Staphylococcus aureus lipoproteins promote abscess formation in mice, shielding bacteria from immune killing

Despite being a major bacterial factor in alerting the human immune system, the role of Staphylococcus aureus (S. aureus) lipoproteins (Lpp) in skin infections remains largely unknown. Here, we demonstrated that subcutaneous injection of S. aureus Lpp led to infiltration of neutrophils and monocytes/macrophages and induced skin lesions in mice. Lipid-moiety of S. aureus Lpp and host TLR2 was responsible for such effect. Lpp-deficient S. aureus strains exhibited smaller lesion size and reduced bacterial loads than their parental strains; the altered phenotype in bacterial loads was TLR2-independent. Lpp expression in skin infections contributed to imbalanced local hemostasis toward hypercoagulable state. Depletion of leukocytes or fibrinogen abrogated the effects induced by Lpp in terms of skin lesions and bacterial burden. Our data suggest that S. aureus Lpp induce skin inflammation and promote abscess formation that protects bacteria from innate immune killing. This suggests an intriguing bacterial immune evasion mechanism.

that of Hoebe et al. (Nature 2005, 433:523-527). The differing roles of TLR2 in skin infection have been suggested to be due to different bacterial strains used and different infectious doses administered in the skin. What doses of bacteria were given to the mice? The methods section only mention bacterial volumes. 4. (Page 10 lines 199-217) Paragraph title "The effect induced by Lpp expression is abrogated by leukocyte depletion". This paragraph focuses on the attenuation of the Δlgt mutant compared to the parent strain depending on the presence of leukocytes when in fact what this shows is that leukocytes are important to contain S. aureus skin infection. 5. Line 206 -"cyclophosphamide-treated mice developed more severe disease", can you please show the statistical significance in CFUs from cyclophosphamide-treated versus untreated mice (Fig. 6B)? 6. Did you verify leukocyte depletion upon cyclophosphamide-treatment? (e.g. Flow data?) 7. Lines 207-208 -"The leukopenic mice also lost significantly more weight compared to ….infection". No figure callout, where are the data?
Minor comments: 1. Line 49 "unique structure" of S. aureus, please rephrase 2. Line 65 S. aureus 'contributes', line 350 Lpp 'plays' 3. Line 120 Lpp 'induces' 4. Why was the immune cell analysis done in infected mouse auricles and not from the skin biopsies (injected mouse backs as per methods)? 5. Line 161-162 -"SA113 Δlgt mutant strain tended to induce lower levels of MPO in the skin tissue than SA113 parental strain (Fig. 3D). But there was no significant difference according to that figure? 6. Lines 337-338 -"Bacterial MAMPs like LPS", what is the relevance to S. aureus? 7. Fig. 5G -MPO levels much lower compared to Fig. 3D, do these fluctuate that much?

Summary
In this study, authors showed Lipoprotein(Lpp) is necessary to recruit leukocytes and induce inflammation through TLR2 signaling. They also showed S. aureus with Lpp enhances skin abscess formation caused by imbalanced coagulation/fibrinolysis hemostasis in the local skin tissue by using Lpp deficient strain(Δlgt). It is interesting idea that abscess formation is the bacterial evasion mechanism. But there is important question how does host immune function fight against Lpp deficient bacteria without innate immune response? Since NewmanΔlgt has slow growth rate in biological conditions (Mohammad, M. et al. The role of Staphylococcus aureus lipoproteins in hematogenous septic arthritis. Scientific Reports 10, 7936, doi:10.1038/s41598-020-537 64879-4 (2020)), it would be difficult to simply compare the number of bacteria after 3 or 10 days. Also, authors use two different strains, SA113(Agr +) and Newman(Agr -), in different experiment. What is the impact from Agr system in this phenotype? It would be better to clarify in result or discussion.

Overall impression
The idea that bacteria take advantage of abscess formation to survive in host homeostasis, and their strategy is straight forward. But they need to address those questions.

