Dual Role of WISP1 in maintaining glioma stem cells and tumor-supportive macrophages in glioblastoma

The interplay between glioma stem cells (GSCs) and the tumor microenvironment plays crucial roles in promoting malignant growth of glioblastoma (GBM), the most lethal brain tumor. However, the molecular mechanisms underlying this crosstalk are incompletely understood. Here, we show that GSCs secrete the Wnt‐induced signaling protein 1 (WISP1) to facilitate a pro-tumor microenvironment by promoting the survival of both GSCs and tumor-associated macrophages (TAMs). WISP1 is preferentially expressed and secreted by GSCs. Silencing WISP1 markedly disrupts GSC maintenance, reduces tumor-supportive TAMs (M2), and potently inhibits GBM growth. WISP1 signals through Integrin α6β1-Akt to maintain GSCs by an autocrine mechanism and M2 TAMs through a paracrine manner. Importantly, inhibition of Wnt/β-catenin-WISP1 signaling by carnosic acid (CA) suppresses GBM tumor growth. Collectively, these data demonstrate that WISP1 plays critical roles in maintaining GSCs and tumor-supportive TAMs in GBM, indicating that targeting Wnt/β-catenin-WISP1 signaling may effectively improve GBM treatment and the patient survival.


Comments
1. Figure1 (f) which GSC culture was used in the experiment is not elaborated? 2. In Figure 1 (g), the method for preparation of conditioned medium is not elaborated. Was it a TCA precipitation method or just the conditioned medium? Please include molecular weight for all the immune-blot experiments. 3. In Figure 1 (e), Since these GSCs were maintaining stemness through Wnt/B-catenin active signaling and WISP1 is a downstream target gene, please include B-catenin expression in the immunoblot. 4. The authors state in that the shWISP1 cells did not develop tumors at all, as there was no sign of bioluminescence. It is not clear then as to why the shWISP1 cells injected mice were dying within 60-70 days after injections as shown in figure 3(c)?. 5. Figure 4h, mention the day of imaging. 6. Figure 5a and 5b, blocking with integrin antibodies may also be an indirect effect of integrin a6B1 receptor function in relation to WISP1 and is not sufficient to make a claim that WISP1 is acting as a ligand for alpha6-beta1 receptor. To confirm that WISP1 is acting as a ligand for integrin a6B1 receptor, it may be important to perform interaction studies to show that WISP1 associates with integrin a6B1 receptor. Also it is necessary to show the specificity of interaction of WISP1 with the receptor using a rescue experiment by using both recombinant WISP1 protein along with antibody at different ratios for the phosphorylation of AKT (Ser473). e7 also recently showed that WISP1 is required for efficient muscle regeneration and controls the expansion and asymmetric commitment of muscle stem cells through Akt signaling. Thus, the role of WISP1 as activator of Akt signaling though not shown in glioma stem cells is already documented for other cell-types. Hence, this manuscript though shows rigor is not novel and hence may not be suitable for consideration for publication in this journal.
Reviewer #3 (Remarks to the Author); expert in macrophages and cancer: Tao et al. describe interesting new results regarding the role of WISP1 in promoting glioblastoma progression. A clear and novel mechanistic framework is provided, arguing that WISP1 is specifically produced by glioma stem cells and provides an autocrine survival signal. In addition, WISP1 would also strongly promote the survival of M2 TAMs. The authors may want to consider the following comments.

