Introduction

Being rapid, inexpensive, easy-to-manufacture and user-friendly, lateral flow assays (LFAs) have usually been the first technology to be considered in a range of fields, including medical diagnostics, bedside analysis, food safety, and environmental safety for biochemical analytes, such as proteins, glycolipids, lipids, and nucleic acids [1,2,3,4,5]. Microfluidics technology brings great potential by providing integration, high-throughput, fast analysis time, portability, and small reagent volume. Taking advantage of the microfluidics, there have been a lot of efforts to integrated lateral flow assays (LAFs) with the microfluidic device to realize integrated assay platform with multifunction, for example, an integrated rotary microfluidic system with DNA extraction, loop-mediated isothermal amplification, and lateral flow strip based detection for point-of-care pathogen diagnostics [6]. At the same time, various microfluidic elements such as hydraulic resistors [7, 8] valves [9], and advanced features for liquid flow control [10] were also introduced into LAFs. However, due to the limit in the space of LFAs and microfluidic chip, the outstanding issues on detection sensitivity remain unsettled, which is an urgent problem in their practical application. Especially in the case of real sample analysis, such as human body fluids, the detection of low-abundance proteins in serum still face a great challenge. Non-integrated sample preparation steps were normally needed, which comprised the efficiency and equipment free advantage owned by LFAs. There were extensive researches in the enhancement of sensitivity without the special sample preconcentration beforehand [11,12,13], such as the improvement of capture reagent immobilization [14], the transport performance [15,16,17,18,19,20,21,22,23], the novel labels [24,25,26,27], signal amplification [28, 29] and readers [1, 13, 30,31,32]. Different from the focuses on the existing LFAs components using complex enzyme/chemical reaction-based signal amplification as literature, in this paper high sensitive microfluidic lateral flow visual assay of protein was realized with the aid of automatic roller to achieve automatic on-stripe repetitive injection and then multiple concentration.

Materials and Methods

Materials and Reagents

Human chorionic gonadotropin (HCG) was ordered from Shanghai Linc-Bio Science Co. LTD. Water was purified and deionized with a Milli-Q system (Millipore, America). The micro peristaltic pump was ordered from LongerPump. AB adhesive was ordered from Ausbond, USA. The double-side adhesive was ordered from 3 M China co. LTD. PVC board purchased from Shanghai Jiening Biological co. LTD. Risym micro gear motor purchased from Shenzhen Kebiwei Electronics co. LTD. Photos were taken by a cellphone ordered from Xiaomi Corporation.

The Structure of the Microfluidic Device

Figure 1a is the 3D design drawing of the entire device. As shown in Fig. 1b, the interior structure consists of two parts: microfluidic section and control section. There are two rollers on the chip, just like the ones in a tape, roller A is the driving one, and roller B is the driven one. The two rollers are connected by cellophane tape, which is made by transparent adhesive tape. All the pads are stuck on the tape. In the forward direction of the belt, the colloidal gold pad is put in the front and followed by an absorption pad with a distance of 11 mm. One colloidal gold pad and one absorption pad is used as a whole during the concentration process. The distance between the gold pad and one absorption pad is 20 mm without mutual interference. The strip without the colloidal gold pad and absorption pad are stuck to a baffle that can be moved powered by an electric machine. The control section consists of two electric machines: one is used to move the strip back and forth and the other one is in charge of the rolling of B. Two buttons are control switches for the electric machines.

Fig. 1
figure 1

a Design drawing of the device; b the structure of the microfluidic machine

Full size image

Results and Discussion

The Structure and Working Principle of the Concentration Integrated Microfluidic LFA

In common LFAs, after adding the solution to the sample pad, the sample migrates with buffer through the strip by capillary forces and dissolves the report reagent in the conjugation pad. The conjugates flow forward in the porous membrane until they are captured by Abs immobilized on the test line (TL), giving a detectable signal. Redundant conjugates in the remaining solution flow through the membrane are attached to the control line (CL), showing the successful running of the test. Since the result of an LFA is related to the optical signal generated at the test line, the more amount of conjugates captured on the test line, the better detection of trace analyte will be. Although there is a comparatively excessive amount of Abs immobilized on the detection line, however, the bottleneck problem lies in the limited amount of conjugates that can be adsorbed and transferred by the sample pad and conjugated pad. After all, one piece of sample pad and conjugated pad with limited length cannot load and transfer much sample.

In this paper, a novel strategy by repeatedly replenishing the colloidal gold pad to realize on-stripe competitive injection and then multiple concentration was adopted with the aid of mechanic roller. The enrichment mechanism is shown in Fig. 2. With the accurate control of the step length and the race of rotation, a consistent length of the pad was assembled each time, guaranteeing the linearity between the amount of enrichment and number of rotation. The sample is added onto the sample pad through the injection pump for precise volume control.

Fig. 2
figure 2

The enrichment mechanism of our microfluidic lateral flow assay

Full size image

The Optimization of Sample Amount

If the injection amount is too small, the amount of sample is not enough to in the detection area to give a detectable signal. If the sample amount is too large, an excessive amount of sample will enter the detection area without binding with colloidal gold and Abs on the test strip in the detection site, giving a lower signal than the real concentration of the sample. Therefore, a series volume of 10, 15, 20, and 25 μL of pure water were applied, respectively, to study the proper sample amount of test strip. It was shown that 10 μL of HCG could not make its way to the test strip, while a sample of 25 μL exceeded the water adsorption capacity of the test strip. Therefore, an injection in 15–20 μL was proper in this experiment, and 15 μL was chosen in the following experiment.