Comments/Suggestion for authors
Major Queries 1, In figure 3 authors showed the SA113Δlgt strain cause smaller infectious lesion and less bacterial counts (Figure3C). It looks small difference and is nuclear the original number of injected SA113. Figure 3D shows SA113Δlgt has similar MPO level to WT strain. But in Figure 4B and D, NewmanΔlgt strain still cause smaller infectious lesion and less bacteria even in TLR2-/-mice. Furthermore, in figure 6, Cyclophosphamide treated leukocyte depletion mouse injected Newman parental strain didn't show higher larger skin lesion compare to PBS treated mice. Taken together it suggest Lpp is necessary to cause inflammation but bacterial survival or overload is due to the different growth speed of each bacteria. In other words, the low number of bacteria is caused by slower growth of Δlgt strain. Could you address how does host fight against Lpp deficient bacteria without innate immune response? Is there any difference in adaptive immunity? Or is it possible to knock in Lpp in Newman or SA113 Δlgt strain fo repeat experiment? 2, In figure 4, Authors use Newman(Agr -) strain instead of SA113(Agr +) for TLR2 Ko mice experiment. But there is no discussion about the impact of Agr sysmen. What is the impact of Agr system?
Minor Queries 1, The authors showed CFU in Figure 3D, 4C, 5C, 6B and 7G, but those are hard to see the difference even though there is statistically significant difference.
Reviewer #4 (Remarks to the Author): Staphylococcus aureus lipoproteins (Lpp) are membrane anchored surface proteins that play important roles in host-bacterial interactions. How Lpp involved in S. aureus pathogenicity remain poorly studied. Mohammad and colleagues showed in their previous study that Lpp induces chronic destructive macroscopic arthritis (PMID: 31226163). In this study, Mohammad and colleagues further investigate the role of Lpp, in particular "lipoprotein-like" Lpl1, in murine skin infection. They find that subcutaneous injection of Lpl1 promote leukocytes infiltration and skin lesion, which are TLR2-dependent. Also, in comparison to wild type strain, lgt-mutant strain causes smaller skin lesion size and lower bacterial loads, which the latter is TLR2-independent. Finally, they find that the skin lesions and bacterial burden induced by Lpp can be abrogated by chemically-depletion of leukocyte or fibrinogen.
This study is potentially interesting, particularly lpp-induced fibrin capsule and abscess formation as a strategy to evade host immune attack, and will be of interest to a broad audience of infectious disease researchers and microbiologists. However, the mechanism underlying the abscess formation to protect S. aureus from immune cells is less convincing, as the conclusion was made largely based on experiments using purified Lpl1 alone or co-injection with live bacteria. More rigorous examinations are required in this regard.
Below are comments/points that may help to improve the manuscript and hope the authors will find them useful.
Major points: 1. The authors showed increased chemoattractant level after purified Lpl1 s.c. injection (Fig 1), thus promoting the infiltration of leukocytes to the injection site (Fig 2). However, the experiments heavily rely on purified Lpl1 treatment, an approach that may undermine other Staph virulence factors that also cause skin inflammation and leukocytes infiltration. It would be great if the authors could use lgt-mutant and complemented mutant lgt strains to validate the significant impact of Staph Lpp in chemoattractant release and leukocytes infiltration at the skin infection site.
2. In general, TLR2 possess a protective function during S. aureus infection, which involves in neutrophils recruitment in response to S. aureus. The bacterial burden in the murine organs (Takeuchi et al. 2000, J Immunol, PMID: 11067888;Yimin et al. 2013, Plos One, PMID: 24058538) and skin (Miller et al. 2006 Immunity, PMID: 16413925) were higher in TLR2-deficient mice than wild type mice. In contrast, this study show a smaller skin lesion size and normal bacterial clearance that is similar to wild type mice (Fig 4). The authors should discuss this discrepancy.
3. What are the levels of MIP-2, KC and MCP-1 of the skin biopsy homogenate in TLR2-/-mice after Lpl1 injection, as well as in WT vs TLR2-/-mice after WT and lgt-mutant strains infection? It would be interesting to see whether these chemoattractant releases are dependent on TLR2.