1)
The authors use the TCGA and Gravendeel databases to examine the expression of WISP1 and other target genes in GBM. It would be interesting to also rely on the recently published human GBM single-cell RNAseq dataset (Neftel et al. 2019 Cell) to assess for the expression of WISP1 (and its putative integrin receptors) across the four GBM cellular states at single-cell resolution.
2) An important claim of the manuscript is that WISP1 signals through the Integrin α6β1 receptor. However, I feel that this needs to be substantiated: a) The author could provide more direct evidence of WISP1-α6β1 interaction, for example via co-immunoprecipitation experiments or more quantitively via surface plasmon resonance or related techniques.
b) In addition, it is not clear why the integrin blocking studies were only performed in the WISP1 overexpression setting. Did the authors examine whether the addition of blocking antibodies inhibits GSC proliferation (without WISP1 overexpression), similar to what is seen in Fig 2b- c) The authors could also use their sha6 construct to silence integrin a6 in GSCs (similar to what they did for U937 cells in Sup Fig 10). This should in theory phenocopy the WISP1 silencing of GSCs.
d) The authors report that silencing integrin a6 inhibits the proliferation of M2-polarized U937 cells. However, it is not clear to me where the WISP1 is coming from in this setting. Are the U937 cells producing WISP1 themselves or was this added to these cultures (which is not mentioned)? If the pro-survival effect in the U937 cells stems from WISP1-α6β1 signaling, then why would just silencing α6β1 in the absence of WISP1 lead to reduced survival?
3) While the results of WISP1 on GSCs are convincing and important. I am more hesitant with the proposed effects of WISP1 on TAMs. First, the reported effects of WISP1 silencing on TAMs seem quite dramatic, with a 60% reduction in total TAMs. It seems as if blocking WISP1 is more effective in obtaining GBM TAM depletion than CSF1R blockade. Indeed, it is reported that blocking CSF1R -one the most important macrophage growth factors -does actually not reduce the total number of TAMs in preclinical GBM (Pyonteck et al Nat Med 2013), showing that the tumor microenvironment can provide compensatory growth and survival signals. Here, the loss of WISP1, which is said to be specifically expressed in GSCs, seems sufficient to deplete the majority of TAMs throughout the tumor. The authors should at least try to speculate on the mechanism: which signaling pathways are disrupted? When the density of TAMs is examined (Fig 6), it is not mentioned at which time point post GSC inoculation tumors were harvested. This is important since shWISP1 tumors grew much slower. Smaller tumors may have less (mature) TAMs, irrespective of paracrine WISP1 signaling. In the same line, WISP1 silencing may result in an altered tumor microenvironment (TME), which may attract less macrophages. Therefore, an alternative explanation for the lower macrophage density may be an altered TME (for example think of low vs. high grade tumors, where the latter contain significantly more TAMs), instead of a direct effect of WISP1 on TAM survival.
Second, the authors report that WISP1 very specifically augments survival of M2 TAMs, while it does not affect M1 TAMs. The macrophage field is increasingly realizing that the M1/M2 dichotomy in tumors (and other in vivo inflammation settings) is a major oversimplification. It needs to be taken into account that markers that are reported to adhere to M1 or M2 in one disease model may not necessarily do so in others (arguing for a spectrum model of macrophage activation, for example see Xue et al Immunity 2014). Additional complexity arises from the fact that the GBM TAM pool can exhibit a mixed ontogeny that partly dictates its transcriptional state and which again does not clearly adheres to M1/M2 (see Bowman et al 2016 Cell Reports, Chen et al 2017 Cancer Res.). Here, the authors use CD206 and CD11c as M2 and M1 markers, respectively. They report that around 60% of TAMs express CD206, while the additional 40% are CD11c+. To exemplify that relying on only a few markers can be problematic, consider mouse syngeneic GL261 GBM tumors, where the majority of TAMs are CD11c+, and a subset of CD11c+ TAMs co-express CD206 (for example see Peterson et al. 2016 PNAS). Therefore, in my opinion, in GL261 it would be problematic to just label CD11c+ TAMs as anti-tumoral M1. Of course, the xenografts reported in this manuscript may behave differently. In any case, to get a better understanding of the effect of WISP1 silencing on TAM heterogeneity in these xenografts, it would be very valuable if the authors were to perform a more in-depth analysis of the tumor myeloid cell pool, instead of relying on only a few markers in isolation. Multi-color flow cytometry can be very useful in this regard, especially when subsequently linked to an unbiased transcriptome analysis.

