Heterogeneity and lyophilization comparison of stool processing for gastrointestinal bile acid measurement by LC-MS/MS

Abstract

Fecal bile acid (BA) analysis is an emerging area of gut microbiome research. However, sample preparation procedures for fecal BA analysis are not standardized. Current fecal BA analysis often utilizes either original or lyophilized aliquot, and fecal BA result difference between these two processing steps remains not systematically investigated. Moreover, the distribution pattern of fecal BA in the collected stool sample also remains unclear but affects interpretation of fecal BA for downstream experiments. To address these two questions regarding effect of lyophilization on fecal BA and fecal heterogeneity, fourteen separate BAs were quantified from 60 aliquots obtained from 10 clinical fecal samples using liquid chromatography-tandem mass spectrometry (LC-MS/MS). BA concentrations in the lyophilized sample were typically 2–4 folds higher than those in the original sample, but were almost identical using a water-adjusted lyophilized BA concentration. The fecal BA compositional profile and four BA ratios were similar utilizing either the original or lyophilized samples. BA concentrations were similar among different aliquots of differing starting mass except for the relatively trace-level BA. Therefore, it is suggested that fecal BA concentrations should be presented as the original sample concentration or water-adjusted lyophilization concentration to allow comparisons between studies. A single aliquot (20–100 mg) of stool can be used to reflect the concentrations in the entire sample. These results help to standardize analyses in this emerging field

Graphical Abstract

1. Introduction

Bile acids (BAs) are end products of cholesterol metabolism and play a crucial role in the emulsification and absorption of dietary fat, steroids, lipophilic drugs, and vitamins and act as signaling compounds to mediate the metabolism of glucose [1, 2]. BAs are excreted into the duodenum with the vast majority (~95%) reabsorbed into the systemic circulation by enterohepatic recirculation [2]. The remaining BAs that are not reabsorbed are biotransformed by specific gut microbiota to secondary bile acids [2]. BA metabolism by gut microbiota has several areas of biomedical interest including liver cancer, cardiovascular disease, and infectious diseases [3–7]. The chromatographic methods (e.g. gas chromatography-mass spectrometry (GC-MS) and liquid chromatography-tandem mass spectrometry and (LC-MS/MS)) to quantitate BAs are well developed for quantifying various bile acids [8, 9]. However, the upstream fecal sample preparation steps have not been standardized. In general, stool sample preparation involves using either the original sample or a lyophilized sample [10–12]. Whether any of these sample processing steps effects BA quantitation is unknown. Likewise, the heterogeneity of BA concentrations within stool samples has not been systematically tested. The goals of this project were to compare BA quantitation utilizing the original samples compared to lyophilized samples and the effect of water-adjusted lyophilized concentrations. A secondary objective was to compare the variation in fecal BA concentrations among aliquots of different weight from the same sample.

2. Materials and experimental section

2.1. Reagents

BA standards, ammonium acetate, and ammonium hydroxide were purchased from Millipore-Sigma (USA); the BA standards included cholate (CA), chenodeoxycholate (CDCA), glycocholate (GCA), taurocholate (TCA), glycochenodeoxycholate (GCDCA), taurochenodeoxycholate (TCDCA)), lithocholate (LCA), deoxycholate (DCA), ursodeoxycholate (UDCA), hyodeoxycholate (HDCA), glycolithocholate (GLCA), taurolithocholate (TLCA), glycodeoxycholate (GDCA), and taurodeoxycholate (TDCA), as well as two internal standards (LCA-d5 and CA-d5). Methanol and water (LC-MS-grade) were purchased from VWR (Radnor, PA, USA).

2.2. Fecal sample preparation

A total of 10 human clinical fecal samples collected during a clinical Phase 1 trial study that evaluated the safety, pharmacokinetics, and fecal microbiome effects of ibezapolstat in healthy subjects were used [13]. Fecal samples were collected, shipped on dry ice to the laboratory, and immediately stored at −80 °C until analysis. After thawing, each original fecal sample was split into two parts. One part was used to prepare three aliquots with the weight ranging from 27.9 to 141.7 mg. Another part was frozen at −80 °C overnight and lyophilized (Labconco, USA) for 30 h after which three aliquots from each lyophilized part were prepared with weight ranging from 9.3 to 51.4 mg.

