Development and Validation of a LC-MS/MS Technique for the Analysis of Short Chain Fatty Acids in Tissues and Biological Fluids without Derivatisation Using Isotope Labelled Internal Standards

Round 1

Reviewer 1 Report

Manuscript details the new method for detection of the short chain fatty acids in different biological matrices with a special focus on the simplicity and productivity of analytical procedures. Compared to well-established methods known in the literature, this method is based on LCMS detection of underivatized fatty acids after their separation on a porous graphitized carbon column.

Certain elements of the manuscript need to be addressed before it can be recommended for publication.

Specific comments:

L182, 183: please rephrase. Responses were zero only in blanks, as one may suggest.

L212, 213: please rephrase.

L340, 355: indicate centrifugal force (g) rather than “high speed”.

L348: what is 100% methanol in 0.5% phosphoric acid? Please clarify.

L366: µm, not µM.

L401, 415: use same dimension, either µM or mM.

L371, 372: gradient details not clear: was washing done at 60% B for 2 min and then equilibration at 0% B for 4 min?

Table 4: H is 10 mM, not 1.

Table 5: spell out the exact time, do not use "short – long" statements.

Sections 2.6-2.7: do not repeat RT. It should be readable from the graphs (see below).

General comments:

Only present results for Hypercarb as this was the column of choice; chromatograms from PFP are irrelevant (Fig 1c, d); no need to mention PFP in Experimental. It would suffice to say that PFP phase was selected over several other chemistries which were not able to separate isobutyrate/butyrate and isovalerate/valerate.

Note that PFP column is not a variant of C18, so make sure not to use ambiguous references to it (e.g., L274, 367).

It is said that lactic acid was detected in negative ESI, but Experimental does not mention negative mode was used. In fact, it is difficult to believe that other acids are best detected in positive ESI, but I leave this at authors’ discretion.

How recovery values given in Table 3 were calculated given the endogenous presence of fatty acids? Or is it recovery in the same sense as in Table 4 (recovery of added amount)?

All figures must be legible. The images are all low-resolution raster and cannot be zoomed to see the actual data. Please redo.

Why a non-volatile phosphoric acid was selected for sample pretreatment, and not formic? One may expect that phosphoric acid will foul MS interface rather quickly. Please comment on that.

Author Response

L182, 183: please rephrase. Responses were zero only in blanks, as one may suggest.

Response: Lines 187-189: Sentence reworded “No signals were detected in any of the blank samples run amongst SCFA-containing samples, indicating that there was little or no carry over occurring.”

L212, 213: please rephrase.

Response: Line 217-218: Sentence rephased to “No significant changes were noticed in sample stability over this time period”.

L340, 355: indicate centrifugal force (g) rather than “high speed”.

Response: Line 373, 384 & 388: High speed changed to 15000 rpm

L348: what is 100% methanol in 0.5% phosphoric acid? Please clarify.

Response 1: Line 222-223: Changed to 0.5% phosphoric acid in methanol”

Response2: Line 372- Added sentence “85% orthophosphoric acid diluted in methanol”.

L366: µm, not µM.

Response: Line 399 & 400 “µM changed to µm”

L401, 415: use same dimension, either µM or mM.

Response: Lines 19, 161, 227, 297, 395 & 435 and also table 2 in LOD and LOQ columns: Units changed to mM

L371, 372: gradient details not clear: was washing done at 60% B for 2 min and then equilibration at 0% B for 4 min?

Response 1: Line 404: added sentence “using 100% mobile phase B

Response 2: Line 405: added sentence “using 100% mobile phase A

Table 4: H is 10 mM, not 1.

Response: H values changed from 1 to 10 mM

Table 5: spell out the exact time, do not use "short – long" statements.

Response 1: Time ranges added to the table

Response 2: Concentration ranges added

Sections 2.6-2.7: do not repeat RT. It should be readable from the graphs (see below).

Response 1: Retention times removed from section 2.6 and 2.7

Response 2: Sentence added to line 225“and retention times for all analytes”

General comments:

Only present results for Hypercarb as this was the column of choice; chromatograms from PFP are irrelevant (Fig 1c, d); no need to mention PFP in Experimental. It would suffice to say that PFP phase was selected over several other chemistries which were not able to separate isobutyrate/butyrate and isovalerate/valerate.

Response: Fig 1c. d removed

Note that PFP column is not a variant of C18, so make sure not to use ambiguous references to it (e.g., L274, 367).

Response: The reviewer is correct, this was stated in error, and has been corrected by removing c18.

It is said that lactic acid was detected in negative ESI, but Experimental does not mention negative mode was used. In fact, it is difficult to believe that other acids are best detected in positive ESI, but I leave this at authors’ discretion.

Response:  Ionisation depends on acidity, polarity and the pH of the mobile phase. We refer the reviewer to the following literature: J Am Soc Mass Spectrom 16(4): 446-455 and J Anal Methods Chem 2017: 5391832. We tested each SCFA empirically. We have added ‘and negative’ to section 4.4.

