Innate scavenger receptor-A regulates adaptive T helper cell responses to pathogen infection

The pattern recognition receptor (PRR) scavenger receptor class A (SR-A) has an important function in the pathogenesis of non-infectious diseases and in innate immune responses to pathogen infections. However, little is known about the role of SR-A in the host adaptive immune responses to pathogen infection. Here we show with mouse models of helminth Schistosoma japonicum infection and heat-inactivated Mycobacterium tuberculosis stimulation that SR-A is regulated by pathogens and suppresses IRF5 nuclear translocation by direct interaction. Reduced abundance of nuclear IRF5 shifts macrophage polarization from M1 towards M2, which subsequently switches T-helper responses from type 1 to type 2. Our study identifies a role for SR-A as an innate PRR in regulating adaptive immune responses.

10. On pg. 8 the authors finally get to the issue of IL-12, which might be responsible for the granuloma phenotype in Figure 1 (e.g. Wynn's data on IL-12 and schistosome granulomas). However, the data here is hard to interpret without enumeration of IL-12 within the granulomas. The experiments shown are indirect and inconclusive.
11. The data concerning IRF5 regulation is too preliminary in nature to interpret. There is no compelling reason to imagine a cell surface scavenger receptor would regulate and archetypal member of the IRF family. A co-crystal structure, or compelling biochemical evidence would be essential.

Reviewer 2 (Remarks to the Author):
The manuscript by Xu et al. is exciting because of novel data identifying a molecular mechanism for macrophage polarization and also because of new information about the role of alternatively activated macrophages in type 2 pathology. The data suggests that M2 polarization is critical for granuloma development and fibrosis in response to S. japonicum at the time point shown.
The authors support their conclusions effectively with well-controlled experiments. They show that macrophages from SR-A KO mice do not exhibit normal polarization to an M2 phenotype in the presence of type 2 stimuli. They explain that the immune dysregulation is due to SR-A from macrophages using adoptive transfer and siRNA transfection/overexpression approaches. They also use sh-IRF5 to demonstrate that IRF5 is necessary for the SR-A KO phenotype.
While I find the data convincing, I have a few comments and questions.  Response: We have followed the reviewer's suggestion to delete the "important" and "critical".
2. In figure 1, which is hardly discussed in the results section, the authors demonstrate the SR-A-deficient mice (proper genetic nomenclature not used, presumably Msr1) have smaller granulomas (that is not explained or quantified properly) and die earlier than controls. These data are questionable. The authors state in the legend that 12 mice were used per group/experiment, and it was done twice. Is the data in 1a the combined data or one experiment? In either case, the 'significance' may be mathematically true, but is biological irrelevant, as both infected groups die between 15-20 weeks. In 1b the authors state they quantified 30 granulomas -is this from 1 mouse, or 30 from all mice? Essentially, and without going in to too much detail, this experimental series was not performed or analyzed properly, and the data cannot be interpreted.

Response:
We are sorry that we had failed to describe the results in detail nor discuss more deeply. In Results section we had only described the granuloma surround eggs in livers and intestines of S. japonicun-infected SR-A -/mice, but had not mentioned more pathology else in detail. In our study, in addition to smaller granuloma in both livers and intestines, the SR-A -/mice showed that the integrity of the gut epithelium was compromised, and the epithelial barrier was broken (Fig 1f), which facilitated the development of lethal endotoxemia in SR-A-/-mice ( Supplementary Fig. 1d). Consistently, studies 1, 2 have already shown that although the egg granuloma size in intestine is milder, stronger M1/Th1 responses (just like what happened in SR-A -/mice, which was shown in Fig. 2a-2b and 3a-3c) significantly lead to breaks in the epithelial barrier that causes septicemia as increased numbers of lumen bacteria and/or toxin penetrate the breach. In conclusion, in our study, a decreased egg granuloma in livers and intestines accompany with severer broken epithelial barrier (Fig. 1f), higher endotoxin-mediated toxicity ( Supplementary Fig. 1d), stronger M1/Th1 responses ( Fig. 2a-2b, 3a-3c), and rapidly succumb to S. japonicum infection (Fig 1a) were showed in the SR-A -/mice. We have modified the Results section (page 6 lines 6-10) and discussed more in our revised manuscript in section of Discussion on page 14 lines 13-22.
Following the reviewer's suggestion, we have added the genetic nomenclature, Msr1, of the SR-A -/mice in Materials and Methods in our revised manuscript on page 20 line 3. Fig. 1a had been repeated twice with similar results but we had only shown a representative experiment (12 mice per group) in two experiments in our previous manuscript. We have combined the two sets of data in new Figure 1a (24 mice per group) in our revised manuscript.

