LOXL1 Is Regulated by Integrin α11 and Promotes Non-Small Cell Lung Cancer Tumorigenicity
Round 1
Reviewer 1 Report
Manuscript ID: cancers-496345
Comments:
The manuscript by Zeltz et al., describes the potential role of LOXL1 in non-small cell lung cancer progression and the correlation between LOXL1 and integrin a-11 in lung adenocarcinoma patient datasets. Below are some comments which could benefit the study.
Major comments:
1) Both LOXL1 and Integrin a-11 are associated with tumor metastasis and thus in turn promote tumor progression. In this scenario, it essential to perform these studies in a metastatic NSCLC model rather than xenograft NSCLC model to better evaluate the role of LOXL1.
2) The authors have evaluated the correlation between LOXL1 and integrin a-11 in 3 lung adenocarcinoma patient datasets which showed inconsistent correlation between ITGA11 and LOXL1. The correlation is further evaluated in mouse embryonic fibroblast. It essential to determine the correlation with respect to the protein levels of LOXL1 and integrin a-11 in these studies to have a clearer understanding.
3) Similarly, the integrin a-11 expression levels (m-RNA and protein) have to be evaluated in the mouse tumor model with Loxl1+/+, Loxl1+/- and Loxl1-/-.
Minor comments:
1) Figure 2: CAF094-shLOXL1#1 wherein the LOXL1 knockdown was not significant in case of secreted protein should be removed from the graphs as it does not add any additional information to the figure.
2) Data showing knockdown of integrin a-11 in MEF cells should be included in the supplementary section.
3) The western blot of Figure 2b, should be added with a house-keeping control protein expression.
4) It is essential to have scale bars on all individual images in Figure 2,3, 4 and 5.
5) Please describe how predicted tumor volumes are calculated in the methods section in detail.
6) Statistics need to be added to Figure 5a and 5b for comparison of tumor volumes at Day 40.
7) The figure legend or methods section need to mention n number for all graphs with statistics.
8) Figure 5c, d – is it the wild type A549 cell line. Please mention.
9) Does knockdown of LOXL1 prolong the survival in tumor bearing mice? A Kaplan-Meier survival plot needs to be included.
Author Response
Reviewer#1
We thank the reviewer for his useful comments. Below we discuss the comments and the changes done in the manuscript (For the figures that we have included in the response, please refer to the uploaded PDF file).
The manuscript by Zeltz et al., describes the potential role of LOXL1 in non-small cell lung cancer progression and the correlation between LOXL1 and integrin a-11 in lung adenocarcinoma patient datasets. Below are some comments which could benefit the study.
Major comments:
1) Both LOXL1 and Integrin a-11 are associated with tumor metastasis and thus in turn promote tumor progression. In this scenario, it essential to perform these studies in a metastatic NSCLC model rather than xenograft NSCLC model to better evaluate the role of LOXL1.
Response: We have previously reported that integrin α11 promotes both NSCLC tumor growth and metastasis. Therefore, we are interested in analyzing the role of LOXL1 in NSCLC tumor growth. We are also interested in studying the role of Loxl1 in metastasis similar to what we did for integrin α11. However, Loxl1-/- SCID mice develop rectal prolapse that often leads to death of the animals prior to the study endpoint. It has been delicate to complete the tumor xenograft study due to this problem, thus limiting our ability to study the metastatic question.
2) The authors have evaluated the correlation between LOXL1 and integrin a-11 in 3 lung adenocarcinoma patient datasets which showed inconsistent correlation between ITGA11 and LOXL1. The correlation is further evaluated in mouse embryonic fibroblast. It essential to determine the correlation with respect to the protein levels of LOXL1 and integrin a-11 in these studies to have a clearer understanding.
Response: We thank the reviewer for his comment. However, in contrast to the comment, our data showed a strong and consistent correlation between ITGA11 and LOXL1 in three lung adenocarcinoma patient datasets, which was not seen with the other LOX family members. We agree with the reviewer that it would be desirable to confirm this correlation at the protein level. However, there is no available protein data in the three datasets. Moreover, the LOXL1 antibody that we have does not work for IHC. We attempted to analyze Loxl1 protein expression by Western blotting in mouse embryonic fibroblasts but the LOXL1 antibody that should recognize both human and mouse LOXL1 (based on its datasheet) did not get any specific band at the expected size from MEF lysates (data not shown).