How recovery values given in Table 3 were calculated given the endogenous presence of fatty acids? Or is it recovery in the same sense as in Table 4 (recovery of added amount)?

Response: Yes, recovery in the same sense as table 4. A description of how the recoveries were calculated has been added to Table 3 as a legend.

Line 207: “*Recovery was calculated as (final calculated concentration-non spike concentration/added known concentration) *100.”

All figures must be legible. The images are all low-resolution raster and cannot be zoomed to see the actual data. Please redo.

Response: Figures now legible

Why a non-volatile phosphoric acid was selected for sample pretreatment, and not formic? One may expect that phosphoric acid will foul MS interface rather quickly. Please comment on that.

Response: This is a fair point – we are aware of the non-volatility of phosphoric acid. We chose phosphoric acid because it can be used for any pH adjustment due to its pKa values. Our aim was to adjust pH to 2-3 ( J Sep Sci 35(15): 1906-1913.). We have kept the injection volume low (2 ul) and the mobile phase flow rate low (0.15 ml/min) which means the small volume of phosphoric acid injected is substantially diluted in the mobile phase and we have not observed any long-term detrimental effects in detector performance.

Author Response File: Author Response.docx

Reviewer 2 Report

I must thank the author and co autor of the mentioned work above for their unique methods of modelling the data, in the most innovative ways in formulating and defining the most optimal and credible models. The conclusion shows that the chosen literature was done in very careful way that suits the topic, especially due to the fact that the literatures are new to the public. 

Author Response

Response: We thank the reviewer for their positive comments to the manuscript. There are no actionable comments to reviewer 2.

Author Response File: Author Response.docx

Reviewer 3 Report

 The proposed new analytical development is an elegant study, well written and discussed. There is a clear need for a direct LC-MS/MS quantitative methodology aiming the analysis of SCFs in biological matrices with serious matrix effects. The method has been validated for the simultaneous quantitation of acetate, butyrate, isobutyrate, isovalerate, lactate, propionate and valerate, although important limitations exit for isomeric compounds and for matrix effects in specific samples. Some major and minor points must be addressed.

• Lines 38-39, there is a paragraph jump.
• Line 44, “Additionally, due to their low, pH, SCFA, are”, the commas should be revised.
• Line 59 and through the manuscript: the authors indistinctly used method and techniques. In line 59, NMR, GC-MS, etc, are all analytical techniques and not methods. Please, revised and amend. Line 83, the authors should change “technique” by “method”.
• The authors must discuss how the low recovery rate obtained for some compounds, as well as the high matrix effects observed, will affect the reliability of the analysis, and if there are alternatives to minimize these limitations.
• Line 222, “sample. Figures 2ga-3l” is very confusing, is there any mistake there?
• I recommend introduce in the discussion section the reference “J. Proteome Res. 2021, 20, 7, 3508–3518” and discuss within lines 258 and 259.
• Table 5 is very interesting, but the references supporting the analytical features of the different methods should be included. Also, LOD and LOQ would be of interest to be further include
• Line 344, high speed? Could the authors you kindly provide rpm or so?
• Line 387, I suggest change “Analytical validation” by “Method validation”

Author Response

The proposed new analytical development is an elegant study, well written and discussed. There is a clear need for a direct LC-MS/MS quantitative methodology aiming the analysis of SCFs in biological matrices with serious matrix effects. The method has been validated for the simultaneous quantitation of acetate, butyrate, isobutyrate, isovalerate, lactate, propionate and valerate, although important limitations exit for isomeric compounds and for matrix effects in specific samples. Some major and minor points must be addressed.

• Lines 38-39, there is a paragraph jump.

Response: We could not find any paragraph jump

• Line 44, “Additionally, due to their low, pH, SCFA, are”, the commas should be revised.

Response: Line 45: commas after low and SCFA removed

• Line 59 and through the manuscript: the authors indistinctly used method and techniques. In line 59, NMR, GC-MS, etc, are all analytical techniques and not methods. Please, revised and amend. Line 83, the authors should change “technique” by “method”.

Response: Method(s) replaced with technique(s) where appropriate throughout the manuscript

• The authors must discuss how the low recovery rate obtained for some compounds, as well as the high matrix effects observed, will affect the reliability of the analysis, and if there are alternatives to minimize these limitations.

Response: Section 3.1 added to lines 309-326

“3.1. Methodology limitations

It is noteworthy that Some areas of method performance are not ideal, for example the matrix effect in the brain, liver and plasma samples may lead to low recovery rate for lactate. However, this should not compromise the accuracy for lactate measurements in these tissue samples as when analyses are carried out in the same matrix, the low recovery rate does not affect quantification accuracy particularly as this effect is counteracted by the use of isotopically labelled internal standards with quantification carried out using a standard curve. Nevertheless, this stresses the requirement that standard curves be carried out in the same matrices. The poor recovery of lactate which was measured simultaneously with SCFAs, is consistent with a previous study where lactate recovery from supernatants of bacterial culture was 25% [40]. In the CG-MS techniques described by Primec et al. although lactic acid was measured from the same samples as SCFA, it is important to note that lactate was measured separately after methylation. Low recovery has no hinderance in measuring the analyte as long as it is consistence across the batches [13]. To our knowledge, no other techniques employing LC/MS/MS have analysed lactate along with SCFAs although, lactate has previously been analysed using LC/MS/MS along with other organic acids [32]. Future studies may wish to examine the recovery of lactate after pretreatment using trichloro acatic acid, formic acid or perchloric acid.”