2.3. BAs extraction and quantification

Each original or lyophilized aliquot was suspended in 50% methanol (1 ml, containing 200 ng/ml of internal standard LCA-d5 and CA-d5), incubated at 4 °C overnight, and then centrifuged at 10000 g for 3 min. One hundred μl of supernatant was diluted by 10 folds with pure water and applied to a 96-well Oasis SPE plate (Milford, MA, Waters). After washing with 5% methanol, BA extracts were eluted with pure methanol, air dried, and re-suspended in 50% methanol solution (1 ml). Bile acids were quantified using liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) on a QTRAP 5500 mass spectrometer (Sciex, Framingham, MA, USA) according to a previously validated method [7, 14]. Briefly, chromatographic separation of BAs was conducted on a C18 column (Phenomenex, Torrance, CA, USA) by gradient elution using mobile phase A: methanol-water (1:1, vol/vol) with 10 mM ammonium acetate and 0.1% (wt/vol) ammonium hydroxide; and mobile phase B: methanol with 10 mM ammonium acetate and 0.1% (wt/vol) ammonium hydroxide). BAs measurements were normalized by the weight of the original or lyophilized aliquot. To compare the difference in BA quantification utilizing the original sample and the lyophilized sample, the fecal BA concentration in lyophilized sample was further corrected to its water-adjusted lyophilized fecal BA concentration using the water percentage (or content) in the original sample. Water percentage in the original sample was calculated by dividing water content (the change in the original sample to the lyophilized sample) by the mass of wet weight of original sample.

Additionally, the following four common BA molar ratios were studied for their potential diagnostic values [1]: (i) the ratio between LCA and DCA (LCA/DCA); (ii) the ratios between glycine-conjugated BAs and the taurine-conjugated BAs (G/T); (iii) the ratio between the free BAs and the conjugated BAs (F/C); and (iv) the ratios between primary BAs and secondary BAs (P/S). Meanwhile, to avoid the biased F/C and P/S ratio values, both UDCA and HDCA were excluded since their corresponding glycine-and taurine-conjugates were not analyzed in this work.

2.4. Statistical analysis

All statistical analysis was performed using NCSS 2007 software (LLC. Kaysville, Utah, USA). A p value of less than 0.05 was considered significant. The difference between total BA concentrations measured from the original sample and the water-adjusted lyophilized fecal BA sample was analyzed using a nonparametric, Wilcoxon Signed-Rank test because of the non-normal distribution of data in normality analysis. The difference in the RSD value between the trace BA concentration (≤1 pmol/mg) and non-trace BA (>1 pmol/mg) was analyzed using the nonparametric, Mann-Whitney U test because of the non-normal distribution of the data in the normality analysis.

3. Results

The fecal BA concentrations from the 10 original samples (pmol/mg) are shown in Figure 1A and the fecal BA concentrations from the lyophilized samples (pmol/mg) are shown in Figure 1B. Fecal BA concentrations from the lyophilized samples were typically 2–4 folds higher than the fecal BA concentrations from the original samples. The water-adjusted lyophilized fecal BA concentrations (pmol/water-adjusted weight, mg; hereafter, pmol/w-mg) are shown in Figure 1C with results in agreement with the original fecal BA concentrations (Figure 2A). The linear regression equation comparing the original to the water-adjusted lyophilized BA concentrations was expressed as LogY=1.002logX-0.006, (R²=0.987, p<0.01), in which X denotes the BA concentration of the original sample and Y denotes the water-adjusted lyophilized BA concentration of its corresponding lyophilized sample. Statistically, there was no significant difference between these two types of BA concentrations (Figure 2B, p>0.05, Wilcoxon Signed-Rank test). The BA molar ratios between the original and lyophilized sample (LCA/DCA, G/T, F/C, and P/S) were similar (Table 2 and 3). The relative standard deviations (RSDs) of the BA ratios between the two types of samples were on average 14% (LCA/DCA), 14% (G/T), 11% (F/C), and 13% (P/S). The correlation of the BA ratios measured from two types of samples were plotted in supplementary Figure 1A–1D. The BA compositional profile patterns measured from two types of samples were almost identical among 9 out of 10 fecal samples (supplementary Figure 2). The exception was a sample from subject 8 in which the proportion of CDCA (51%) measured from lyophilized sample was higher than that (27%) measured from the original sample.