Responses to Reviewer's Comments
We thank all reviewers for the critical evaluation of our manuscript. We greatly appreciate the insightful comments and helpful suggestions from the reviewers. In response to their major comments, we have performed a large amount of additional experiments and extensively revised the manuscript to address the critical issues. We believe that our manuscript has been significantly improved and strengthened now. Below, please find our point-by-point responses to the reviewers' comments. We hope that our responses have adequately addressed the important issues raised by the reviewers.
Reviewer #1 (Remarks to the Author); expert in glioblastoma and mouse models: Response: We thank the reviewer for the helpful suggestion and agree that necessary controls are critical. We used two independent shRNAs (shWISP1-1 and shWISP1-2) in this study. We checked sequences of these two shRNAs and found that shWISP1-2 targets the 3'-end noncoding region (3'-UTR) of WISP1 mRNA to disrupt WISP1 expression. As the WISP1 overexpression construct does not contain the 3'-end non-coding sequence, we were able to simultaneously silence endogenous WISP1 and overexpress exogenous WISP1 in GSCs.
Immunoblot analysis demonstrated that ectopic expression of WISP1 (pCDH-WISP1) in GSCs rescued the decreased Akt phosphorylation (pAkt-Ser473) caused by knockdown of endogenous WISP1 (Please see Figure R1 below).  We have added this additional data in Supplementary Fig. 4c and described the results in our revised manuscript. Please see Line 215 at Page 9, the 4 th part in the "Results" section: "rWISP1 treatment also rescued the decreased Akt phosphorylation (pAkt-Ser473) caused by WISP1 disruption in a dose-dependent manner ( Supplementary Fig. 4c)."

Reviewer #1: 3) effects of inhibition of alpha 6 and beta 1 by neutralizing antibodies on p-Akt;
Response: We thank the reviewer for the suggestion. We have performed the additional experiment and found that inhibiting Integrin α6 or β1 by neutralizing antibody reduced Akt phosphorylation (Ser473) in GSCs (Please see Figure R3 below). However, inhibiting Integrin β4, the other binding partner of Integrin α6, had no effect on Akt phosphorylation (pAkt-Ser473) in GSCs (Please see Figure R3 below). Integrin blocking antibody (5 μg/ml) or isotype IgG control for 12 hours.
We have added the new data in Supplementary Fig. 5g and described the results in our revised manuscript. Please see Line 255 at Page 11, the 5 th part in the "Results" section: "Immunoblot analysis confirmed that inhibiting Integrin α6 or β1 by blocking antibody reduced Akt phosphorylation (pAkt-Ser473) in GSCs, while inhibiting Integrin β4 had no effect on Akt phosphorylation (pAkt-Ser473) ( Supplementary Fig. 5g)." Reviewer #1: 2. Although the rationale of focusing on alpha 6 beta 1 integrin was based on previous reports, the controls of blocking other possible integrins, in particular, possible partners of beta subunit that known associate with alpha 6 integrin should be included.

Response:
We thank the reviewer for the important suggestion. Integrin α6 forms heterodimers with Integrin β1 or β4 1, 2 . As shown in Figure R3 (above), inhibiting Integrin α6 or β1 by its neutralizing antibody reduced Akt phosphorylation (pAkt-Ser473) in GSCs, but inhibiting Integrin β4 has no effect on Akt phosphorylation (pAkt-Ser473). In addition, we examined the effects of inhibiting Integrin α6, β1 or β4 by blocking antibody on GSC proliferation and tumorsphere formation. The results showed that inhibiting Integrin α6 or β1 significantly decreased GSC proliferation and tumorshpere formation, while inhibiting Integrin β4 had no effect on GSC proliferation and tumorshpere formation (Please see Figure R4a, b below). Consistently, previous study has shown that Integrin β4 is not expressed in GSCs 2 . Thus, it is resonable that inhibiting Integrin β4 does not impact GSC proliferation.

