Transcriptome Analysis of Particulate Matter 2.5-Induced Abnormal Effects on Human Sebocytes

Round 1

Reviewer 1 Report

This revised manuscript may have useful information, but it exhibits obvious flaws. 

[1] The mention in the abstract of an arsenic hyper-accumulating plant seems irrelevant and distracting. 

[2] Similarly, paragraph 3 in the introduction has nothing to do with the rest of the manuscript.

[3] The hypothesis that arsenic induces AQP3 in the skin (which may contribute to the pathogenesis in vivo) is only very weakly supported by the cultures in Figure 2, where the findings are not statistically significant and thus not compelling.  Unfortunately, since these cultures are used in Fig 3-5, the reader must wonder whether they are a valid model.

[4] Auphen is claimed to be nontoxic on the basis of Figure 5a, which does not even tell what is being measured.  Toxicity in vivo depends on many possible ways a chemical can interfere with normal function, all of which a single characteristic cannot mimic.  Reference 28 is cited as support for the nontoxicity, but that reference actually does not give any toxicity information.

[5] Line 194 states that “AQP3 is critical for As-BD and uptake of arsenic into keratinocytes.”  No evidence is provided that AQP3 is critical for As-BD, which is really a hypothesis that the authors are exploring.

[6] Effects on beclin-1, LC3 and p62 are assumed to relate to apoptosis and autophagy, but the connection is not explained.  The reader must wonder how conclusive an indication these markers are. 

[7] It is possible to enlarge figures 2 and 5 to read the fine print on the computer screen, but as they are printed from the pdf of the manuscript, they are impossible to read.

Author Response

This revision of Reviewer#1 is not about our paper. Please check this situation.

Reviewer 2 Report

This research has an important aspect of the effects of PM2.5 on the SZ95 sebocytes via transcriptome analysis.. The work of this paper is practical and logical. However, there are some problems to be further improved as well:

1.In figure 2b,  The phosphorylation results of MAP2K5 shown by WB do not adequately explain the problem, please provide other pictures.

Author Response

1. In figure 2b, The phosphorylation results of MAP2K5 shown by WB do not adequately explain the problem, please provide other pictures.
Answer: Thank you for your meticulous review. Following the reviewer’s suggestion, we changed the results of MAP2K5 by WB (Figure 2b).

Reviewer 3 Report

Methods in this paper lack clarity and detail. 

1. Where and how was the PM2.5 collected. It is not clear if the apparatus described in the Materials and Methods, was generating the PM2.5 or collecting from a real-life polluted site. If it is former, what is the content of PM 2.5 (PAH, metals, pesticides etc). If later, then describe the type of polltants collected (Traffic, combustion etc).

2. Culture conditions of Sebocytes for Transcriptomics analysis is not clear. What was the duration of exposure of sebocytes with PM 2.5 before harvesting.

3. What is the rationale for exposure to 100 microgram/ml of PM2.5. If we consider real life situation of exposure to PM 2.5 through inhalation and then entering the systemic route, the concentrations at the cellular levels would be in nanograms/ml. How would authors justify such high concentrations. 

Author Response

Methods in this paper lack clarity and detail. 
1. Where and how was the PM2.5 collected. It is not clear if the apparatus described in the Materials and Methods, was generating the PM2.5 or collecting from a real-life polluted site. If it is former, what is the content of PM 2.5 (PAH, metals, pesticides etc). If later, then describe the type of polltants collected (Traffic, combustion etc).
Answer: Thank you for pointing this out. We collected PM2.5 from a real-life polluted site (the rooftop of the Amorepacific Corporation R&D building located in Yongin, Korea [37°15' N, 127°06' E]) from January to April, 2018. We add this information of PM2.5 to the Materials and Methods (line 285-288). We analyzed the metal contents of PM2.5 (Pb, As, Cr, etc.) by ICP-MS analysis, although the results were not included in this manuscript. The detailed methods were described in a previous study (Toxicology letters 2017, 273, 26-35). Also, we compared PM2.5 with SRM 2786 (Standard Reference Material; Fine Particulate Matter [<4 μm]) via coherent anti-Stokes Raman scattering (CARs) and two-photon excitation fluorescence (TPEF) imaging (Int. J. Mol. Sci. 2021, 22, 5199.).

2. Culture conditions of Sebocytes for Transcriptomics analysis is not clear. What was the duration of exposure of sebocytes with PM 2.5 before harvesting.
Answer: We agree with this. Sebocytes were treated with PM2.5 (100 μg/mL) in phenol-red free DMEM/F-12 with 1% penicillin/streptomycin for 24 h for transcriptome analysis. We add these descriptions to the Materials and Methods (line 296-298).

