Minimizing the variables of voiding spot assay for comparison between laboratories

Mice

Female C57BL/6J mice used in the laboratory at Beth Israel Deaconess Medical Center (BIDMC, Boston, MA, USA) and the laboratory at Southwest Medical University (SWMU, Luzhou, Sichuan, China) were respectively purchased from the Jackson Laboratory (Bar Harbor, ME, USA) and GemPharmatech Biotechnology (Chengdu, Sichuan, China). The mice were respectively labeled as “mice-BIDMC” and “mice-SWMU” in this study for convenience. All mice in two laboratories were maintained on a 12 h light and dark cycle at 25 °C and 20%–50% relative humidity. A maximum of five adult mice are group housed per cage with free access to standard laboratory food and water. Mice used in this study were females aged 12–16 weeks old. All procedures were approved by the BIDMC Institutional Animal Care and Use Committee (026-2016) or SWMU Animal Care and Use Committee (swmu20220301-008).

Types of daily housing cage

Four types of daily housing cages were used in this study. The first type of cage is used in the laboratory of BIDMC, and it is the polycarbonate AN75 mouse cage with stainless steel lid and microfilter top (Ancare, Bellmore, NY, USA). This cage is transparent with dimensions 28.50 cm (length) × 17.50 cm (width) × 12.00 cm (height). This type of cage is used for daily housing and voiding spots assay (VSA) in the laboratory at BIDMC. We call this cage “standard clear cage-BIDMC.” The second type of cage is the polycarbonate CP5 mouse cage with stainless steel lid and microfilter top (Huanqiu, Beijing, China). It is transparent with dimensions 26.50 cm (length) × 18.50 cm (width) × 13.00 cm (height). This cage has a similar appearance and size to AN75. Unless otherwise specified, this type of cage is used for daily housing and VSA in the laboratory at SWMU. We call this cage “standard clear cage-SWMU”. The third type of cage is the Polypropylene R3 cage (Zeya, Suzhou, Jiangsu, China). It is white opaque with dimensions 46.00 cm (length) × 30.00 cm (width) × 18.00 cm (height). This cage is the biggest cage for housing, we call this cage “big opaque cage”. The last type of cage is the Polypropylene M1 cage and it is white opaque (Zeya, Suzhou, Jiangsu, China), the dimensions of which is 29 cm (length) × 17.8 cm (width) × 16 cm (height). we call this cage “small opaque cage”.

Experiment design

To compare results of VSA in wildtype of C57BL/6J mice between laboratories from two geographical locations, VSA was performed in two laboratories, which are located in BIDMC and SWMU, respectively. The identical VSA procedures as previously described were used in both laboratories for VSA testing to minimize as many variables as possible (Rajandram et al., 2016; Chen et al., 2017). Individual mice (BIDMC: n = 49; SWMU: n = 61) were gently taken from their daily housing cage and put in a standard clean and empty mouse cage (BIDMC: AN75; SWMU: PC5) with filter paper (BIDMC: Blicks, Catalog no. 10422-1005; SWMU: Guangxiangyu, catalog no. PA-60-90A). The filter paper was cut to size and placed on the floor of the standard clear cage to absorb voided mouse urine. During the VSA testing, water was withheld and standard dry mouse chow was available. VSA testing was performed for a duration of 4 h between 9:00 am and 2:00 pm. Mice were returned to their home cages afterward and used to repeat the experiment the next day. The filter papers were collected for imaging after the urine dried.

To test whether the type of daily housing cage influences the voiding patterns in the standard cage, three types of cages were used, including the “big opaque cage”, “small opaque cage”, and “standard clear cage-SWMU”. The same cohort of 15 mice was used in this study and sequentially placed in three types of cages. Each type of cage allowed a maximum of five adult mice and housed them for 4 days before VSA. After 4 days of acclimation, mice were subjected to VSA testing in standard clear cage-SWMU for two consecutive days between 9:00 am and 2:00 pm.

To test how long the mice required to return to the baseline levels of voiding patterns after transport, we simulated the transport stress state as previously described (Wan et al., 2014; Shen et al., 2016). Five mice in one standard cage-SWMU were put on a shaker at 60rpm for 2 h at 25 °C. 20 mice were used in this experiment and water and standard dry mouse chow were provided for the duration. Mice were subjected to VSA testing for two consecutive days before transport treatment as the baseline levels. Afterward, mice were subjected to stimulation and then allowed to rest for 24 h (day 0). VSA testing was performed on days 1, 3, 5, and 7 between 9:00 am and 2:00 pm in standard cage-SWMU.