Response:
We appreciate the critical concern. We have tried very hard to find the WISP1 inhibitor for this study, but there is no available WISP1 inhibitor so far. Thus, we had to target the Wnt/β-catenin, the upstream signaling of WISP1. We understand that Wnt/β-catenin signal may induce multiple downstream targets to promote tumor growth. Thus, the inhibition of GBM growth by carnosic acid may be a comprehensive result. Nevertheless, carnosic acid treatment reduced WISP1 expression in vitro and in vivo, suggesting that WISP1 inhibition may at least partially contribute to the therapeutic effect of carnosic acid. We have discussed it in the "Discussion" section. Please see Page 20, the 4 th paragraph in the "Discussion" section: "Because there is no available WISP1 inhibitor so far and Wnt/β-catenin signaling is activated in GSCs ……….
suggesting that WISP1 inhibition at least partially contribute the therapeutic effect of carnosic acid." In addition, blood-brain barrier (BBB) prevents most anti-cancer agents from penetrating GBM tumors and limit therapeutic efficacy 3 . As carnosic acid can penetrate the blood brain barrier well 4,5 , and it has been reported to improve the treatment of medulloblastoma in mouse models 5 , we selected carnosic acid to test its effect of on GSCs, TAMs and GBM tumor growth in our models.
We have described the reason of selecting carnosic acid for treatment in our revised manuscript.
Please see Page 16, the last part in the "Results" section: "We selected the small molecule carnosic acid in our preclinical study, because it can penetrate the blood brain barrier and it has been reported to improve the treatment of medulloblastoma in a mouse model." Furthermore, we performed immunoblot analysis and found that both total β-cantenin and active β-cantenin are enriched in all isolated GSC populations relative to matched non-stem tumor cells (NSTCs) (Please see Figure R5 below). We have added the new data in Fig. 1e and described the result in our revised manuscript.
Please see Page 6, the 1 th part in the "Results" section: "Immunoblot analysis showed that WISP1, active β-catenin, total β-catenin and GSC markers including SOX2 and OLIG2 were preferentially expressed in GSCs relative to matched NSTCs (Fig. 1e)." This results provode us with the rationality of targeting Wnt/β-catenin signaling for GBM treatment. We have discussed it in the "Discussion" section. Please see Page 20, the 4 th paragraph in the "Discussion" section: "Because there is no available inhibitor of WISP1 so far and Wnt/β-catenin signaling is activated in GSCs ………suggesting that WISP1 inhibition at least partially contribute to the therapeutic effect of carnosic acid." Response: We regret that the method for the conditioned medium was not elaborated. We collected conditioned media from GSCs or NSTCs cultured in the Neurobasal medium without supplements and growth factors, and then concentrated conditioned media by using the Eppendorf Concentrator plus / Vacufuge vacuum centrifugation system. We have added the detailed description in the "Methods" section in the revised manuscript. Please see Line 591 at Page 24 in the "Conditioned Medium Preparation" section. In addition, we have added the molecular weight for all immunoblots.  We have added the new data in Fig. 1e and described the results in our revised manuscript.
Please see Page 6, the 1 th part in the "Results" section: "Immunoblot analysis showed that WISP1, active β-catenin, total β-catenin and GSC markers including SOX2 and OLIG2 were preferentially expressed in GSCs relative to matched NSTCs (Fig. 1e)." Reviewer #2: 6. Figure 5a and 5b, blocking with integrin antibodies may also be an indirect effect of integrin a6B1 receptor function in relation to WISP1 and is not sufficient to make a claim that WISP1 is acting as a ligand for alpha6-beta1 receptor. To confirm that WISP1 is acting as a ligand for integrin a6B1 receptor, it may be important to perform interaction studies to show that WISP1 associates with integrin a6B1 receptor. Also it is necessary to show the specificity of interaction of WISP1 with the receptor using a rescue experiment by using both recombinant WISP1 protein along with antibody at different ratios for the phosphorylation of AKT (Ser473).