3) Similarly, the integrin a-11 expression levels (m-RNA and protein) have to be evaluated in the mouse tumor model with Loxl1+/+, Loxl1+/- and Loxl1-/-.
Response: We have analyzed the expression of integrin α11 at both mRNA and protein levels in some xenograft tumors from Loxl1+/+, Loxl1+/- and Loxl1-/- SCID mice (see Figure 1 below). We found that integrin α11 expression in these tumors was independent of Loxl1 expression, indicating that regulation of LOXL1 expression by integrin α11 is not reciprocal.
Figure 1: Analysis of integrin α11 expression in xenograft tumors from Loxl1+/+, Loxl1+/- and Loxl1-/- mice. (a) Levels of integrin α11 protein were analyzed by Western blot. b-actin was used as loading control. (b) Itga11 expression has been analyzed by RT-qPCR in A549 (circle) and HCC4006 (triangle) xenografts from Loxl1+/+ (green), Loxl1+/- (red) and Loxl1-/- (blue) mice. Loxl1 mRNA expression confirmed the genotype of the three groups of loxl1 SCID mice.
Minor comments:
1) Figure 2: CAF094-shLOXL1#1 wherein the LOXL1 knockdown was not significant in case of secreted protein should be removed from the graphs as it does not add any additional information to the figure.
Response: We appreciate the reviewer’s concern. However, we believe including the shLOXL1#1 result is useful to demonstrate the varying efficiency of two shRNAs both at both mRNA and protein levels (Figure 2 a, b and c). This is in the context of effect of the specific shLOXL1#2 leading to a significant decrease in both mRNA and protein levels of LOXL1 in the functional gel contraction assay (Figure 2d).
2) Data showing knockdown of integrin a-11 in MEF cells should be included in the supplementary section.
Response: Expression of integrin α11 in α11+/+ MEFs, α11-/- MEFs and α11-/- KI MEFs has previously been characterized (Lu et al., PMID: 25076207). We have now added this reference to the manuscript. We have confirmed the expression of integrin α11 in these cells by RT-qPCR (see Figure 2 below).
Figure 2: Analysis of integrin α11 expression in MEFs by RT-qPCR. (a) Absence of Itga11 in α11-/- MEFs have been confirmed by RT-qPCR. (b) ITGA11 expression has been analyzed by RT-qPCR in MEF α11-/- KI using primers specific to human ITGA11.
3) The western blot of Figure 2b, should be added with a house-keeping control protein expression.
Response: There are no house-keeping proteins that are secreted and that could be used as a loading control. Instead, to normalize the expression of secreted proteins, the common practice is to stain the blot membrane with either Ponceau S or Coomassie blue, of which we used the latter.
4) It is essential to have scale bars on all individual images in Figure 2,3, 4 and 5.
Response: We realized that we have not incorporated scale bars in some of the images. We apologize for this and have now included scale bars in all the images.
5) Please describe how predicted tumor volumes are calculated in the methods section in detail.
Response: We agree with the reviewer that this method was missing. Predicted log tumor volumes were calculated using mixed effects modeling with the main effects Day, Loxl1 type, and their interaction, and a random intercept for each mouse. The predicted log tumor volumes were transformed to predict the tumor volume. We have now added this description in the method section.
6) Statistics need to be added to Figure 5a and 5b for comparison of tumor volumes at Day 40.
Response: The statistics related to tumor growth of HCC4006 and A549 in Loxl1+/+, Loxl1+/- and Loxl1-/- SCID mice were indicated in Supplementary Tables 4 and 5. We have now added the comparison of tumor volumes at day 40 (see Figure 3 below). These data have been included in supplementary data.
Figure 3: Analysis of tumor volume at day 40 of HCC4006 (n=11-16) and A549 (n=5-7) xenografts in Loxl1+/+, Loxl1+/- and Loxl1-/- mice. The differences between these groups were tested using Mann-Whitney test.
7) The figure legend or methods section need to mention n number for all graphs with statistics.