4. Despite the possibility of increased neutrophils and monocytes infiltration, higher bacterial burden was observed on day 3 after s.c. skin infection of both wild type SA113 and Newman strains compared to lgt-mutant strains (In Fig 3C and 4C). It is not convincing that lgt-mutant burden is significantly lower than wild type strain at day 3 as the difference is marginal (less than 2-fold). The authors also suggest that Staph Lpp provoked less bacterial clearance in wild type SA113 compared 6. In Fig 5C and 7G, additional time-point bacterial count at day 1 or 2 will be necessary to justify that Lpp and fibrin capsules protect bacteria from immune cells killing.
7. The authors claim that cyclophosphamide-treated mice lost weight and developed more severe diseases during the course of infection (Line 205 -208), but no evidence were shown. The authors should include the data for proper interpretation.
Minor issues: 1. In Abstract, "Lpp-deficient S. aureus strains exhibited smaller lesion size and reduced bacterial loads than their parental strains; this altered phenotype was TLR2-independent.", but this is not consistent with the data, which suggest the skin lesion severity was TLR2-dependent at the early infection.
2. It would be great to include the FACS dot plots of immune cells isolated from auricular skin tissue following s.c. injection of PBS or Lpl1 in TLR2-/-mice. 3. It would be interesting to add the role of IL-1R and MyD88 to the discussion. 4. Have the authors performed the bacterial cfu count in the supernatant of skin biopsy homogenates on day 10 post-infection with wild type or lgt-mutant SA113 strains ( Fig 3C). As the authors showed the bacterial counts of Newman strain on day 10 in Fig 4C, it would be interesting to see if SA113 strain has similar effect to the Newman strain. 5. Have the authors accessed the levels of tissue factor and PAI-1 after s.c. injection of Lpl1(-sp)?
Gothenburg, January 4 th , 2021 We would like to thank referees for the insightful examination of our manuscript. In order to strengthen our conclusions, eight independent in vivo experiments and several immunological assays were performed as reviewers advised including: 1) The bigger sample size of wild-type and TLR2 -/mice for FACS analysis are now used (please see Results page 7 and figure 2). 2) To study whether adaptive immunity is involved in Lpp induced disease phenotype in skin infections, SCID and control mice were s.c. infected with Newman and Newman Δlgt strains. Our data suggest that adaptive immunity plays minor role regarding the Lpp induced phenotype in skin infections (please see Results page 11 and figure 6). 3) To study whether the complementation of Δlgt mutant restores the phenotype, complemented strain SA113Δlgt (pRB474::lgt) was used. Our data demonstrated that complementation fully restored the phenotype that was abrogated in SA113Δlgt.
(Please see results page 8-9 and figure 3). 4) To understand whether lower bacterial doses change the Lpp induced effect in skin infection model, 20 times lower bacterial dose of both SA113 and SA113Δlgt were used. Our results showed that SA113 induced more severe skin lesions and higher bacterial counts than SA113Δlgt even with lower doses (Please see results page 8 and supplementary figure 2). 5) To study whether fibrinogen depletion promotes the bacterial metastasis, the bacterial counts from different organs (kidneys, spleen, liver, lungs, and blood) were performed. Our results suggest that fibrinogen depletion did not promote the bacteria dissemination to the other organs. (Please see results page 13 and supplementary table 1) 6) To study whether the chemokine release induced by Lpp in local skin is TLR2 dependent, the WT and TLR2-/-mice were s.c. injected with purified Lpl1. Our data clearly showed that in vivo chemokine release induced by Lpp in local skin is TLR2 dependent (Please see results page 6 and figure 1I-K). 7) To verify whether lipid moiety of Lpp is of importance for the effect induced by Lpp in the skin infection caused by Lpp mixed with SA113 Δlgt, Lpl1(-sp) that lacks the lipid moiety was mixed with SA113 Δlgt instead and s.c. injected to the mice. Our data suggest that the lipid moiety of Lpp is crucial for the disease phenotype observed such as bigger skin lesions, higher bacterial loads, higher chemokine and MPO levels (Please see results page 11 and figure 5H-M). 8) To understand the effect of Lpp in skin infections at the early disease course, control substance or Lpl1(+sp) was mixed with live SA113Δlgt strain and s.c. injected to the mice. Our results demonstrate that at the early timepoints when the fibrin capsules were not formed yet, the bacterial load in the local skins was similar in two groups (Please see results page 10 and supplementary figure 4).
Below please find our responses to all the queries, point by point.
Manuscript ID Ref.