Response:
We thank the reviewer for the helpful suggestions.
(1) To validate that Integrin α6β1 is a receptor for WISP1, we performed co-immunoprecipitation (CoIP) assay to confirm their binding as suggested by the reviewer. To increase the potential binding amounts, we overexpressed WISP1 in GSCs and then performed CoIP with anti-Integrin α6 or β1 antibody. Anti-Integrin α6 antibody pulled down the Integrin α6 along with WISP1 and Integrin β1 (Please see Figure R6a   "To validate that Integrin α6β1 is a receptor for WISP1, we performed co-immunoprecipitation (CoIP) assay to confirm their binding……….In addition, the anti-Integrin β1 antibody also pulled down the Integrin β1 along with WISP1 and Integrin α6 ( Fig. 5m and Supplementary Fig. 5l)." (2) To test the specificity of the interaction between WISP1 and the receptor Integrin α6β1, we examined Akt phosphorylation in GSCs treated with recombinant human WISP1 (rWISP1) protein along with Integrin blocking antibody at different ratios. Immunoblot analysis revealed that 5 μg/ml Integrin α6 or β1 blocking antibody dramatically prevented 0.2 μg/ml rWISP1-induced Akt phosphorylation (pAkt-Ser473), while this dose of antibody had a relatively little effect on 0.8 μg/ml rWISP1-induced Akt phosphorylation (pAkt-Ser473) (Please see Figure R7 below).
The results showed that Integrin α6β1 is relatively specific to WISP1. Response: We appreciate the critical concern. We have addressed the similar issue in Comment #4. Since the luciferase signals from the brains of mice bearing the xenografts expressing shWISP1 were very weak within 30 days after transplantation, we could not detect obvious signals under bioluminescent imaging on day 14 and day 21 as shown in Figure 3a, although one of these mice in the shWISP1-1 group showed the luciferase signal at day 21. This did not indicate that there were no micro-tumors in the brains of mice bearing xenografts expressing shWISP1. GSCs expressing shWISP1 proliferate slowly and need longer time to develop tumors in mouse brains. Indeed, the group of mice bearing xenografts expressing shWISP1 developed detectable tumors within 40-70 days after transplantation. Thus, the mice bearing xenografts expressing shWISP1 still developed large tumors although they took a relatively longer time ( Figure 3c). We collected the brains bearing GBM xenografts from mice when neurological signs occur. This time point is usually two to three days before the death of mice and the size of tumors is large at this time point. Although we collected tumors from the control group (shNT) and shWISP1 groups at different times after transplantation, tumor sizes from these three groups (shNT, shWISP1-1 and shWISP1-2) were similar. Thus, these tumors from control and experimental groups were comparable for further analyses shown in Figure 6. We have described the collection time of tumor xenografts in the Legend part of Figure 6 (Page 43) and "Methods" section (Page 25) in our revised manuscript.

Response:
We are sorry for that. We have tried our best to rearrange images or figure panels in sequential manner. Due to too many data panels in each figure, it is really hard to have all images or panels arranged sequentially in all figures. We sincerely hope that the reviewer understands the situation. Response: We appreciate the evaluation and helpful suggestions by the reviewer. We understand that the regulation of Akt signaling by WISP1 has been reported in other cell types.
However, our study focused on the dual role of WISP1 in promoting both GSC and M2 TAM maintenance, thus supporting malignant growth of GBM. The role of WISP1 in maintaining the tumor-supportive TAMs (M2) to promote GBM tumor growth has not been reported. This is one of novel points of this manuscript. In addition, we first found that Integrin α6β1 is the receptor of WISP1 in both GSCs and M2 TAMs in GBMs, and identified that WISP1-α6β1-Akt signaling is responsible for GSC-promoted survival of TAMs in tumor microenvironment, which is another new point. Furthermore, we found targeting the Wnt/β-catenin-WISP1 signaling with carnosic acid potently inhibited GBM tumor growth and extended the survival of tumor-bearing mice, suggesting that targeting this signaling axis represents an attractive therapeutic strategy.
Therefore, we believe that our findings in this manuscript are significant and contain novel points, which will make it interesting to general readers. Response: We thank the reviewer for the time and effort to review our manuscript. We appreciate the concerns and suggestions raised by the reviewer. We have performed a large amount of additional experiments to address the reviewer's concerns. We believe that this manuscript is significantly improved after extensive revision in response to the constructive suggestions.