To test whether the voiding pattern changes with time during the daytime, we performed VSA with the same cohort of 15 mice in two periods with 4 days intervals. VSA was first performed in the morning between 9:00 am and 2:00 pm. After 4 days interval, the same cohort of mice was subjected to VSA testing in another period between 2:00 pm and 7:00 pm in the afternoon.

Image acquisition and analysis

Filters were imaged under ultraviolet light at 365 nm (BIDMC: Chromato-Vue C-75 system; UVP, CA, USA; SWMU: JY02S system; Junyi, Beijing, China) with an onboard Canon digital single-lens reflex camera (BIDMC: EOS Rebel T3; Canon, Tokyo, Japan; SWMU: EOS 90D; Canon, Tokyo, Japan). The same operator visually examined overlapping voiding spots in VSA images and manually separated them by outlining and copying and pasting each spot onto a nearby empty space using the Fiji version of ImageJ software (Image J version 2.9.0). All of the 20 VSA images were randomly selected and subject to analysis by MATLAB software (UrineQuant) and Fiji software (Image J version 2.9.0 with macro plugin) for the comparison in the quantification of VSA parameters between the two software. The rest of VSA images were analyzed by the Fiji software. The result table, which contains the area of each individual voiding spot and the number of spots, was imported into Excel for statistical analysis. A 1 mm² voiding spot represents 0.283 µl of urine on the filter paper from the laboratory of BIDMC (Keil et al., 2016). A volume-area standard curve (y = 0.245x –2.934, R² = 0.995) determined that a 1 mm² voiding spot on the paper from the laboratory of SWMU represents 0.245 µl of urine. Volume of voiding spot ≥ 20 µl was categorized as PVS and Volume of voiding spot <20 µl was categorized as SVS. All volume of <0.245 µl were excluded from the analysis.

Statistical analysis

All data are reported in the text as mean ± standard error unless otherwise indicated. The data in the figures are represented as boxes and whiskers. The centerline is the median of the data set, the box represented 75% of the data, and the bars indicate whiskers from minimum to maximum. Regression analysis was performed using the Passing Bablok method and Pearson’s correlation coefficients. Bland-Altman analysis was used to evaluate the agreement. Parametric or nonparametric tests were employed based on the results of the Kolmogorov-Smimov normality tests. Parametric data were tested using Student’s t-test or one-way ANOVA. Nonparametric data were tested using the Mann–Whitney U test or Kruskal-Wallis Test. P ≤ 0.05 was considered statistically significant. Statistical analyses were performed using GraphPad Prism 9 (GraphPad, San Diego, CA, USA) and MedCalc 21 (MedCalc Software, Ostend, Belgium).

Fiji and MATLAB are good agreement for the quantification of VSA

Laboratory-specific software was reported to induce variability in VSA quantification among laboratories (Wegner et al., 2018). Fiji and MATLAB software are widely used in the quantification of VSA, Fiji with its plugin Macros is freely available and automates VSA image batch processing and data extraction. UrineQuant is a module in MATLAB, specially designed by Harvard Imaging and Data Core for VSA analysis. It is unclear whether Fiji and MATLAB have a systemic bias in VSA quantification. Here, we compared the two software in terms of four VSA parameters, including total urine area, PVS area, PVS number, and SVS area for 20 images. For the quantification of total urine area, Passing Bablock regression analysis between the two software generated y = 0.99x + 14.78 (y = MATLAB; x = Fiji). 95% confidence interval for the intercept was -24.19 to 39.65 mm² and the slope was calculated to be 0.97 to 1.03. Comparison analysis displayed a significant correlation between the two software, with coefficient r = 0.99, p < 0.01 (Fig. 1C). Bland-Altman analysis showed a mean difference of −11.10 mm² without statistical significance and that 95% limits of agreement were from −65.24 to 43.03 mm² (Fig. 1D). For the quantification of the PVS size, Passing Bablock regression equation between two software was: y = 1.00 × –1.66. 95% confidence interval for intercept and slop were −3.53 to 0.38 mm² and 0.99 to 1.00. Correlation coefficient was r = 0.99, p < 0.01 (Fig. 1E). Bland-Altman analysis showed a small mean difference level of −0.31 mm² without statistical significance and 95% limits of agreement from −13.60 to 12.98 mm² (Fig. 1F). Fiji and MATLAB are perfect fit and well agree for PVS number determination, displaying regression equation: y = 1.00 × (Figs. 1G–1H). For the quantification of the SVS area, Passing Bablock regression analysis generated equation: y = 0.93 ×+ 0.95. 95% confidence interval for intercept was −0.57 to 1.95 mm² and for the slope was 0.85 to 0.97. Comparison analysis showed coefficient r = 0.98, p < 0.01 (Fig. 1I). Bland-Altman analysis showed that a mean difference was small with a value of −0.62 mm² and statistically not significant and that 95% limits of agreement were from −4.18 to 2.93 mm² (Fig. 1J).