• Line 222, “sample. Figures 2ga-3l” is very confusing, is there any mistake there?

Response: Line 225 & 226: 2ga-3l has been corrected to 2A-2L.

• I recommend introduce in the discussion section the reference “J. Proteome Res. 2021, 20, 7, 3508–3518” and discuss within lines 258 and 259.

Response 1: Reference added

Response 2: Lines 262-265: added “Further, LC-MS techniques requiring sample derivatisation have also been used, some involving a very short run time (14 mins) and covering a wide range of gut derived metabolites but once again require longer sample preparation including sample filtration which is time consuming and laborious [21, 34-36]”.

Response three: Line 268-269 changed to “Additionally this technique did not involve simultaneous analysis of stereo isomers and lactate in a wide range of matrices

Response 4: Line 272-273 added: “The LC-MS techniques reported to date for quantifying SCFAs have not been tested or validated for use with a wide range of”

Response 5: Line 276-277 added: “Here we provide a technique with a short run time and good sensitivity for the analysis of SCFAs, stereo isomers and lactate in a wide range of matrices.”

• Table 5 is very interesting, but the references supporting the analytical features of the different methods should be included. Also, LOD and LOQ would be of interest to be further include

Response: We refer the reviewer to the comprehensive review by Primec et al.Who discussed LOQ for the GC-MS based method but in the current review it is not possible to comment on LOQ and LOD from other techniques as different are used and also, in other techniques LOQ and LOD are not described. Therefore, the sensitivity in table 5 is described as a range.

• Line 344, high speed? Could the authors you kindly provide rpm or so?

Response: Lines 373, 384, 388: High speed changed to 15000 rpm

• Line 387, I suggest change “Analytical validation” by “Method validation”

Response: Line 421 “Analytical validation” changed to “Method validation”

Author Response File: Author Response.pdf

Reviewer 4 Report

This referee has carefully studied the validation of the method proposed by the authors for the determination of fatty acids in different matrices and I must admit that the authors have done a great job. Nevertheless I still do not see the relevance of the work vs others described previously in the literature.  The work lacks discussion and it seems more a summary of the results obtained.

Table 5 must be rewritten adding the figures of merits on each case. 

My opinion is that the present work could be accepted as a short communication in the present form, sending some information to the supplementary material. But it is not strong enough for a full paper. 

Other considerations: Selection of the HPLC column should be explained.

Line 427: 5 mM instead 5mM

Author Response

This referee has carefully studied the validation of the method proposed by the authors for the determination of fatty acids in different matrices and I must admit that the authors have done a great job. Nevertheless, I still do not see the relevance of the work vs others described previously in the literature.  The work lacks discussion, and it seems more a summary of the results obtained.

Response: We thank the reviewer for their comments. However as described in the introduction and discussion section, the current technique is novel in comparison to the previously published techniques as here we offer a very quick and less laborious sample preparative approach that does not require derivatisation and filtration and is therefore very high throughput, with SCFA analysis in a wide range of matrices.  We therefore deem this manuscript important to the reader.

Table 5 must be rewritten adding the figures of merits on each case.

My opinion is that the present work could be accepted as a short communication in the present form, sending some information to the supplementary material. But it is not strong enough for a full paper.

Response: We thank the reviewer for their suggestion. However, we deem this manuscript apt as a full paper, as it covers various aspects of methodology development, it is novel and provide substantial material similar to other methodology validation articles, published as full articles in Molecules. We Refer the reviewer to the following references as examples (Molecules 26(19): 5870., " Molecules 26(18): 5561.

Other considerations: Selection of the HPLC column should be explained.

Response: As explained in section 2.2, various columns were tested but due to poor analyte separation in particular the stereoisomers, the Thermo Scientific Hypercarb (porous graphitic carbon, PGC) 3 µm (50 x 2.1 mm) column was selected. Our aim was to develop a technique with a short runtime and more efficient chromatography method to analyses SCFA. Therefore, we first selected 100 X 2.1 mm column for short run time and 1.8 um particle size for more efficiency.  Common C18 different column chemistry columns from different provider were tried because of column packing techniques are different. Porous graphitic carbon column is more retentive than C18 columns. Therefore, the column length was chosen 50 X 2.1 mm instead of 100 X 2.1 mm length column.

Line 427: 5 mM instead 5Mm

Response: Line 461: 5mM changed to 5 mM

Round 2

Reviewer 1 Report

Thanks for taking your time on working on the revised version and addressing the issues raised.