**3. Results**

We have attempted to analyze the angiogenin levels in burn blister fluids to assess burn depth using P-ELISA. To determine optimal primary antibody concentration, we examined colorimetric responses to different dosages of recombinant angiogenin antibody on test paper impregnated with blister fluid. The results showed that 1 μg /mL of anti-angiogenin antibody provided the best signal. (Figure S1). To achieve the best staining with minimal background interference, we tested three different concentrations (0.05, 0.04, and 0.033 μg/mL) of secondary antibody. After testing serial dilutions of secondary antibody, we found that 0.033 μg/mL demonstrated the highest signal-to-background ratio

for our primary antibody (Figure S2). We washed the test paper twice, first with 5 μL of TBST solution and then again with 10 μL of TBST solution, to remove unbound material. To determine optimal BSA concentration during the blocking step, we coated filter paper with di fferent concentrations (1%, 0.5%, 0.1%) of blocking solution and compared reaction results to blister fluid harvested from DPTB patients and control blood serum. Mirroring the expected conventional plate ELISA results, our data showed that 1% of BSA displayed the best blocking e ffect for burn blister fluid when compared to our healthy human blood serum control (Figure S3). In addition to optimizing blocking, determining the best washing volume is essential for removing unbound reagents and reducing background signal. Unlike the conventional plate ELISA wash volume of 300 μL, P-ELISA requires only minute wash volumes. We tested the e fficacy of three di fferent wash solution amounts, 15 μL, 10 μL, and 5 μL, and found that three separate washes (the typical number of washes for conventional ELISA) with 5 μL removed all unbound nonspecific material (Figure S4). While the typical number of washes with conventional ELISA is three, we discovered that two cycles, one with 5 μL of wash bu ffer and one with 10 μL of wash bu ffer, provided the lowest background and strongest signal strength for P-ELISA (Figure S5).

Paper-based diagnostics provide a low-cost and easy-to-handle approach for studying a variety of target factors. Here, we demonstrated this approach by adding burn blister fluids and select reagents to test zones to detect angiogenin levels. As shown in Figure 1, burn fluid was added by hand, reagen<sup>t</sup> was spotted onto each test zone, and the reaction was allowed to proceed for several minutes. Primary and secondary antibody were then added and optimized washing processes carried out to remove unbound antibody from the paper. Colorimetric responses were digitally recorded before and after color development. By analyzing the change in grayscale intensity of before and after images, we could determine the amount of angiogenin in burn blister fluids.

**Figure 1.** Procedure of paper-based ELISA (P-ELISA) for angiogenin analysis in burn blister fluid. Three microliters of burn fluid was loaded onto the test zone, and the test paper was incubated for 10 min. Three microliters of 1% BSA was then added for blocking, and the test paper was incubated for 10 min. The first image was recorded after loading 3 μL primary antibody, and the test paper was incubated for 10 min. The test paper was washed with 5 μL and then 10 μL TBST (Tris-buffered saline + 0.05% Tween 20) before adding 3 μL of horseradish peroxidase (HRP)-conjugated secondary antibody. The final procedure was to wash again with TBST and add 3 μL of substrate solution. After 10 min of incubation, we recorded the second image.

Relative mean intensity for angiogenin from SPTB burn blister fluids (Figure 2a) and DPTB burn blister fluids (Figure 2b) were tested and compared. In keeping with our previous study results, we found that angiogenin levels were significantly higher in DPTB fluids (4.2 ± 1.5, 95% confidence interval, 1.3–6.5, *n* = 6) than SPTB fluids (1.4 ± 0.3, 95% confidence interval, 0–4.0, *n* = 6) and healthy blood serum samples (0.9 ± 0.3, 95% confidence interval, 0.4–1.2, N = 7, *p* < 0.01) as measured by P-ELISA (Figure 2c). This demonstrates the reliability of P-ELISA to determine burn severity by evaluating angiogenin levels in burn blister fluids. Burn fluid samples have previously been tested via conventional ELISA plates for angiogenin and VEGF. Angiogenin is a downstream molecule of VEGF-regulated angiogenesis [23]. Observation of the role of VEGF in burn wound determination is interesting. Although no significant di fferences in VEGF were observed (SPTB: 31.5 ± 3.6 ng/mL, 95% confidence interval, 20.0–42.9 ng/mL, DPTB: 47.1 ± 10.9 ng/mL, 95% confidence interval, 16.8–77.5 ng/mL), there was a trend for higher angiogenin concentration in DPTB fluids compared with SPTB fluids as determined by conventional plate ELISA analysis (SPTB: 119.2 ± 28.4 ng/mL, 95% confidence interval, 28.8–209.6 ng/mL, DPTB: 331.5 ± 81.0 ng/mL, 95% confidence interval, 123.3–539.7 ng/mL, Figure 3, Table S1). No significant di fferences in VEGF level were observed between two di fferent burn fluids, as shown in Figure 3, which was consistent with our previous study and another study indicating that VEGF was not responsible for di fferentiation of circulating angiogenic cells [9] or tumor growth and angiogenesis [23]. The consistent comparability of our P-ELISA results with conventional plate ELISA results was further supported by a correlation test using Pearson's correlation analysis. A moderate positive correlation (rho = 0.5906, *p* = 0.0722) between the titer of the plate ELISA and the relative intensity of P-ELISA was observed and is displayed in Figure 4.

**Figure 2.** Analysis of angiogenin levels from superficial partial-thickness burn (SPTB) and deep partial-thickness burn (DPTB) blister fluids using P-ELISA. (**<sup>a</sup>**,**b**) Clinical pictures of superficial partial-thickness burn (SPTB, **a**) and deep partial-thickness burn (DPTB, **b**) wounds. (**c**) Comparison of angiogenin levels from two di fferent burn fluids and healthy human blood serum as the control (*n* = 6 in SPTB and DPTB, *n* = 7 in control, mean ± S.D, \*\* *p* < 0.01).

**Figure 3.** Analysis of angiogenin and VEGF (Vascular endothelial growth factor) concentrations in SPTB and DPTB blister fluids with conventional plate ELISA. A trend toward higher angiogenin concentration in DPTB fluids was detected, compared to SPTB (*n* = 4 in SPTB, *n* = 6 in DPTB, mean ± S.D.; *p* = 0.07). No significant difference in VEGF levels was observed between two different blister fluids (*n* = 4 in SPTB, *n* = 5 in DPTB, mean ± S.D.; *p* = 0.26).

**Figure 4.** Correlation of the angiogenin detection between conventional plate ELISA and paper-based ELISA in burn blister fluids. The data show a moderate correlation between the results of P-ELISA and conventional plate ELISA (*r* = 0.5906, *p* = 0.0722).

In our laboratory, we have developed and explored the capacities for P-ELISA made of simple filter paper. To support the concept of P-ELISA material and process viability, we used patterned filter paper to collect directly absorbed blister fluid from human burn wounds (Supporting Movie) and had the test paper delivered to our laboratory on dry ice within 24 h, where we used it to successfully detect angiogenin levels via P-ELISA (Figure 5).

**Figure 5.** Clinical examination of angiogenin concentration by paper-based ELISA. Test paper was designed to absorb blister fluid from burn patients. Angiogenin signals were captured and analyzed. The mean intensity of detected angiogenin in burn blister fluid was significantly higher than that in normal serum control (mean ± S.D.; \*\*\*\* *p* < 0.0001, *n* = 6).