Reviewers' comments:
Reviewer #1 (Remarks to the Author): In "Staphylococcus aureus lipoproteins induce skin inflammation and promote abscess formation, shielding bacteria from immune killing", Majd Mohammad and colleagues examine immune effects of lipoproteins of the human pathogen Staphylococcus aureus. They show that lipoproteins as proinflammatory, and bacteria deficient in them are less inflammatory. These observations are not entirely novel, but of interest and potential importance. However, key experiments are lacking adequate controls and do not fully support the conclusions. Specific notes follow:

Major comments
Altogether, significance appears to be usually reached based on animal number, and not necessary biological mechanism. Power functions need to be used to determine appropriate sample numbers which should be applied more consistently. In some experiments large numbers of mice are used to make fairly minor points (eg 32 mice in Fig 5C), while others are very underpowered (Figure 2 makes conclusions on the importance of TLR2 based on groups of 3 mice). Many appropriate comparisons appear that they would reach significance with adequate mouse numbers (for example, monocyte and macrophage number induced changes induced by Lpl1 in TLR2 mice, which does not agree with the proposed molecular mechanisms). In many paired datasets, the number of animals changes, suggesting samples are being omitted or other factors are confounding analysis. Its also not clear how many experiments are performed, when samples are combined, or are representative, or these differences decided.
: We are very sorry for confusing data presentation. In figure 5C, we performed three independent experiments, which exhibited similar results (2 of 3 with p<0.05 and the third one was also close to significance), and data were thus pooled. Only samples from one representative experiment were used for chemokines and MPO assays. We have now added this information into the figure legend. Lpp deficient bacteria, expectedly, have decreased viability and growth, for reasons including deficiencies in iron acquisition. These mutants are more susceptible to killing by host immune effectors and are rapidly cleared -any mutant that so fundamentally impacts bacterial survival will decrease inflammation, but this is very non-specific and not a mechanism. This has not been sufficiently discussed or deconvolved as a contributor to the in vivo observations. The attenuation of Lgt mutants also appears highly variable between strains and experiments. Prior literature is all consistent with these mutants being highly attenuated, which makes sense, but why this is seen in some experiments and not others is of concern. Potentially there are issues with mutants, or these are not fundamental mechanisms, but a phenomenology of certain strains that would at least require additional explanation.
: We agree. Now we have adjusted the discussion. The fibrin capsule formation is another mechanism above of the iron acquisition mechanism.
Line 147, it is not clear that the wound healing process was "faster", as stated. The data are consistent with healing starting sooner, because the infection is resolved soon, since the bacteria are attenuated. Line 176; similar comment. The apparent role of lpp early in infection is likely due to attenuation and reduced bacterial burden; evidence is not consistent with roles of TLR2 signaling in this. The slopes of lesion size (figure 4, for example) suggest wound healing is not different, just starting from a different size.
: We have rephrased our conclusion that is more objective. We hope you agree with us.
Line 216. Also consistent with this model is that cyclophosphamide depletion of important immune cells reduces the killing of staph, even leading to a partial attenuation reversal of the lpp mutant strain.
Fibrin is a barrier to bacterial dissemination; therefore, ancrod depletion of fibrinogen is expected to promote staph dissemination. Figure 7 shows smaller lesions with ancrod treatment, with correspondingly fewer bacteria, but it is not clear that there just isn't fewer bacteria due to spread to other sites. No details on purity of Lpl1 : Lpl1 was prepared as previously described. We have now added the information about the purity of Lpl1 as reviewer suggested.