Reviewer #3: 1) The authors use the TCGA and Gravendeel databases to examine the expression of WISP1 and other target genes in GBM. It would be interesting to also rely on the recently published human GBM single-cell RNAseq dataset (Neftel et al. 2019 Cell) to assess for the expression of WISP1 (and its putative integrin receptors) across the four GBM cellular states at single-cell resolution.
Response: We thank the reviewer for the important suggestion. We examined the expression of  We have added the new data in Supplementary Fig. 5n and described the results in our revised manuscript. Please see Line 283 at Page 12, the 5 th part in the "Results" section: "A recent study has shown that malignant cells in human GBM exist in four main cellular states that recapitulate neural-progenitor-like (NPC-like), oligodendrocyte-progenitor-like (OPC-like), astrocyte-like (AClike), and mesenchymal-like (MES-like) states…….These data suggest that WISP1 and Integrin α6β1 are co-expressed by some AC-like and MES-like cells in GBM (Supplementary Fig. 5n)." Reviewer #3: 2) An important claim of the manuscript is that WISP1 signals through the Integrin α6β1 receptor. However, I feel that this needs to be substantiated: a) The author could provide more direct evidence of WISP1-α6β1 interaction, for example via coimmunoprecipitation experiments or more quantitively via surface plasmon resonance or related techniques.

Response:
We thank the reviewer for the insightful suggestion. To confirm that Integrin α6β1 is a receptor for WISP1, we performed co-immunoprecipitation (CoIP) assay to confirm their binding.
To increase the potential binding amounts, we overexpressed WISP1 in GSCs and then performed CoIP with anti-Integrin α6 or β1 antibody. The anti-Integrin α6 antibody pulled down the Integrin α6 along with WISP1 and Integrin β1 (Please see Figure R6a   "To validate that Integrin α6β1 is a receptor for WISP1, we performed co-immunoprecipitation (CoIP) assay to confirm their binding……….In addition, the anti-Integrin β1 antibody also pulled down the Integrin β1 along with WISP1 and Integrin α6 ( Fig. 5m and Supplementary Fig. 5l)." Fig 2b-d when silencing WISP1? Response: We thank the reviewer for raising this important point. Integrin α6 forms heterodimers with Integrin β1 or β4 1, 2 . We also used Integrin β4 blocking antibody to perform the experiment following the suggestion from another reviewer. We examined the effects of inhibiting Integrin α6, β1 or β4 by blocking antibody on GSC proliferation and tumorsphere formation. The results showed that inhibiting Integrin α6 or β1 significantly decreased GSC proliferation and tumorshpere formation, while inhibiting Integrin β4 had no effect on GSC proliferation and tumorshpere formation (Please see Figure R4a, b below). Consistently, previous study has shown that Integrin β4 is not expressed in GSCs 2 . Thus, it is resonable that inhibiting Integrin β4 showed no effect on GSC proliferation. Figure R4. a, Cell viability assay of GSCs treated with Integrin α6 blocking antibody (5 μg/ml) or isotype IgG for 6 days. b, Tumorsphere formation of of GSCs treated with Integrin α6 blocking antibody (5 μg/ml) or isotype IgG for 6 days. Data are shown as means ± s.d. ***p＜0.001, twotailed unpaired t-test.