Response: We apologize for this omission and now have added "n" number for all graphs with statistics.
8) Figure 5c, d – is it the wild type A549 cell line. Please mention.
Response: We only used A549 wild type cells in this study. We have now changed the Figure 5c and d accordingly.
9) Does knockdown of LOXL1 prolong the survival in tumor bearing mice? A Kaplan-Meier survival plot needs to be included.
Response: We appreciate the reviewer’s comment, however, due to a technical limitation we believe that this point cannot be addressed at this time. As we described in response to comment#1, Loxl1-/- SCID mice develop lethal rectal prolapse during tumorigenicity study, thus making difficult to accurately assess tumor-specific survival.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments Cancers-496345
Date: 4/17/19
The current study entitled “LOXL1 is regulated by integrin α11 and promotes non-small cell lung cancer tumorigenicity” is a continuous study from their previous publication entitled “Integrin α11β1 regulates cancer stromal stiffness and promotes tumorigenicity and metastasis in non-small cell lung cancer ” (doi:10.1038/onc.2015.254). In this study, the authors discovered effects of loss and gain of function of LOXL1 on tumor cells’ invasion using genetic approaches to knockdown and overexpress LOXL1 in cancer-associated fibroblast cell lines immortalized with hTERT. They further tested xenograft tumor cell growth in Loxl1 knockout mouse. In mechanism, the authors tested the effect of loss of function of LOXL1 on extracellular matrix (collagen) reorganization. The authors claimed that LOXL1 promotes non-small cell lung cancer tumor invasion and proliferation through regulating extracellular matrix in cancer-associated stromal cells. The hypothesis make sense and the data are solid. My major comments are as followings:
1. Title: the authors claim their discovery in non-small cell lung cancer, but the data only include figures from lung adenocarcinoma (in figure 1) and adenocarcinoma cell lines (A549) and large cell carcinoma cell line (H661: https://www.atcc.org/Products/All/HTB-183.aspx). Either change the title or add more data including squamous cell carcinoma data in Figure 1.
2. In Figure 1, the authors claim that LOXL1 is regulated by integrin α11 (ITGA11). However, how integrin α11 regulates LOXL1 was not mentioned. Is it through regulating promoter activity of LOXL1? Does overexpression of LOXL1 in ITGA11 knockdown cells rescue the functional effects induced by ITGA11 knockdown?
3. In Figure 1b, which dot represent CAF094? What characterizations (i.e. pathology) of the NSCLC cell lines and NSCLC patients? What are the differential expressions for LOXL1 and ITGA11 in a comparison between normal lung-derived fibroblasts and NSCLC cell lines or cancer-associated fibroblasts?
4. In Figure 2, does overexpression LOXL1 in ITGA11 knockdown cells could rescue reduced collagen matrix reorganization induced by ITGA11 knockdown?
5. In Figure 2b, the authors demonstrated secreted LOXL1 protein levels in cell supernatants from CAF094 cell cultures. However, there is no functional data of co-culture of tumor cells and supernatants derived from CAFs with LOXL1 knockdown or overexpression.
6. What are expression levels for LOX, LOXL2, LOXL3, LOXL4 and ITG11A when LOXL1 is knocked down or overexpressed? Do those changes enhance or decrease the effects induced by knockdown or overexpression of LOXL1?
7. In Figure 4, why pick the large cell carcinoma cell line H661 in this study? Large cell carcinoma is less common compared to adenocarcinoma or squamous cell carcinoma in non-small cell lung cancer?
8. In Figure 4, co-culture of CAF with H661 promotes proliferation of H661 within 3 weeks in organotypic culture. How to distinguish effects of LOXL1 overexpression or knockdown in CAFs on invasion and /or proliferation of lung cancer cell line?
9. In Figure 5b, if co-injection of CAFs overexpressing Loxl1 together with A549 (or H4006) could rescue decreased xenograft tumor growth in LOXL1 knockout mouse model?
10. In Methods, how CAFs are derived?
11. In methods-4.2 Animals and Cell lines, HCC4006 was not mentioned in main text.
12. In discussion, the author did not consider distribution of LOXL1 in non-fibroblasts (such as tumor cells or endothelial cells) and their potential effects on tumor invasion/growth.