Experiment in
Given that one worm pair of S. japonicum laid approximately 2,200 eggs per day 3 , it is far (up to six folds) more than Schistosoma mansoni worm pairs. In our study, each mouse was infected with 12 cercariae of S. japonicum, which is an extremely severe infection for mice, if considering worm number/kg host weight (equal to 24000-30000 worms in a person). Therefore, it is impossible for mice to survive too long (or even for ever) without immediate Praziquantel treatment. In this case, the extension of mice survival span from 15 to 20 weeks is biologically/medically significant and relevant. The similar situation was also happened in field of cancer treatment research or some other fields.  3. Line 3, pg. 3. 'or in the cytoplasm of macrophages'. This is not sufficiently specific.
Response: Given that SR-A expressed on the Golgi apparatus of macrophages 1, 2 , we have descript this sentence more specific as described below. 'or on the Golgi apparatus of macrophages' in our revised manuscript page 3 line3.  7. pg. 5, line 22. 'which was reported to elicit a Th1 response' seems unnecessary.
Response: As suggested by the reviewer, we have removed this sentence.
8. The data concerning the heat maps in figure 3 are impossible to interpret. The authors need to test macrophage polarization from macrophages isolated (or stained) from granulomas. Why were peritoneal macrophages used? There are no criteria for 'differences' between genotypes, and it is not clear how reproducible the data was.
Why wasn't IL-4 used as a control?
Response: Although it is techniquely much more difficult for us to isolate and obtain sufficient number of live macrophages with high-purity, especially from uninfected control mice, both mice liver and peritoneal CD16/32-macrophages and CD206macrophages had been FACS analyzed and similar results had been obtained. As suggested by the reviewer, we have replaced with the RT-PCR results with liver macrophages (Figure 3c), which showed that the expression of M1 related genes was increased but M2 related genes were decreased in liver macrophages of SR-A −/− mice compared to those of WT mice after S. japonicum infection.
We had detected the level of IL-4 but we had put it in Fig. 4f of previous manuscript as a separate figure. Following the reviewer's suggestion, we have added the control data of IL-4 both in the Figure 3c and 3d of our revised manuscript.
9. pg. 7. A mixed bone marrow chimera is required to understand the SR-A intrinsic versus extrinsic effects. The siRNA experiments are difficult to interpret, if not impossible.
Response: We are sorry that we are unable to carry out the mice experiment with mixed bone marrow chimera under the limitation of our experimental conditions. As a replacement, we have carried out additional adoptive transfer experiments to understand the SR-A intrinsic versus extrinsic effects. Results in Supplementary   Fig.4g-n in our revised manuscript showed that transferring of SR-A -/macrophages to S. japonicum-infected WT mice dampened the granuloma size and Th2 response, but increased Th1 response both in spleen and liver, indicated a SR-A intrinsic role for macrophages in the development of liver granuloma. However, transferring of WT macrophages to S. japonicum-infected SR-A -/mice increased the liver granuloma size and Th2 response, but decreased Th1 response. We have added these results in our revised manuscript in page 9 lines 7-15.
10. On pg. 8 the authors finally get to the issue of IL-12, which might be responsible for the granuloma phenotype in Figure 1 (e.g. Wynn's data on IL-12 and schistosome granulomas). However, the data here is hard to interpret without enumeration of IL-12 within the granulomas. The experiments shown are indirect and inconclusive.
Response: As suggested by the reviewer, we have detected the expression of IL-12 in liver homogenates by ELISA. Results showed that the expression of IL-12 in the liver of SR-A -/mice was higher than that in WT mice. We have added the result in Supplementary Fig. 5g 11. The data concerning IRF5 regulation is too preliminary in nature to interpret. There is no compelling reason to imagine a cell surface scavenger receptor would regulate and archetypal member of the IRF family. A co-crystal structure, or compelling biochemical evidence would be essential.
Response: Studies have shown that SR-A not only expressed on the membrane of macrophages, but also in the cytoplasm of macrophages because: 1) the cell surface SR-A proteins are activated upon ligand binding and internalized into cytoplasm via coated pits and traffic the endocytic pathway, 2) recycling between early endosomal compartments and the cell surface 1, 2, 3 . Consistent with our result, studies have shown that many other cytolpasmic molecules, such as TRAF6 or polyethylene glycol cholesteryl ester (PEG-Chol), also interact with SRA in the cytoplasm 3,4 .We have added these in the section of Discussion in page 18 line 5-7. In addition, in our study, SR-A was detected both on the membrane (by FACS) (Supplementary Fig.8g, 8i, and 8n) and in the cytoplasm (by confocal) (Figure7e, Supplementary Fig.8c and 8j), and a SR-A-IRF5 interaction by confocal within the cytoplasm of macrophages in Figure7e and Supplementary Fig.8g.
It is a pity that we are not able to complete the co-crystalstructure analysis by our current technology and condition. Instead, we used STRINE database (http://string.embl.de/) to predict the possible interaction of SR-A and IRF5 and found that there is a possibility of two protein interactions (as follows). However, it needs further studied.
The software predicted that SR-A and IRF5 could be combined