Reviewer #3: 2)-(b) In addition, it is not clear why the integrin blocking studies were only performed in the WISP1 overexpression setting. Did the authors examine whether the addition of blocking antibodies inhibits GSC proliferation (without WISP1 overexpression), similar to what is seen in
We have added the new data in Fig. 5g, h and Supplementary Fig. 5e, f and described the results in our revised manuscript. Please see Page 10, the 5 th part in the "Results" section: "Moreover, treatment of GSCs with Integrin α6 or β1 blocking antibody significantly decreased GSC proliferation and tumorsphere formation (Fig. 5g, h and Supplementary Fig. 5e, f). However, blocking Integrin β4, the other binding partner of Integrin α6, had no effect on GSC proliferation and tumorsphere formation (Fig. 5g, h and Supplementary Fig. 5e, f)." We also examined Akt phosphorylation (pAkt-Ser473) in GSCs treated with Integrin blocking antibody. The results showed that inhibiting Integrin α6 or β1 by blocking antibody reduced Akt phosphorylation (pAkt-Ser473) in GSCs, while inhibiting Integrin β4, the other binding partner of Integrin α6, had no effect on Akt phosphorylation (pAkt-Ser473) in GSCs (Please see Figure R3 below). Figure R3. Immunoblot analysis of Akt phosphorylation (pAkt-Ser473) in GSCs treated with Integrin blocking antibody (5 μg/ml) or isotype IgG control for 12 hours.
We have added the additional data in Supplementary Fig. 5g and described the results in our revised manuscript. Please see line 255 at Page 11, the 5 th part in the "Results" section: "Immunoblot analysis confirmed that inhibiting Integrin α6 or β1 by blocking antibody reduced Akt phosphorylation (pAkt-Ser473) in GSCs, while inhibiting Integrin β4 had no effect on Akt phosphorylation (pAkt-Ser473) (Supplementary Fig. 5g)."

Reviewer #3: 2)-(c) The authors could also use their sha6 construct to silence integrin a6 in
GSCs (similar to what they did for U937 cells in Sup Fig 10). This should in theory phenocopy the WISP1 silencing of GSCs.

Response:
We appreciate the helpful suggestion. We examined the effects of Integrin α6 disruption by shRNA on GSC proliferation and Akt phosphorlation. shRNAs targeting α6 significantly decreased Integrin α6 expression and Akt phosphorylation (pAkt-Ser473) in GSCs (Please see Figure R9a below). Disruption of Integrin α6 also significantly inhibited GSC proliferation and tumorsphere formation (Please see Figure R9b, c below).   showing that the tumor microenvironment can provide compensatory growth and survival signals.
Here, the loss of WISP1, which is said to be specifically expressed in GSCs, seems sufficient to deplete the majority of TAMs throughout the tumor. The authors should at least try to speculate on the mechanism: which signaling pathways are disrupted? When the density of TAMs is examined (Fig 6), it is not mentioned at which time point post GSC inoculation tumors were harvested. This is important since shWISP1 tumors grew much slower. Smaller tumors may have less (mature) TAMs, irrespective of paracrine WISP1 signaling. In the same line, WISP1 silencing may result in an altered tumor microenvironment (TME), which may attract less macrophages.
Therefore, an alternative explanation for the lower macrophage density may be an altered TME The reviewer mentioned that silencing WISP1 may result in an altered tumor microenvironment, which may contribute to decreased TAM number. According to our data, we cannot rule out this possibility. However, our in vitro data suggest that WISP1 has a direct effect on the survival of macrophages. It would be interesting to further investigate whether WISP1 can regulate the tumor microenvironment in GBM in the future study. We have discussed these issues in the "Discussion" section. Please see Line 454 at Page 18, the 3 rd paragraph in the "Discussion"