Author Response
Reviewer#2
We thank the reviewer for his useful comments. Below we discuss the comments and the changes done in the manuscript (For the figures that we have included in the response, please refer to the uploaded PDF file).
The current study entitled “LOXL1 is regulated by integrin α11 and promotes non-small cell lung cancer tumorigenicity” is a continuous study from their previous publication entitled “Integrin α11β1 regulates cancer stromal stiffness and promotes tumorigenicity and metastasis in non-small cell lung cancer ” (doi:10.1038/onc.2015.254). In this study, the authors discovered effects of loss and gain of function of LOXL1 on tumor cells’ invasion using genetic approaches to knockdown and overexpress LOXL1 in cancer-associated fibroblast cell lines immortalized with hTERT. They further tested xenograft tumor cell growth in Loxl1 knockout mouse. In mechanism, the authors tested the effect of loss of function of LOXL1 on extracellular matrix (collagen) reorganization. The authors claimed that LOXL1 promotes non-small cell lung cancer tumor invasion and proliferation through regulating extracellular matrix in cancer-associated stromal cells. The hypothesis make sense and the data are solid. My major comments are as followings:
1. Title: the authors claim their discovery in non-small cell lung cancer, but the data only include figures from lung adenocarcinoma (in figure 1) and adenocarcinoma cell lines (A549) and large cell carcinoma cell line (H661: https://www.atcc.org/Products/All/HTB-183.aspx). Either change the title or add more data including squamous cell carcinoma data in Figure 1.
Response: The term "non-small cell lung cancer" (NSCLC) is commonly used to distinguish them from small cell lung cancer and is still commonly used in clinical and pathology literatures. We agree with the reviewer that genomically, adenocarcinoma has many features that are distinct from squamous cell carcinoma, but significant overlaps also occur. Large cell carcinoma is used to designate NSCLC that clear phenotypic features of adenocarcinoma and squamous cell carcinoma are not apparent, but by gene expression profiling, it overlaps with both adenocarcinoma and squamous cell carcinoma. We have also performed a correlation analysis in squamous cell carcinoma as well (See Figure 1 below) and observe a strong correlation between ITGA11 and LOXL1 in two datasets and a moderate one in the third dataset. These data have been included in supplementary materials.
Figure 1: Spearman correlation between ITGA11 and LOXL1 gene expression in NSCLC squamous cell carcinoma has been analyzed from RNA Seq data of the TCGA patient dataset (n=501) or microarray data of UHN (n=43, GSE50081) and SKKU (n=75, GSE8894) patient dataset.
2. In Figure 1, the authors claim that LOXL1 is regulated by integrin α11 (ITGA11). However, how integrin α11 regulates LOXL1 was not mentioned. Is it through regulating promoter activity of LOXL1? Does overexpression of LOXL1 in ITGA11 knockdown cells rescue the functional effects induced by ITGA11 knockdown?
Response: We believe that integrin α11 regulates LOXL1 through its promoter activity, since mRNA levels are impacted. However, this mechanism needs further intensive investigation. This is one of our future projects.
Regarding the effect of LOXL1 overexpression in ITGA11-/- cells, please refer to the response to comment#4 below.
3. In Figure 1b, which dot represent CAF094? What characterizations (i.e. pathology) of the NSCLC cell lines and NSCLC patients? What are the differential expressions for LOXL1 and ITGA11 in a comparison between normal lung-derived fibroblasts and NSCLC cell lines or cancer-associated fibroblasts?
Response: We thank the reviewer for this comment. We have now added and highlighted CAF094 in green in the Figure 1b of the main manuscript. We have also added expression of ITGA11 and LOXL1 in normal fibroblasts (see Figure 2 below). The characterizations of the NSCLC patients and cell lines have been added to the Supplementary materials.
Figure 2: Integrin α11 (ITGA11) and LOXL1 are mainly expressed in CAFs. Expression of ITGA11 and LOXL1 has been analyzed by RT-qPCR in normal lung fibroblasts (n=9), in CAF extracted from NSCLC patient tumors (n=20) and in different established NSCLC cell lines (n=31). CAF094 are highlighted in green.