3. What is the rationale for exposure to 100 microgram/ml of PM2.5. If we consider real life situation of exposure to PM 2.5 through inhalation and then entering the systemic route, the concentrations at the cellular levels would be in nanograms/ml. How would authors justify such high concentrations. 
Answer: Thank you for your pointing out this. The exposure concentration of PM2.5 was set to sub-cytotoxic level with 80-90% cell viability (Figure S1). When studying with environmental factors such as ultraviolet B (UVB), it's like treating with a high level of minimal erythema dose (MED). Although the concentration is higher than in real life situation, it may be suitable to get clear responses and potential mechanisms. There are many studies on the short term effects of high concentration of PM (mainly 1-400 μg/mL; Particle and Fibre Toxicology (2020) 17:35 & Inhalation Toxicology, 20:399–414, 2008). Also, we confirmed the PM2.5 effects on sebocytes using human skin models with PM2.5 (20 μg/cm2) topically applied on the surface (Figure 4). The conditions are closer to real life. According to our calculation, the amount of PM2.5 used (20 μg/cm2) is similar to the amount accumulated on the skin (25.2 μg/cm2), assuming that all PM touching the skin surface for 1 h accumulate on the skin surface on a day with high concentration of PM2.5 (35 μg/m3; average air flow of 2 m/s).

Editor comments

(I) Please check that all references are relevant to the contents of the manuscript.
Answer: We checked all references and edited several references. We marked up using the “Track Changes” function.

(II) Any revisions to the manuscript should be marked up using the “Track Changes” function if you are using MS Word/LaTeX, such that any changes can be easily viewed by the editors and reviewers.
Answer: We marked up using the “Track Changes” function to easily view the changes.

(III) Please provide a cover letter to explain, point by point, the details of the revisions to the manuscript and your responses to the referees’ comments.
Answer: We submit this cover letter with point by point responses to referees’ comments.

(IV) If you found it impossible to address certain comments in the review reports, please include an explanation in your rebuttal.
Not Applicable

(V) The revised version will be sent to the editors and reviewers.Thank you for your consideration. I look forward to hearing from you.

Round 2

Reviewer 1 Report

Transcriptome analysis of particulate matter 2.5-induced abnormal effects on human sebocytes 

This manuscript has a wealth of data.  The overall approach is appealing.  A generic problem with such studies is how to relate the observations to exposure in vivo.  The discussion provides a well thought out interpretation of how the results could relate to observed skin effects.  However, the model has weaknesses, some of the effects seem of questionable magnitude, and some aspects of the data could be explained more fully.

[1] Although the SZ95 cells are derived from human sebocytes and retain some properties of the normal cells, the reader must wonder how closely the regulation is to normal.  Since it is a T-antigen immortalized line with highly aneuploid chromosome number, a direct extrapolation to normal physiology is not justified without specific rationale.  The readers should be made aware of this limitation.

[2] The text neglects to mention the passage number, important since the degree of departure from normal generally increases with passage.

[3] The text does not indicate the degree of confluence, which can influence the state of differentiation for those properties that are retained.

[4] Some of the characterization of responses (for example, the GO analysis) is formulaic without providing useful insight.  If the authors wish to keep it, the text would be improved by relegating it to supplementary material.

[5] The idea of confirming transcriptional responses by western blotting is good, but the results vary in their ability to convince the reader.  In figure 1d, for example, effects on IL1B and Claudin1 are clear, but those on CYP1A1, SCD, HMGCS1 are not.  Similarly, the reader is hard pressed to see some of the differences claimed in Fig 2a and b.  Perhaps quantitation would make these more convincing.  It is not clear what the reader is supposed to see in Fig 3d.

[6] Fig 4 is quite difficult to decipher.  The difference between the line labeled x200 and the line below it are not clear.  While staining with CK7 and MUC1 is obvious, what the reader is supposed to see in the other sections is not.  Perhaps some arrows to direct the reader’s attention would help.

[7] Some of the tabulated items could use further explanation. For example, in Table 1, the reader has no idea of the importance of the values in the columns “Ratio” and “z-score”.  Similarly, in Table 2, neither the “Activation z-score” nor the “overlap” explained in the next column are explained.  Table 3 has similarly unexplained columns.

[8] The cell viability by the CCK-8 assay employed shows little if any sensitivity toward the samples, rationalizing the high exposure levels in the cultures.  This seems quite unusual, leading the reader to wonder whether the CCK-8 assay is sufficiently sensitive (what it measures in not explained).  Reduction of exposure level to 20 ug/ml in Figure 4 is puzzling.

Author Response

Transcriptome analysis of particulate matter 2.5-induced abnormal effects on human sebocytes

This manuscript has a wealth of data.  The overall approach is appealing.  A generic problem with such studies is how to relate the observations to exposure in vivo.  The discussion provides a well thought out interpretation of how the results could relate to observed skin effects.  However, the model has weaknesses, some of the effects seem of questionable magnitude, and some aspects of the data could be explained more fully.