Response:
We appreciate the important concern raised by the reviewer. We agree with the reviewer that the M1/M2 dichotomy is an oversimplification of TAMs in tumors. We also think that there is a heterogeneity of TAMs in GBM tumors. In this study, we used the term "M2 TAMs" to indicate the tumor-supportive macrophages that may contain several subpopulations, and used "M1 TAMs" to indicate the tumor-suppressive macrophages that may also contain several subpopulations. The M1/M2 dichotomy used in this manuscript does not mean that there are only two types of TAMs in GBM tumors. We also agree that M1 or M2 markers in one disease model may not necessarily be the same in other disease models, as the previous work showing that the majority of TAMs are CD11c + , and a subset of CD11c + TAMs co-express CD206 6 . In our study, we used CD206 and CD163 as M2 TAM markers, and CD11c and CD16/32 as M1 TAM markers.
According to our previous results, CD206 + and CD163 + TAMs may represent one major subpopulation of TAMs, and CD11c + and CD16/32 + TAMs are the another major sub-population. To further verify this point, we performed immunofluorescent staining in GBM xenografts using these markers. We found that more than 90% CD206 + TAMs express CD163 (Please see Figure R12a, b below), and more than 90% CD11c + TAMs express CD16/32 (Please see Figure R12c, d below). These data indicate that CD206 + and CD163 + TAMs are almost the same population, and CD11c + and CD16/32 + TAMs are nearly the same population.  We next performed immunofluorescent staining in GBM xenografts using CD206 and CD11c antibodies and found that less than 6% CD206 + TAMs express CD11c (Please see Figure R13 below). This data further confirm that CD206 + /CD163 + and CD11c + /CD16/32 + TAMs represent very different sub-populations of TAMs. Our previous work has demonstrated that CD163 + TAMs promote GBM tumor growth in our xenograft models (T4121 and T387 GSC-derived xenografts) 7,8 . Therefore, these studies further confirm that silencing WISP1 indeed reduced the number of tumor-supportive macrophages (M2 TAMs) in our xenograft models. According to these results, we think that M2/M1 TAMs indeed represent two major but functionally different macrophage populations (Tumor-supportive and tumor-suppressive macrophages) in our tumor models, although we can't rule out that each major population (M2 or M1) may contain several subpopulations.
Taken together, all these data demonstrated that silencing WISP1 indeed decreased tumorsupportive M2 macrophages in GSC-derived xenografts, while had little effect on tumorsuppressive M1 macrophages. We believe that our results will provide some useful information for therapeutic targeting of tumor-supportive macrophages (M2 TAMs) in GBMs. We have added these new data in Supplementary Fig. 7c-h and Supplementary Fig. 8g-l and revised the description of this results in our revised manuscript. Please see Line 320 at Page 13, the 6 th part in the "Results" section: "We used specific M2 markers (CD206, CD163, Arg1 and Fizz1) and M1 markers (CD11c, CD16/32, iNOS and MHCII) for the study, as those markers have been used to distinguish M2/M1 TAMs in our GBM xenograft models and some other GBM xenograft models." In addition, we added the following in the Discussion part at Page 18 (Line 441): "We fully recognized that the M1/M2 dichotomy is an oversimplification of TAMs in tumors. In this study, we just used the term "M2 TAMs" to indicate the tumor-supportive macrophages that may contain several subpopulations, and used "M1 TAMs" to represent the tumor-suppressive macrophages that may also contain subpopulations. The M1/M2 dichotomy used here does not mean that there are only two simple types of TAMs in GBM tumors. We believe that there is a heterogeneity of TAMs in GBM tumors. However, our studies confirmed that silencing WISP1 indeed reduced tumor-supportive macrophages (M2 TAMs) in our xenograft models. According to our previous studies and current data, it is reasonable to conclude that M2/M1 TAMs indeed represent two major but functionally different macrophage populations (Tumor-supportive and tumorsuppressive macrophages) in our tumor models, although we can't rule out that each major population (M2 or M1) may contain several subpopulations. It will be interesting to further analyze subpopulations in M2 TAMs and M1 TAMs in GBMs in the future."

Response:
We are sorry for missing the important reference. We have discussed it in the "Discussion" section and cited it. Please see Page 17, the 2 nd paragraph in the "Discussion" section: "Recent study has also demonstrated that WISP1 is a novel oncogene in GBM……..the origin of WISP1 in GBM and the role of WISP1 in regulating of GSC properties remain unclear."

REVIEWERS' COMMENTS:
Reviewer #1 (Remarks to the Author): In the revised manuscript, the authors have satisfactorily addressed all the comments I raised with strong new data and/or reasonable explanation and corresponding changes in the result descriptions and/or discussions. This is a much improved study with further strengthened significance. This study is sufficient for its publication in Nature Communications.