4. In Figure 2, does overexpression LOXL1 in ITGA11 knockdown cells could rescue reduced collagen matrix reorganization induced by ITGA11 knockdown?
Response: To address this question, we have overexpressed LOXL1 in C2C12 cells, which lack integrin α11, and have performed a collagen gel contraction assay (see Figure 3 below). As previously published (Tiger et al., PMID: 11518510), C2C12-α11 cells displayed more collagen contraction compare to C2C12 wild-type cells. Interestingly, overexpression of LOXL1 in C2C12 cells rescued the reduced collagen matrix reorganization induced by the absence of integrin α11. We believe that the observed collagen matrix contraction mediated by C2C12-LOXL1 cells, in the absence of integrin α11, is LOXL1-dependent and is mediated by other integrins that contribute to cell-collagen interactions. In long-term collagen gel contraction assay, integrin α5β1 and αvβ1 are candidates to mediate contraction through indirect binding to collagen (Zeltz et al., PMID: 24361615). The data have been included in Supplementary materials.
Figure 3: Effect of LOXL1 overexpression in C2C12-mediated collagen remodeling. (a) Analysis of human LOXL1 expression in C2C12s by RT-qPCR. (b) C2C12 cell lines were embedded in attached collagen gel and collagen gel diameter was measured 12 days later. Collagen gel contraction was assessed photographically. Scale bar: 1.5mm. Statistics were performed using Mann-Whitney (**, p<0.01, n.s., not significant).
5. In Figure 2b, the authors demonstrated secreted LOXL1 protein levels in cell supernatants from CAF094 cell cultures. However, there is no functional data of co-culture of tumor cells and supernatants derived from CAFs with LOXL1 knockdown or overexpression.
Response: As shown in Figure 4 of the main manuscript, we have performed a functional assay where tumor cells were co-cultured with CAF094, secreting varying levels of LOXL1 (i.e. knocked down or overexpressed). Since LOXL1 affects collagen cross-linking, these cells were co-cultured using a collagen/matrigel matrix (i.e. organotypic culture model) to determine whether the secreted LOXL1 affects tumor cell invasion into this matrix. We showed that overexpression of LOXL1 in CAF094 significantly increased tumor cell invasion, whereas knockdown of LOXL1 decreased this process, as a result of collagen remodeling. We also found that the presence of CAFs is essential for the invasion of tumor cells to occur.
We believe that this is the most physiologically relevant functional assay, as it allows to study the interaction of LOXL1, secreted directly by CAFs, with collagen matrix, and its effect on tumor cell invasion in a three-dimensional environment. Thus, we do not think that co-culturing tumor cells with CAF supernatant alone would further improve the manuscript.
6. What are expression levels for LOX, LOXL2, LOXL3, LOXL4 and ITG11A when LOXL1 is knocked down or overexpressed? Do those changes enhance or decrease the effects induced by knockdown or overexpression of LOXL1?
Response: We have analyzed the expression levels of the LOX family members and ITGA11 in CAFs with LOXL1 knockdown or overexpression (see Figure 4 below). In LOXL1-knockdown CAF, there is no significant difference in expression of the different genes that were analyzed. In LOXL1-overexpressing CAF, we surprisingly observed a decrease of ITGA11 expression that we believe is compensated by other integrins. We also observed an increase of LOXL3 expression, which is however not significant.
Figure 4: Analysis of LOX family members and ITGA11 expression in CAF094 cell lines. Gene expression has been measured by RT-qPCR. The values were normalized using the housekeeping gene RPS13.
7. In Figure 4, why pick the large cell carcinoma cell line H661 in this study? Large cell carcinoma is less common compared to adenocarcinoma or squamous cell carcinoma in non-small cell lung cancer?
Response: We have selected the H661 NSCLC cell line for two reasons: In a previous screening study using a 3D migration assay, H661, among other NSCLC cell lines, showed significant migration in collagen. Moreover, H661 is a slow-growing cell line that is suitable for the organotypic culture model to study invasion.
8. In Figure 4, co-culture of CAF with H661 promotes proliferation of H661 within 3 weeks in organotypic culture. How to distinguish effects of LOXL1 overexpression or knockdown in CAFs on invasion and /or proliferation of lung cancer cell line?