: Thank you for the constructive suggestion. It is obviously important to rule out this possibility. In order to address this question, we repeated the experiment and assessed the bacterial load in various organs, such as the blood, kidneys, liver, spleen, lungs, and also collected larger skin biopsies (double-sized biopsies than usual) of the mice treated with either PBS or Ancrod followed by subcutaneous injection of S. aureus Newman parental strain or Newman
Line 167, discussion on the significance of agr+ would be useful : We would like to thank review once again for insightful examination. The same comment was addressed by reviewer 3 and 4. Now we have added more discussion on it.
Are cytokines normalized by lesion size? Methods on sample collection could use more detail. Cytokines, for example, on Y-axis of graphs appear to be normalized to tissue mass. However, methods indicate skin was consistently collected by 8mm biopsy punch. This would be irrespective of lesion size, so small lesions will inherently have the cytokine numbers diluted out by normal tissue. Obviously, the majority of MCP-1 comes from the lesion and the relationship remains the same between total MCP-1 amount (indicated in our figures) and the amount calculated/released from skin lesion.

Histology lacks scale bars
: Thank you for the comment. This is now added in the figure legends of figure 1 and 7, respectively.

Reviewer #2 (Remarks to the Author):
This study shows that the lipid moiety of S. aureus lipoprotein (Lpp) contributes to the pathogenesis of skin infection by stimulating inflammation and abscess formation, using purified Lpp (Lpl1) and S. aureus lacking lipidated Lpp (Δlgt mutant) in mice. The authors also suggest that Lpp-induced abscess formation contributes to staphylococcal immune evasion given that fibrinogen depletion promoted WT S. aureus clearance but not the Δlgt mutant. While interesting, these are not novel concepts (Schmaler et al. J Immunol June 1, 2009, 182 (11) 7110-7118, Cheng et al., FASEB J. 2009 Oct; 23(10): 3393-3404). I have the following comments: 1. The interpretation of data in the manuscript is often confusing or at times not quite accurate (see point 4 below please). The discussion section should be more concise.
: Now we have revised the discussion section. We hope the discussion is now acceptable.
2. The inconsistency in the role of TLR2 stimulation by Lpp upon the use of purified Lpp (Lpl1) versus during bacterial infection was not clearly explained. While Lpl1-induced skin lesion was dependent on TLR2, the Δlgt mutant was attenuated in both the WT and TLR2-/mice. The latter observation could be due to the fact that the Δlgt mutant has a growth defect under nutrient limitation compared to the parent strain as previously shown by Stoll et al.  3. Line 120 Lpp 'induces' : Modified as suggested.
4. Why was the immune cell analysis done in infected mouse auricles and not from the skin biopsies (injected mouse backs as per methods)?
: The skin biopsies from backs are extremely difficult to digest, prepare the single cells, and thereafter perform the FACS analysis. We have therefore utilized the well-established method of using mouse auricles. This system closely resembles the skin tissues and is much easier to work with in order to understand the immune response in vivo.
5. Line 161-162 -"SA113 Δlgt mutant strain tended to induce lower levels of MPO in the skin tissue than SA113 parental strain (Fig. 3D). But there was no significant difference according to that figure? : Correct, we have now clarified that in the result section. Please see line 194-196. 6. Lines 337-338 -"Bacterial MAMPs like LPS", what is the relevance to S. aureus? : It is a very good point and we fully agree. This part has now been removed.
7. Fig. 5G -MPO levels much lower compared to Fig. 3D, do these fluctuate that much? : These are two independent experiments with different concentrations of bacteria. Higher MPO levels are associated with higher dose (1.25x10 8 cfu/site) of bacteria in figure 3D (updated to Fig 3G). In figure 5G (2020)), it would be difficult to simply compare the number of bacteria after 3 or 10 days. Also, authors use two different strains, SA113(Agr +) and Newman(Agr -), in different experiment. What is the impact from Agr system in this phenotype? It would be better to clarify in result or discussion. : It is a very good comment (review 1 and 4 also commented on this). Now we have added the discussion in the manuscript.