Types of daily housing cages do not alter the voiding patterns

Voiding patterns are easily affected by housing and husbandry conditions, such as mouse housing density (Chen et al., 2017; Keil et al., 2016). Types of housing cages are directly related to mouse housing density, the variety of which adopted among laboratories may complicate cross-study comparisons, we tested whether the types of daily housing cages influence the mice voiding patterns. The same cohort of 15 female mice was sequentially housed in standard clear cages, big opaque cages, or small opaque cages, for 4 days, and then was subjected to VSA in standard cages (Figs. 2A–2C). Surprisingly, mice from the standard clear cages, big opaque cage, and small opaque cage had an average total volume of 732.60 ± 48.21 µl, 636.40 ± 42.26 µl, and 747.50 ± 47.66 µl, showing no significant differences (Fig. 2D). A few differences were observed in the frequency of the total urine volume but over 70% of mice from three cages produced a total volume of 300–900 µl, displaying some similarities (Fig. 2E). In addition, mice from three cages produced almost identical total voiding numbers, which were 9.13 ± 0.50, 9.90 ± 0.87, and 10.13 ±  0.67, and most of the mice voided 6–15 times (Figs. 2F and 2G). Unexpectedly, types of daily housing cages did not induce PVS volume and number changes, As shown in Figs. 2H and 2J, mice from three cages perceptively produced 5.63 ± 0.36, 5.23 ± 0.33, and 5.20 ± 0.25 PVS, corresponding with the average PVS volume of 130.10 ± 6.40 µl, 119.30 ± 6.67 µl, and 142.40 ± 7.76 µl. The mice from three cages had similar total PVS volume, which were 726.80 ± 48.67 µl, 626.80 ± 42.09 µl, and 740.60 ± 48.05 µl, respectively (Fig. 2L). Furthermore, the frequency of PVS volume and number displayed almost identical among the mice from three cages (Figs. 2I and 2K).

Voiding patterns are different during the morning vs afternoon

VSA testing is usually performed in the morning or afternoon for a 4-hour duration. The time of day when voiding assays are conducted among laboratories was reported to be a variable, we next compared VSA parameters generated by mice during the morning vs afternoon. The same cohort of 15 female mice was tested either in the morning or afternoon (Figs. 3A and 3B). Mice in the morning produced total urine of 717.00 ± 49.33 µl in 4 h, which was more than 499.40 ± 28.93 µl produced in the afternoon (Fig. 3C). Approximately 60% of mice in the afternoon produced total urine of 300–500 µl and 40% of mice voided more than 500 µl, while 27% of mice in the morning produced total urine of 300–500 µl and 73% of mice voided exceeding 500 µl, indicating that the total urine volume differed with the periods (Fig. 3D). Interestingly, mice in the two periods voided similar times, which were 9.13 ± 0.50 and 7.90 ± 0.57, respectively, and 57%–60% of the mice in the two periods voided 6–10 times (Figs. 3E and 3F). However, mice in the morning produced a greater number of PVS and a smaller PVS volume compared to the mice in the afternoon (5.70 ± 0.33 PVS with 124.70 ± 6.04 µl vs 3.10 ± 0.22 PVS with 159.00 ± 7.89 µl) (Figs. 3G and 3I), indicating that PVS parameters changed with the time of day. The conclusion was further supported by the frequency of PVS volume and number. 85% of PVS produced by the mice in the morning was 20–200 µl and 15% of PVS was more than 200 µl, while 72% of PVS in the afternoon was 20–200 µl and 28% of PVS exceeded 200 µl. In addition, only 27% of mice in the morning produced PVS numbers less than 4, and 73% of mice produced PVS numbers more than 5, while 83% of mice in the afternoon had PVS numbers less than 4 and 17% of mice had the PVS numbers more than 5 (Figs. 3H and 3J). Furthermore, mice in the morning generated a total PVS volume (711.00 ± 49.84 µl) more than mice in the afternoon (499.40 ± 28.93 µl) (Fig. 3K).