Figure 1:

(A) Original fecal BA concentrations (pmol/mg); (B) Lyophilized fecal BA concentrations (pmol/mg); (C) Water-adjusted lyophilized fecal BA concentrations (pmol/w-mg).

Figure 2:

(A) Scatter plot of original fecal BA concentration (pmol/mg) versus the water-adjusted lyophilized fecal BA concentration (pmol/w-mg). (B) Wilcoxon Signed-Rank test comparison between original and water-adjusted fecal BA concentrations (n.s.: not significant (p>0.05)).

Table 2.

Comparison of P/S and G/T ratio values measured between original and lyophilized samples.

Sample	Ratio (P/S) a					Ratio (G/T) b
	Original sample		Lyophilized sample		Comparison	Original sample		Lyophilized sample		Comparison
	(Mean ± SD)	RSD (%)	(Mean ± SD)	RSD (%)	RSD c (%)	(Mean ± SD)	RSD (%)	(Mean ± SD)	RSD (%)	RSD c (%)
1	0.006 ± 0.000	0	0.005 ± 0.000	0	16.7	-	-	2.0 ± 0.3	15	-
2	0.002 ± 0.000	0	0.002 ± 0.000	0	0	32 ± 2	6.3	40 ± 2	5	15.8
3	0.007 ± 0.002	28.6	0.02 ± 0.02	100	64.3	11 ± 3	27.3	11 ± 3	27.3	0
4	0.013 ± 0.001	7.7	0.012 ± 0.003	25	7.7	93 ± 18	19.4	148 ± 66	44.6	32.3
5	1.69 ± 0.02	1.2	1.56 ± 0.06	3.8	5.5	10 ± 1	10	10.6 ± 0.2	1.9	3.9
6	13 ± 1	7.7	13.4 ± 0.3	2.2	2.3	197 ± 75	38.1	165 ± 31	18.8	12.5
7	33 ± 3	9.1	32 ± 2	6.3	2.2	56 ± 25	44.6	58 ± 5	8.6	2.5
8	2.8 ± 0.4	14.3	3.3 ± 0.6	18.2	12.9	11 ± 2	18.2	14 ± 5	35.7	16.8
9	26 ± 1	3.8	22 ± 1	4.5	11.7	5 ± 1	20	3 ± 1	33.3	35.4
10	61 ± 6	9.8	57 ± 1	1.8	4.7	1.9 ± 0.0	0	2.0 ± 0.2	10	3.6

^aP/S, the ratio between primary and secondary BA)

^bG/T, the ratio between glycine-conjugated and taurine-conjugated BA

^cRSD, relative standard derivation based on two mean values measured from original and lyophilized samples.

Table 3.

Comparison of F/C and LCA/DCA ratio values measured between10 original and lyophilized samples.

Sample	Ratio (F/C) a					Ratio (LCA/DCA) b
	Original sample		Lyophilized sample		Comparison	Original sample		Lyophilized sample		Comparison
	(Mean ± SD)	RSD (%)	(Mean ± SD)	RSD (%)	RSD c (%)	(Mean ± SD)	RSD (%)	(Mean ± SD)	RSD (%)	RSD c (%)
1	293 ± 19	6.5	230 ± 82	35.7	17	0.25 ± 0.00	0	0.19 ± 0.01	5.3	18.2
2	97 ± 3	3.1	101 ± 5	5	2.8	0.48 ± 0.02	4.2	0.43 ± 0.01	2.3	8.7
3	148 ± 8	5.4	166 ± 20	12	8.1	0.52 ± 0.10	19.2	0.40 ± 0.03	7.5	17.4
4	31 ± 2	6.5	29 ± 5	17.2	4.7	1.50 ± 0.04	2.7	1.67 ± 0.16	9.6	7.5
5	180 ± 22	12.2	186 ± 20	10.8	2.3	0.22 ± 0.01	4.5	0.22 ± 0.01	4.5	0
6	42 ± 3	7.1	45 ± 2	4.4	4.8	0.002 ± 0.000	0	0.002 ± 0.000	0	0
7	82 ± 6	7.3	77 ± 3	3.9	4.4	0.003 ± 0.001	33.3	0.002 ± 0.001	50	33.3
8	199 ± 31	15.6	354 ± 208	58.8	39.6	0.39 ± 0.03	7.7	0.21 ± 0.13	61.9	43.3
9	202 ± 72	35.6	220 ± 3	1.4	6	0.008 ± 0.000	0	0.009 ± 0.000	0	11.1
10	159 ± 3	1.9	124 ± 9	7.3	17.5	0.01 ± 0.00	0	0.01 ± 0.00	0	0

^aF/C, the ratio between free and conjugated BA

^bLCA/DCA, the ratio between LCA and DCA

^cRSD, relative standard derivation based on two mean values measured from original and lyophilized samples.