1. Although the SZ95 cells are derived from human sebocytes and retain some properties of the normal cells, the reader must wonder how closely the regulation is to normal. Since it is a T-antigen immortalized line with highly aneuploid chromosome number, a direct extrapolation to normal physiology is not justified without specific rationale. The readers should be made aware of this limitation.

Answer: Thank you for pointing this out. SZ95 sebocytes are the immortalized human sebaceous gland cell line with simian virus 40 large T antigen. The SZ95 sebocytes retain fundamental characteristics of normal human sebocyte, such as expression of sebocyte lineage markers and differentiation by accumulating lipid synthesis (J Invest Dertnatol 113:1011-1020, 1999, J Invest Dermatol 120:175-181, 2003, and Dermatoendocrinol. 2009 Mar;1(2):92-5). We added this information to the manuscript (line 65-69). Although the sebaceous gland cell line might have limitations such as incomplete differentiation and the possibility of interference in cellular processes by the immortalizing genes, the SZ95 sebocytes are widely used in sebocyte research including lipogenesis and innate immunity (Experimental Dermatology. 2018;27:484–488 and Experimental Dermatology. 2020;29:833–839). In addition, we tried to verify the results using human skin models because there are limitations using cell lines. 

2. The text neglects to mention the passage number, important since the degree of departure from normal generally increases with passage.

Answer: Thank you for your meticulous review. SZ95 sebocytes were used at 30-39 passages and at 80-90% confluence for the experiments. We added this information to the Materials and Methods (line 312-313). According to Zouboulis et al, SZ95 sebocytes show a similar epithelial morphology and can produce squalene and wax esters, as well as triglycerides and free fatty acids, even after 25–40 passages (J Invest Dermatol 113:1011-1020, 1999 & Experimental Gerontology 42 (2007) 879–886).

3. The text does not indicate the degree of confluence, which can influence the state of differentiation for those properties that are retained.

Answer: SZ95 sebocytes were used at 30-39 passages and at 80-90% confluence for the experiments. We add this information to the Materials and Methods (line 312-313).

4. Some of the characterization of responses (for example, the GO analysis) is formulaic without providing useful insight. If the authors wish to keep it, the text would be improved by relegating it to supplementary material.

Answer: Thank you for pointing this out. We transferred the results of GO analysis to supplementary material (Figure S2).

5. The idea of confirming transcriptional responses by western blotting is good, but the results vary in their ability to convince the reader. In figure 1d, for example, effects on IL1B and Claudin1 are clear, but those on CYP1A1, SCD, HMGCS1 are not. Similarly, the reader is hard pressed to see some of the differences claimed in Fig 2a and b. Perhaps quantitation would make these more convincing.  It is not clear what the reader is supposed to see in Fig 3d.

Answer: Thank you for pointing this out. The intensity of protein bands by western blotting was quantified (Figure 1c and Figure 2). The fluorescence intensity (blue) in Figure 3d was also calculated using the Zen software and normalized to the number of cells. We added these results to the manuscripts (line 207-208).

6. Fig 4 is quite difficult to decipher. The difference between the line labeled x200 and the line below it are not clear. While staining with CK7 and MUC1 is obvious, what the reader is supposed to see in the other sections is not.  Perhaps some arrows to direct the reader’s attention would help.

Answer: Thank you for pointing this out. We added arrows pointing to representative stained regions (Figure 4; line 213).

7. Some of the tabulated items could use further explanation. For example, in Table 1, the reader has no idea of the importance of the values in the columns “Ratio” and “z-score”. Similarly, in Table 2, neither the “Activation z-score” nor the “overlap” explained in the next column are explained. Table 3 has similarly unexplained columns.

Answer: Thank you for pointing this out. We added the explanation of the terms (Table 1, Table 2 and Table 3; line 130-134, 156-159 and 163-166). 

8. The cell viability by the CCK-8 assay employed shows little if any sensitivity toward the samples, rationalizing the high exposure levels in the cultures. This seems quite unusual, leading the reader to wonder whether the CCK-8 assay is sufficiently sensitive (what it measures in not explained). Reduction of exposure level to 20 ug/ml in Figure 4 is puzzling.

Answer: Thank you for pointing this out. CCK-8 assay is a sensitive method based on WST-8 which is widely used in the detection of cell cytotoxicity (Anal. Commun., 1999, 36, 47–50).

PM2.5 was applied on the human skin model at the concentration of 20 μg/cm² in Figure 4. The concentration is similar to those used in other experiment using SZ95 sebocytes. Although there are slight differences depending on the size of cell culture plates, the concentration of 100 μg/ml can be converted into the amount close to 20 μg/cm² (17.2 μg/cm² for 100 mm plate or 20.8 μg/cm² for 6-well plate). According to our calculation, the amount of PM2.5 (20 μg/cm²) is similar to the amount accumulated on the skin (25.2 μg/cm²), assuming that all PM touching the skin surface for 1 h accumulate on the skin surface on a day with high concentration of PM2.5 (35 μg/m³; average air flow of 2 m/s).