The Sample Analysis with the Proposed Device

The detection limit of commercial LAFs using the traditional method is 7.85 ng/mL. Using the proposed device in this paper, if the sample with a concentration of 7.85 ng/mL was five times diluted to 1.57 ng/mL, the sample could be detected. As shown in Fig. 4, in the first round of analysis of HCG sample in the concentration of 1.57 ng/mL, the text-line (T-line) was almost invisible to the naked eye. After the five rounds of enrichment, red bands were seen with the naked eye. According to the detection limit of commercial LAFs, five-time enrichment of sample with the initial concentration of 1.57 ng/mL after five rounds of rotation had been realized.

The Study of the Linearity Between the Signal Intensity and Enrichment Times

The linearity between the signal intensity and enrichment times with the proposed device was further studied with the samples in the concentrations of 1.57 ng/mL and lower concentration of 1.26 ng/mL. The LAF images were analyzed with ImageJ software. The corresponding gray value of each band was given in Tables 1 and 2. As can be more easily observed in Fig. 3, there is a linear relationship between T-line and the control line (C-line), respectively, with the number of enrichment when the rounds of enrichment are smaller than five times. The R2 value of 0.9468 and 0.9257 for the samples of 1.57 and 1.26 ng/mL, respectively, indicate that the enrichment effect was linearly and consistently correlated with the number of enrichment. The good linearity also indicated the super reproducibility of this mechanic device. When the enrichment times for the sample of 1.57 ng/mL were more than five times, there was no further enrichment due to the adsorption saturation of antibody on the C line. And it can also be observed, the range of linearity in the sample of 1.57 ng/mL was wider than the sample of 1.26 ng/mL. The detection in the lower concentration is more challenging. When the sample of 1.26 ng/mL was analyzed for several times, the approximate grey value of the C line obtained after five rounds of enrichment indicates the good reproducibility of the device. At the same time, the fixed amount of sample loss reflected also indicates that there is systemic error which makes it easier for future improvement.

Table 1 Gray analysis of the detection image of HCG (1.57 ng/mL)
Full size table
Table 2 Gray analysis of the detection image of HCG (1.26 ng/mL)
Full size table
Fig. 3
figure 3

a1 LAF images of HCG in the concentration of 1.57 ng/mL; a2 the linearity between the gray value of bands and the times for enrichment; b1 LAF images of HCG in the concentration of 1.26 ng/mL; b2 the linearity between the gray value of bands and the times for enrichment

Full size image

The Study of the Relationship Between the Predicted Accumulation Quantity of HCG and the Gray Value of Bands

The results of multiple HCG enrichment integrated analysis were further compared between the different samples in the concentration of 1.26 and 1.57 ng/mL, respectively. According to the initial concentration of samples and times of concentration, a predicted accumulation quantity of HCG after five rounds of enrichment and the gray value of each band in the different concentration rounds are given together in Table 3. As can be observed in Fig. 4, the real grey value reflecting the actual amount of HCG is in the linearity with the predicted accumulation quantity of HCG, demonstrating again the good linear enrichment performance of the system and feasibility of semi-quantitative detection after further optimization.

Table 3 The comparison of HCG enrichment with the concentration of 1.26 and 1.57 ng/mL
Full size table
Fig. 4
figure 4

The relationship between the predicted accumulation quantity of HCG and the gray value of bands

Full size image

The Investigation of Detection Limit and Performance Comparison with Literature

Based on the above study, a series of samples in the concentration of 1.26, 0.94, and 0.63 ng/mL, respectively, lower than 1.57 ng/mL were studied. It was found that the sample in the concentration of 1.26 ng/mL could be detected. While there was not an obvious band in the analysis of the sample in the concentration of 0.94 ng/mL. And then, the detection limit of this proposed method for HCG was 1.26 ng/mL. The performance of this method is compared with LAFs for protein detection in the latest literature. As shown in Table 4, without the requirement of an extra instrument such as surface-enhanced resonance Raman scattering [5, 10], the detection limit obtained through the integration of ordinary LAFs with the simple microfluidic device in this paper is comparable with these method using special nanoparticles [6, 7, 11, 12].

Table 4 The summary of the LAFs detection limit for protein in literature
Full size table

Conclusions

The main focus of this technology nowadays is how to substantially improve LFA sensitivity without sacrificing its advantages. The existing improvements in LFA could be categorized into reaction, transport and signals [1]. In this paper, different from the existing way of improvement, with the advantage of chip integration, the sample pad was repetitively replaced to realize multi-enrichment of analyte onto the detection limit. This device was applied for enrichment detection of HCG. The minimum concentration of the sample initially determined by this device is 1.26 ng/mL, which is more than six times lower than the detection limit of the classic strip method (7.86 ng/mL) and also comparable with literatures using special nanoparticles. Together with good sensitivity, the linear relationship between the signal intensity vs the times of concentration and the predicted accumulation quantity of HCG vs the gray value of bands, lay a foundation for quantitative and semi-quantitative detection of low-abundant targets in real sample samples. This cost-effective, reusable, naked eye or smartphone readable, and robust device holds great potential as a candidate for medical diagnosis, home testing, food and environment monition.