetization, the hybrid was separated and 3 μL supernatant was titrated on the slides for DFM imaging.
For inhibitor screening, the concentration gradients of BIBR 1532 or curcumin were mixed with the extracts of 1000 HeLa cells, 43,44 respectively. CHAPS buffer and hightemperature inactivated extracts of HeLa cells were used as blank controls. 15,45 For detection in serum, MBs-TS, dNTPs and telomerase from cells lysis were mixed in TRAP buffer containing 10% (v/v) normal human serum. And for detection in urine samples, MBs-TS, dNTPs and telomerase from urine lysis were mixed in TRAP buffer. After reaction for 4 h at 37 • C, the visualized detection was carried out as the above-described method.

Proof of the strategy
Scheme 1 shows the visualized automatic particle counting strategy to detect telomerase activity. MBs were modified with bio-TS (MBs-TS) through a streptavidin-biotin interaction. The presence of telomerase and dNTPs allowed the TS primer to be repeatedly amplified with (TTAGGG) n to produce the telomerase extension product (MBs-TEP). Modification of 5'-thiolated cDNA to the surface of AuNPs formed AuNPs-SH-cDNA complexes by activating the thiol group and salt aging. Therefore, the elongated TEP sequence could hybridize with multiple AuNPs-SH-cDNA complexes based on complementary base pairing, constructing MBs-TEP AuNPs-SH-cDNA hybrid. The free AuNPs-SH-cDNA complexes were removed by magnetic separation to reduce the false-positive signal. The dropped AuNPs from MBs by NaOH solution (0.15 M, pH = 12.1) were collected for DFM imaging, in which bright dots were automatically counted using Matlab software. The telomerase activity is quantitatively reflected by the counts of AuNPs, achieving the goal of high sensitivity and visualized detection. On the contrary, in the absence of telomerase, the repeat sequence (TTAGGG)n couldn't extend at the 3' end of the TS primer. Then, almost no AuNPs will be captured for visual imaging.
To assess the ability of the DFM system for detecting telomerase activity, the counts of AuNPs under four conditions were tested to further eliminate false-positive signals. Samples were prepared under different condi-tions ( Figure 1). MBs-TS was incubated with telomerase extract of 1000 Hela cells (a), extract of 1000 HL-7702 cells (b), telomerase extract of Hela cells without dNTPs (c), or telomerase extract of Hela cells heated at 95 • C for 10 min (d), respectively. When all components existed (a), the DFM image exhibited visualized dense bright spots, with a high count of AuNPs reaching almost 670. Most tumors (85%) highly express telomerase activity, while the telomerase activity of normal somatic cells is inhibited resulting in the low signal in sample b. In comparison, the counts of AuNPs remained low for all other samples, demonstrating that the counts of AuNPs reliably reflect the activity of telomerase. The results strongly indicate that it is feasible to quantify telomerase activity by SPD based on DFM.

Optimization of experimental conditions
To achieve the maximum bright spots in the DFM field, six conditions affecting the signal value were optimized, including exposure time of DFM, the number of the repeat sequence (CCCTAA), TS primer concentration, concentration of AuNPs, telomerase extension time and hybridization time. Since the exposure time of DFM directly affects the signal-to-noise ratio (S/N). The optimal range was 100-700 ms. As shown in Figure S2A, the S/N reached a maximum value of 200 ms. The increase in exposure time led to an increase in the background signal and ultimately a decrease in S/N. Therefore, 200 ms was selected as the optimal exposure time for DFM.
Because AuNPs-SH-cDNA complexes are complementary with the repeat (TTAGGG) n of MBs-TEP, the number of repeat sequences (CCCTAA) directly affects the hybridization between AuNPs-SH-cDNA complexes and MBs-TEP. If the length of the cDNA sequence is too short, it will cause AuNPs aggregation during the salt aging process. In Figure S2B, it could be found that the counts of AuNPs were greatest when n reached 5. Therefore, (CCCTAA) 5 was used as the best experimental condition.
In the presence of telomerase, MBs-TS undergoes an extension reaction. Too much TS primer will affect the efficiency of telomerase extension, and too little will have a negative impact on the signal value. The amount of TS primer attached to MBs plays a crucial role in the effect of telomerase extension. As shown in Figure S2C, the counts of AuNPs reached a peak when the TS primer concentration was 100 nM, which was decided as the optimal TS concentration.
It is obvious that the concentration of AuNPs directly affects the counts of bright dots for DFM imaging. As shown in Figure S2D, the optimal concentration of AuNPs was determined to be 20 pM. To achieve a time-saving and efficient system, the optimum extension time of telomerase and DNA hybridization time were tested. Figure S2E,F revealed the best extension and hybridization times were 4 and 1.5 h, respectively.