Reviewer 2 (Remarks to the Author):
The manuscript by Xu et al. is exciting because of novel data identifying a molecular mechanism for macrophage polarization and also because of new information about the role of alternatively activated macrophages in type 2 pathology. The data suggests that M2 polarization is critical for granuloma development and fibrosis in response to S. japonicum at the time point shown.
The authors support their conclusions effectively with well-controlled experiments. They show that macrophages from SR-A KO mice do not exhibit normal polarization to an M2 phenotype in the presence of type 2 stimuli. They explain that the immune dysregulation is due to SR-A from macrophages using adoptive transfer and siRNA transfection/overexpression approaches. They also use sh-IRF5 to demonstrate that IRF5 is necessary for the SR-A KO phenotype.
While I find the data convincing, I have a few comments and questions. 1) Macrophages play major roles as sentinels for first line alerts and as mediators that shape the adaptive immune responses during infection. SR-A is constitutively expressed in most macrophages (not only limited to M1 and M2, but also BMDMs), in addition, the expression level of SRA could be regulated 1, 2, 3, 4 . We have modified our Introduction section on page 4 lines 3-5.
2) As suggested by the reviewer, we have carried out additional experiments regarding M2 subsets. Results showed that S. japonicum infection significantly induced a M2b-dominant macrophages (TNF-α high , IL-10 high , MR), but not M2a or M2c (TNF-α low , IL-10 low , IL-6 low ) in WT mice. Interestingly, SR-A-deficiency resulted in phenotypic M1-dominant (TNF-α high , IL-10 low , MR low , IL-12 high ) macrophages present in the liver and peritoneal cavity during S. japonicum infection. We have added these results in our manuscript page 7 lines 16-20.
3) Yes, it is very likely that there exists the SR-A/IRF5-independent mechanism of polarization towards an M2 phenotype, as we had observed that the blocking of SR-A/IRF5 interaction (eg. through knockout of SR-A in Fig. 3) only resulted in a partial, not a complete inhibition of M2 differentiation. However, the underlying mechanism needs further study.

4)
We had also found that SR-A-deficiency induced M1-dominant macrophages in heatinactivated M. tuberculosis immunized mice (Supplementary Fig.3b). In addition, study shows that SR-A-deficiency also results in increased M1-dominant macrophages in myocardial infarction-induced cardiomyocyte necrosis model 5 . All these results suggest this phenomenon may not be S. japonicum specific.