To explore whether BAs were evenly distributed in fecal samples, the variation coefficient (expressed as relative standard derivation, RSD) of each BA concentration was compared in three different aliquots from the same sample. The average RSDs of the BA concentration in the original and lyophilized samples were 19% and 17%, respectively. Generally, Low concentration BA had larger variation compared to higher concentrations BA (Figure 3A and 3C). A cutoff criterion for trace level fecal BA concentration was set at ≤1 pmol/mg (wet weight), the RSD values of the relatively trace BAs averaged 40% compared to 14% for no-trace BAs (Mann-Whitney U test, p<0.001). Similar results were observed with lyophilized samples (Figure 3B and 3D). In the present study, these trace-level BAs and non-trace-level BAs appeared to be related to the subject and BA variant itself. While certain bile acids (e.g. LCA) can be at the relatively trace level in some subject samples but abundant in other subject samples, some bile acids (e.g. TCDCA and TDCA) were consistently at the relatively trace level across all subject samples. For example, the concentrations of LCA in the volunteer sample 1, 2, 3, 4, 5, and 8 (original sample type) were high to 129, 550, 735, 320, 78, and 50 pmol/mg, respectively. By contrast, they are below 1 pmol/mg in the sample 6, 7, 9, and 10. Both TCDCA and TDCA were below 1 pmol/mg across all 10 subject samples. Of note, in lyophilized aliquots, high variation values (49–109%) was occasionally observed in certain hydrophobic BAs (e.g. CDCA in the lyophilized aliquots from the sample 3).

Figure 3:

Scatter plots of relative standard deviations (RSDs, %) based on (A) original fecal BA concentrations and (C) water-adjusted lyophilized fecal BA concentrations. Mann-Whitney U test comparison of RSD values between (B) trace BA (≤ 1 pmol/mg) and (D) non-trace BA (>1 pmol/mg).

4. Discussion

Sample preparation is an important step in analysis of fecal BA. However, no standardized procedure for BA analysis sample preparation has been proposed. Fecal BA analysis in gastrointestinal studies generally use the original sample [10] or lyophilized samples [7]. After quantification by chromatographic methods (e.g. LC-MS/MS or GC-MS), BA concentrations are usually normalized by the sample weight. Due to the weight difference between original and lyophilized samples and variation in fecal water content [15], weight normalized samples between original and lyophilized samples will generate different numerical results. In addition, microbiome analyses of fecal samples will often involve other assays that require using the same stool sample. This requires splitting the fecal sample into several aliquots for subsequent experiments. However, the heterogeneity of BA concentrations in different aliquots of the same stool sample has not been systematically studied. This work addressed these two common concerns relating to (i) the effect of lyophilization on the fecal BA measurement and (ii) the heterogeneity of BA in fecal samples in the gastrointestinal microbiome study. This study demonstrated that using either the original stool sample or a lyophilized sample yielded comparable performance in quantifying the fecal BA concentrations, BA ratios, and BA compositional profiles. BA concentrations generally showed low variance in different aliquots of original or lyophilized sample.