Determination of telomerase in HeLa cells
As an important biomarker of bladder cancer, the sensitivity is critical. Herein, the telomerase extracts from 1-1000 HeLa cells were evaluated under the optimal experimental conditions. Figure 2 clearly exhibited that the counts of AuNPs corresponded to different numbers of cells. According to the previous studies, 15 These results suggested that the limit of detection (LOD) for telomerase was down to 1 HeLa cell based on the 3σ/k rule, which was comparable to or lower than that reported in other works (Table S2). These results fully prove that SPD based on DFM is an effective platform for the visualized measurement of telomerase activity at the single-cell level.

Specificity of DFM assay
To test the specificity, the telomerase extracts from several kinds of cell lines were studied, involving HeLa, MCF-7, MDA-MB-468, A549, HepG2, A431, and HL-7702 cells. Figure 3A shows the counts of AuNPs in 1000 cells of each cell line, in which HeLa cells exhibited the highest counts of AuNPs. A normal human liver cell (HL-7702) line was used as a negative control, and the signal value was equivalent to the blank, indicating that the telomerase activity is very low in normal cells. These results are comparable to those of previous reports (Table S3). Subsequently, enzyme specificity was investigated using Exonuclease I/III (Exo I/III), Bacteriophage λ exonuclease (λ exo ), T7 Endonuclease I (T7), Taq DNA Polymerase (Taq), and Hap II. As shown in Figure 3B, except for telomerase specifically elongating TS primer, the other six enzymes had similar counts of AuNPs with the blank, thus suggesting that this strategy is highly selective for telomerase.

Detection of inhibition efficiency
To further explore the biological application of this strategy, two telomerase inhibitors, BIBR 1532 and curcumin, were chosen to evaluate the inhibitor screening ability.
Referring to the previous research, 38,44,46 the inhibition (%) was calculated as follows: where Au c,0 , Au c,1 , and Au c,2 are the counts of AuNPs in MBs-TS with CHAPS, MBs-TS with 1000 HeLa cells extracts and inhibitors, and MBs-TS with 1000 HeLa cells extracts, respectively. Figure 4 shows the curves of inhibition rate with the increase of BIBR 1532 and curcumin concentration, the inhibition (%) gradually increased and finally leveled off. According to the results, the 50% inhibiting concentration (IC 50 ) values of BIBR 1532 and curcumin were calculated to be 269 ± 8 nM and 12 ± 0.2 μM, respectively, which are equivalent to the previously reported experimental results (Table S4). These findings confirm that the developed biosensor has the potential to screen telomerase inhibitors.

Detection of telomerase activity in human serum and urine samples
To analyze the anti-interference of the system, the telomerase activity in human serum was tested. Different amount of HeLa cell lysates was diluted with CHAPS buffer and added to complex samples containing 10% human serum. As shown in Figure S3, the counts of AuNPs detected in serum and cell lysate were similar, further demonstrating the strong anti-interference capability of this strategy in complicated biological samples.
Large numbers of bladder cancer cells are present in patients' urine, 47 which is currently the second most common malignant tumor of the genitourinary system. 37 At present, the mainstream methods for early-stage detection of bladder cancer are clinical cytology and regular cystoscopy. 48 However, these diagnostic tools still have disadvantages, such as low sensitivity and detrimental invasiveness. 49 Therefore, the development of a noninvasive and sensitive monitoring platform is essential. Since the high expression of telomerase in urine is closely related to bladder cancer, 39 we explored whether this method could achieve noninvasive testing of bladder cancer from patients' urine.
Urine samples from 18 patients (14 with bladder cancer and 4 with other diseases, i.e., urinary incontinence, inguinal hernia, hydrocele and varicocele), were collected from Shenzhen People's Hospital. A previously reported method was used to extract telomerase from the urine samples. 38 As can be seen from Figure 5 and Table S5, the number of AuNPs detected in most samples from bladder cancer patients was more than 100, while detected in the samples of 4 nontumor patients was less than 50, forming a sharp contrast. These results indicate that this assay can identify bladder cancer among urinary diseases, thus demonstrating its potential in noninvasive early screening diagnosis of bladder cancer.

CONCLUSION
In summary, we developed a visualized automatic particle counting strategy for single-cell level telomerase activity quantification. The proposed assay exhibits several obvious advantages. Firstly, due to the sensitive LSPR effect and automatic counting of AuNPs, telomerase activity could be visualized using DFM imaging, achieving a LOD as low as 1 HeLa cell. Secondly, this strategy could discriminate differences of telomerase activity in several kinds of cell lines and successfully screen the common telomerase inhibitors (BIBR1532 and curcumin). Furthermore, this method possesses strong anti-interference in complex blood samples. More importantly, visibly identify bladder cancer among various urinary diseases indicates its further potential applications in early and noninvasive diagnosis for bladder cancer.

A C K N O W L E D G M E N T S
The authors acknowledge the financial support from the State Key Laboratory of Chemical Oncogenomics, the National Key R&D Program of China, Synthetic Biology Research (grant number: 2019YFA0905900) and Shenzhen Municipal government (grant number: JCYJ20220530142812029).

C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare no conflicts of interest.