Voiding patterns recovered within three days after transportation

For research purposes, mice are often transported between institutions, which may elicit stress to influence voiding patterns. Unexpectedly, mimicking transportation did not change the total urine volume (Figs. 4A–4E). Mice after transportation on days 1, 3, 5, and 7, respectively generated a total urine volume of 674.7 ±36.61µl, 690.40 ± 40.73 µl, 634.00 ± 40.38 µl, and 605.50 ± 35.63 µl, all of which were similar to baseline levels (675.1 ± 55.19 µl) (Fig. 4F). However, urine more than 1,100 µl was only observed in 5% of mice before transportation, indicating that transportation reduced the total urine volume in some mice (Fig. 4G). Mice originally voided 12.45 ± 1.41 times during 4 h period, which significantly increased to 23.85 ± 4.45 times on day 1 after transportation, then returned and maintained to baseline levels on days 3, 5, and 7 (11.20 ± 1.51, 12.35 ±  1.30 and 11.65 ± 1.33) (Fig. 4H). Mice before transportation produced 4.70 ± 0.32 PVS with an average volume of 122.60 ± 6.58 µl in 4 h of testing. After transportation, the mice produced 7.45 ± 0.50 PVS with an average volume of 83.82 ± 4.62 µl on day 1. The PVS volume and number returned to baseline levels on day 3 (5.50 ±0.55 PVS with 110.90 ± 6.37 µl). No significant changes in PVS parameters were observed on days 5 or 7 (day 5: 5.35 ± 0.41 PVS with 110.20 ± 5.37 µl; day 7: 5.50 ± 0.40 PVS with 108.10 ± 5.92 µl) (Figs. 4J and 4L). In addition, the frequency of PVS volume and number displayed the marked differences on day 1. A total of 60% of mice on day 1 generated PVS numbers more than 6 and 75% of them were 20–100 µl. More than 70% of mice on the other days generated PVS numbers less than 6 and below 63% of PVS were 20–200 µl, indicating that PVS parameters were significantly influenced by transportation but baseline levels were attained within three days (Figs. 4K and 4M). Mimicking transportation did not change the total PVS volume. Mice produced a total PVS volume of 615.70 ± 32.55 µl, 678.30 ± 41.28 µl, 627.50 ± 40.11 µl, and 591.3 ± 34.83 µl, respectively on days 1, 3, 5, and 7, which were similar to baseline levels (611.40 ± 40.20 µl) (Fig. 4N).

Limited VSA results between laboratories are comparable

A comparison study of published VSA results from two laboratories in the USA indicated some variability in VSA parameters of the same strain mice, including C57BL/6J (Bjorling et al., 2015), which may be resulted from different VSA procedures. By using the same procedure for performing VSA in two laboratories in the USA and China, we evaluated the comparability of VSA results of C57BL/6J mice from different laboratories. A total of 49 mice were subjected to VSA in BIDMC, which generated 98 filter papers, and these mice averagely produced total urine of 446.50 ± 18.30 µl during 4 h. A total of 61 mice in SWMU were subjected to VSA and produced 122 filter papers, and these mice averagely generated total urine of 695.80 ± 23.16 µl (Figs. 5A–5C), indicating that mice in SWMU voided more urine than mice in BIDMC. A total of 98% of mice in SWMU voided total urine more than 300 µl and 7% of mice voided exceeding 1,100 µl, while only 76% of mice from BIDMC voided more than 300 µl and no mice were showed voiding more than 1,100 µl (Figs. 5D). Interestingly, mice in the two laboratories had similar total voiding spot numbers, which were 11.65 ± 0.52 and 13.39 ± 0.86 (Fig. 5E). A total of 70% of mice in SWMU voided 6–15 times, 11% of mice voided 1–5 times, and the rest 19% of mice voided 16–30 times. A total of 37% of mice in BIDMC voided 6-15 times, 21% of mice voided 1–5 times, and the rest 42% of mice voided 16–40 times (Fig. 5F). Mice from BIDMC produced 3.62 ± 0.21 PVS with an average volume of 117.00 ± 4.83 µl in a 4-hour period, which is consistent with the previous study. Mice in SWMU produced 5.47 ± 0.19 PVS with an average volume of 125.40 ± 3.08 µl (Figs. 5G and 5I). Mice in the two laboratories had similar PVS volume but showed differences in PVS numbers. The volume–frequency chart of PVS showed similar voiding volume distributions between mice in both laboratories (Fig. 5H). A total of 65% of mice in SWMU had PVS numbers more than 5, while 72% of mice in BIDMC had PVS numbers less than 5 (Fig. 5J). The differences in PVS number contributed to the variety in total PVS volume, mice in SWMU produced a total PVS volume of 685.70 ± 23.33 µl, more than 424.00 ± 17.51 µl produced by mice in BIDMC (Fig. 5K).