In this work we measured the water percentage in the original stool sample and BA concentrations in the original and lyophilized stool samples. Water percentage ranged from 68% to 86% (Table 1), and the lyophilized fecal BA concentrations were typically 2–4 folds higher than the fecal BA concentration of the original sample. After normalization for water percentage, the water-adjusted lyophilized BA concentrations were comparable to BA concentrations in the original sample (Figure 1–2, Table 1). Our study also demonstrated that lyophilization did not significantly affect ratios of BA sub-components (Table 2 and 3, supplementary Figure 1). This finding is supported by previous studies of fecal BA analysis, in which spiked BA [16], or ¹⁴C-labeled BA[17, 18] applied prior to lyophilization was associated with satisfactory BA recovery. Similarly, in a recent study of neonatal fecal samples, Shen et al (2021) also demonstrated lyophilization had minimal effect on quantifying the relative abundance of fecal metabolites (including bile acids) using the original or lyophilized sample [19]. A study by Shafaei et al (2021) reported that lyophilized aliquots resulted in lower BA concentrations and this problem was addressed by spiking internal standards into the original samples prior to lyophilization [9]. The reason for this discrepancy between this observation and others including ours is unclear. Possibilities include differing lyophilization conditions and sample types. Our present study analyzed 60 aliquots from 10 unique subject samples extracted with 50% methanol (containing internal standards) involving ultrasonication and overnight incubation. Shafaei et al utilized different methodology using 8 aliquots from a homogenized sample combined from 6 individual wet stools that may have resulted in different results [9]. Further study will be needed to clarify these differing findings.

Table 1.

Comparison of total bile acid (BA) concentrations measured between original frozen and lyophilized clinical samples.

Sample	Original sample			Lyophilized sample			Comparison between BA in original and water-adjusted BA in lyophilized
	Water content a (%)	Total BA	RSD	Total BA	Water-adjusted total BA	RSD	RSD d
		(Mean ± SD) b (pmol/mg)	(%)	(Mean ± SD) (pmol/mg)	(Mean ±SD) (pmol/w-mg) c	(%)	(%)
1	69	677 ± 25	3.7	2755 ± 183	854 ± 57	6.7	16.4
2	73	1837±31	1.7	6773 ± 496	1829±134	7.3	0.3
3	71	2244 ± 286	12.7	8266 ± 687	2397±199	8.3	4.7
4	68	650 ± 66	10.2	1957±212	626 ± 68	10.9	2.7
5	81	1341±51	3.8	7147 ± 286	1358±54	4	0.9
6	80	2493 ± 305	12.2	11136±171	2227 ± 34	1.5	8
7	76	1983±216	10.9	7981±617	1915±148	7.7	2.5
8	77	725 ±127	17.5	6372 ± 3065	1466 ± 705	48.1	47.8
9	77	2432 ±176	7.2	9007 ± 185	2072 ± 42	2.0	11.3
10	86	1068 ± 42	3.9	5061±73	709 ± 10	1.4	28.6

^aWater content in original fecal sample as expressed in percentage (%)

^bSD, standard deviation

^cw-mg, water-adjusted weight for lyophilized sample

^dRSD, relative standard derivation based on the mean value of total BA in original sample and the mean value of water-adjusted BA in lyophilized sample.

The heterogeneity of fecal BAs is another research question that first began when Setchell et al (1987) noted inter-day variation in fecal BA concentration among healthy volunteers [20]. In this study, we demonstrated that for most cases, BAs had low variance (<20%) among different aliquots whether from original or lyophilized samples. This builds on previous work by Neuberger-Castillo et al who demonstrated similar BA concentrations between the aliquots of either snap frozen samples or stabilized samples in a commercially available storage tube [21]. A higher degree of variation for relatively trace concentrations of BAs (≤1 pmol/mg) was observed in this work and it suggests that multiple aliquots with higher weight be prepared if experiments are targeting these trace level BAs to accurately quantitate these BA concentrations.

In summary, this study demonstrated comparable BA concentrations between original and lyophilized samples after water-adjustment of the lyophilized samples. Therefore, either original or lyophilized samples can be used for BA analysis. A single aliquot of the original stool sample is adequate for BA analysis except for in depth investigations of trace BA. This work has important and practical implications towards standardizing sample preparation in the clinical fecal BA analysis.

Supplementary Material

1

Supplemental Figure 1: Scatter plots of (A) LCA/DCA ratio, (B) G/T ratio, (C) F/C ratio, (D) P/S ratio between the original and lyophilized sample.

2

Supplemental Figure 2: BA proportion in the original and lyophilized sample.

Highlights.

Fecal bile acid (BA) analysis is an emerging area of gut microbiome research. Sample preparation procedures for fecal BA analysis are not standardized. Using a standardized approach, after correcting for water loss fecal BA analysis from original and lyophilized results are comparable with minimal variability within the sample A single aliquot or original or lyophilized sample can be used for fecal BA analysis