Overall impression
The idea that bacteria take advantage of abscess formation to survive in host homeostasis, and their strategy is straight forward. But they need to address those questions.

Comments/Suggestion for authors
Major Queries 1, In figure 3 authors showed the SA113Δlgt strain cause smaller infectious lesion and less bacterial counts (Figure3C). It looks small difference and is nuclear the original number of injected SA113. Figure 3D shows SA113Δlgt has similar MPO level to WT strain. But in Figure 4B and D, NewmanΔlgt strain still cause smaller infectious lesion and less bacteria even in TLR2-/-mice. Furthermore, in figure 6, Cyclophosphamide treated leukocyte depletion mouse injected Newman parental strain didn't show higher larger skin lesion compare to PBS treated mice. Taken together it suggest Lpp is necessary to cause inflammation but bacterial survival or overload is due to the different growth speed of each bacteria. In other words, the low number of bacteria is caused by slower growth of Δlgt strain. Could you address how does host fight against Lpp deficient bacteria without innate immune response? Is there any difference in adaptive immunity? Or is it possible to knock in Lpp in Newman or SA113 Δlgt strain fo repeat experiment? : We would like to thank reviewer very much for the constructive suggestion. 1) Now we have investigated the impact of adaptive immunity on the Lpp induced phenotype in skin infections using SCID mice who lack of both B and T cells. The Lpp induced phenotype was retained in SCID mice, suggesting adaptive immunity played minor role there. Please see figure 6C and 6D. 2) Importantly, all phenotypes were restored when Lpp was knocked into SA113 Δlgt strain. Please see figure 3H-M 2, In figure 4, Authors use Newman(Agr -) strain instead of SA113(Agr +) for TLR2 Ko mice experiment. But there is no discussion about the impact of Agr sysmen. What is the impact of Agr system? : Once again, very good comment (review 1 and 4 also commented on this). Now we have added the discussion about the impact of Agr system in the manuscript.
Minor Queries 1, The authors showed CFU in Figure 3D, 4C, 5C, 6B and 7G, but those are hard to see the difference even though there is statistically significant difference. : The CFU data were presented in logarithmic scale Y axis, the significant differences were not so obvious in some cases. We have now tried to adjust the scales of Y axis in order to make the differences clearer. We hope the figures are now acceptable.
Reviewer #4 (Remarks to the Author): Staphylococcus aureus lipoproteins (Lpp) are membrane anchored surface proteins that play important roles in host-bacterial interactions. How Lpp involved in S. aureus pathogenicity remain poorly studied. Mohammad and colleagues showed in their previous study that Lpp induces chronic destructive macroscopic arthritis (PMID: 31226163). In this study, Mohammad and colleagues further investigate the role of Lpp, in particular "lipoprotein-like" Lpl1, in murine skin infection. They find that subcutaneous injection of Lpl1 promote leukocytes infiltration and skin lesion, which are TLR2-dependent. Also, in comparison to wild type strain, lgt-mutant strain causes smaller skin lesion size and lower bacterial loads, which the latter is TLR2-independent. Finally, they find that the skin lesions and bacterial burden induced by Lpp can be abrogated by chemically-depletion of leukocyte or fibrinogen.
This study is potentially interesting, particularly lpp-induced fibrin capsule and abscess formation as a strategy to evade host immune attack, and will be of interest to a broad audience of infectious disease researchers and microbiologists. However, the mechanism underlying the abscess formation to protect S. aureus from immune cells is less convincing, as the conclusion was made largely based on experiments using purified Lpl1 alone or coinjection with live bacteria. More rigorous examinations are required in this regard.
Below are comments/points that may help to improve the manuscript and hope the authors will find them useful.