Response: We agree with the reviewer that distinguishing cell proliferation from invasion may be important. To exclude cell proliferation, we have selected a slow-growing cell line for this assay as we mentioned in the response to comment#7. We have stained H661-organotypic culture sections with MIB1 antibody, a marker for cell proliferation (see Figure 5 below). We did not see any differences among different conditions, indicating that in this model, LOXL1 only influences H661 invasion.
Figure 5: Proliferation of H661 lung tumor cells was studied in an organotypic model in presence of CAF in which LOXL1 is overexpressed (CAF094-LOXL1) or knocked down (CAF094-shLOXL1). Acellular matrix (no CAF) and matrix populated with mock-CAF (CAF094-mock) were used as control. H661 cells have been stained using MIB1 antibody. Xenograft lung tumor section has been used as positive control and normal lung section as negative control. Scale bar: 100µm.
9. In Figure 5b, if co-injection of CAFs overexpressing Loxl1 together with A549 (or H4006) could rescue decreased xenograft tumor growth in LOXL1 knockout mouse model?
Response: The reviewer raises an interesting point. As shown in Figure 5c of the main manuscript, when we co-injected A549 with CAFs in which LOXL1 is either overexpressed or knocked down, into SCID mice (Loxl1+/+), we showed different levels of intra-tumoral collagen organization that depend on LOXL1 expression in CAFs and not of basal Loxl1 levels in this mouse model. This indicates that the initial presence of the injected CAFs has a lasting effect on intra-tumoral collagen organization. Thus, if we were to co-inject CAFs overexpressing LOXL1 into Loxl1-/- mice, we would expect these CAFs to prevail over the mouse genetic background and to increase the tumor growth.
10. In Methods, how CAFs are derived?
Response: We apologize that we inadvertently did not include our published data as a reference for the fibroblasts isolation method (Navab et al., PMID: 21474781). Essentially, pairs of lung cancer resection specimens and corresponding grossly normal appearing lung tissues at least 5 cm away from the tumor were harvested within 30 min after surgical resection. Only tissues in excess of those required for clinical diagnoses were harvested for this study. Harvested tissues were placed in DMEM supplemented with 10% FBS and antibiotics (Invitrogen Corporation) for immediate transportation on ice to the laboratory. Tissues were minced into small pieces and digested for 1 h at 37 °C in DMEM containing 10% FBS and 1 mg/mL collagenase type I (Roche Molecular Biochemicals). The cell suspension was centrifuged at 180 × g for 5 min, and the pellet was resuspended in the fresh DMEM containing 10% FBS and plated onto 100- mm tissue-culture plates. The cells were kept in culture with twice media exchange for one week. The initial number of cells in the beginning varies between the tumors. However, by changing the media twice a week we were able to have a confluent 100- mm tissue-culture plate (~5x106- 8x106 cells), which was enough for passaging into five to eight 10 cm petri dishes as passage 1. At this passage, we used immunocytochemistry and we showed that both CAF and NF stain positive for vimentin but negative for cyokeratin (Navab et al., PMID: 21474781). We never observed tumor cell contamination and by changing the media for one week we were able to get pure fibroblasts populations.
11. In methods-4.2 Animals and Cell lines, HCC4006 was not mentioned in main text.
Response: In the main text we mentioned H4006 instead of HCC4006. We apologize for this confusion and we have now changed the figure and main text accordingly.
12. In discussion, the author did not consider distribution of LOXL1 in non-fibroblasts (such as tumor cells or endothelial cells) and their potential effects on tumor invasion/growth.
Response: We did not discuss the contribution of non-fibroblasts in LOXL1 production, because ultimately the role of LOXL1 would remain unchanged regardless of the cell type that secretes it. LOXL1 promotes collagen remodeling, which in turn enhances tumor invasion and tumor growth. However, integrin α11 is only expressed on CAF in NSCLC and we have shown that integrin α11 regulates, in part, LOXL1 expression. Furthermore, we showed low or no expression of LOXL1 in NSCLC cell lines. Thus, CAFs appear to be the main source of LOXL1 in the NSCLC tumor stroma.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments have been satisfactorily addressed.
Reviewer 2 Report
I am satisfied with the revised data.
