Loss of kallikrein‐related peptidase 7 exacerbates amyloid pathology in Alzheimer's disease model mice

Abstract Deposition of amyloid‐β (Aβ) as senile plaques is one of the pathological hallmarks in the brains of Alzheimer's disease (AD) patients. In addition, glial activation has been found in AD brains, although the precise pathological role of astrocytes remains unclear. Here, we identified kallikrein‐related peptidase 7 (KLK7) as an astrocyte‐derived Aβ degrading enzyme. Expression of KLK7 mRNA was significantly decreased in the brains of AD patients. Ablation of Klk7 exacerbated the thioflavin S‐positive Aβ pathology in AD model mice. The expression of Klk7 was upregulated by Aβ treatment in the primary astrocyte, suggesting that Klk7 is homeostatically modulated by Aβ‐induced responses. Finally, we found that the Food and Drug Administration‐approved anti‐dementia drug memantine can increase the expression of Klk7 and Aβ degradation activity specifically in the astrocytes. These data suggest that KLK7 is an important enzyme in the degradation and clearance of deposited Aβ species by astrocytes involved in the pathogenesis of AD.

Is the 50% reduction of KLK7 mRNA in the Japanese AD brains also reflected on protein levels?
What means "Mock" in Fig. 3? Is this just a buffer control or an isotype control (what would be required)? Is the inhibitory activity of the antibody dose dependent?
In Fig. 3D numerous bands are observed. How can the authors be sure that they detect indeed the indicated dimers and trimers?
Why are the KLK7 driven effects in Fig. 3B so weak (hard to see). Was there a significant overexpression of KLK7? There are two novel findings reported. First, kallikrein-related peptidase 7 (KLK7) is shown to exacerbate amyloid pathology in transgenic mouse models of human amyloid pathology. Second, for the first time astrocytes are identified as source of Aß-degrading proteolytic KLK7 and most important the NMDA antagonist memantine, a drug improved for the treatment of Alzheimer's disease is shown to induce KLK7 mRNA expression. Interestingly, the NMDA antagonist MK-801 has no effect on the regulation of astrocytic KLK7 suggesting that the subunit composition of astrocytic NMDA receptors differ from those of neurons since the latter bind MK-801. Regarding the technical quality, the animal models used throughout are crucial for what has been achieved. Both mouse models used throughout were most appropriate, the homozygous App-NL-G-F/NL-G-F mice that were produced by homologous recombination with a human App-NL-G-F gene that carries three famial mutations causing autosomal dominant Alzheimer's disease (AD) (FAD) and even so important and appropriate, the establishment and use of Klk7-/-knockout mice and their crosses with homozygous App-NL-G-F mice.
Referee #2 (Remarks): There are several minor points that should be considered and incorporated in the final manuscript. First and most important, the statement in the discussion that "increased expression and/or activity of KLK7 would provide beneficial effects for cognitively normal individuals with senilie plaques" is not compatible with what has been published by Randy Bateman and coworkers (see for instance ANN NEUROL 2015;78:439-453). This group showed that in inviduals with amyloid positivity Aß42 is selectively and rapidly removed from the CSF. About 50% of CSF Aß42, corresponding to ≈30ng, is bound per hour to parenchymal Aß42 aggregates. These ≈30ng correspond to ≈5 of total Aß42 produced per hour, a 50% reduction in Aß42 by proteolysis mediated by KLK7 would not affect progression of AD pathology. For this at least 97.5% of Aß42 need to be removed (see the discussion in the aforementioned publication). Nevertheless, memantine has a great potential in combinatorial therapy, for instance in addition to immunotherapy or even as monotherapy after successfull removal of amyloid by immunotherapy. Please amend.
Second, the first two sentences of the discussions need to be rewritten. As an example: "In this study, we identified KLK7 as a crucial astrocytic ..... Since Klk7 deletion attenuates Thioflavin Spositive amyloid deposition, this indicates that astrocytes are involved in regulation of Aß pathogenesis via the KLK 7 pathway. It was reported .....(Shropshire et al. 2014) but the pathophysiological role of KLK7 in the etiology of AD and the origin of KLK7 remained obscure." Third, Figure 6D: LPS effect does not reach significance but it should be mentioned in the text that despite the current lack of significance a role in regulation of Klk7 cannot be excluded.
Fourth, Introduction: Change your misleading statement regarding "activation of remaining neurons to elimate Aß deposition" to what I would suggest: "remaining neurons by removal of Aß deposits might not be suitable".. Fifth, use more appropriate terminology for the age of mice -replace "mice at 3-month-old"by "mice 3-months-old" or "mice 3 months of age" Referee #3 (Comments on Novelty/Model System): Too many experiments use insufficient n numbers in terms of mice to draw firm conclusions, and often statistical correction for multiple comparisons are not done.
Referee #3 (Remarks): The manuscript by Kidana provides strong evidence that KLK7 is a novel Abeta degrading protease likely produced by astrocytes that can modulate amyloid deposition in vivo in the brain of mice. Evidence is provided that KLK7 expression may be decreased in the AD brain, but that in mice it increases as amyloid accumulates. These results are for the most part solid and well done though a few concerns are noted below. Some data on memantine is provided that suggests that this AD drug, which is used for cognitive enhancement in AD, could lower Abeta levels by increasing KLK7 expression. Though this data is intriguing, it is generated with 30 uM Memanie and the effect size in vivo is quite small and based on a very small group size. A number of suggestions regarding this data are made below. 1. It would be highly useful to actually generate kinetic data (Km, Vmax etc) on the enzymology of ABeta degradation by KLK7. This data would help to understand its relative potential for turnover of Abeta by KLK7 compared to wells tudied abeta degrading enzymes such as NEP, IDE and ECE1/2. 2. Figure 3 in the &PA2 cells it is not celar that all of the "oligomeric bands" are truly oligomers but represent novel ABeta immunoreactive cleavage products reported by Hass and colleagues. 3. The data on fibril degradation by KLK7 is important but buried in supplemental information. I think this should be in the main part of the manuscript. 4. Figure 5.The mouse data looks convincing in terms of effect isze but is poorly documented. An n of 3-4 is simply insufficient to draw firm conclusiosn. This data needs ot be repeated with a larger group size. 5. Figure 6. A t-test is not appropriate as multiple comparison are being made the data needs to be adjusted for multiple comparisons. 6. Both the human KLK7 expression data and the mouse data could be greatly improved by mining public data sets now available at SAGE AD-AMP portal. This data is freely accessible and contains hundreds of smaples form AD and control brain and longitudinal data from APP and tau mice. 7. Memantine effects. What is the rational for using 30 uM? A dose response should be shown on primary astrocytes. The expression data in vivo in figure 7c is based on too few of an n number, and again is not adjusted for multiple comparisons. The effect size is marginal. I don't think this data is robust enough to publish in its current form. Toxicity controls are lacking and it is not clear that the mechanism here is worked out. Again mining public data for memantine effects on Abeta levels in CSF would be a logical implication of this data that could be rapidly done through ADNI. Unfortunately the most relevant preclinical experiment would be to look at whether memantine lowers Abeta in an APP mouse crossed into a KLK7 background. 8. The data on tau again shows a modest effect size and is based on to low of n number. RESPONSE TO REVIEWER 1: Comment 1: Abeta produced from another source (human cells or even better primary neurons from APPPS1 mice) may be investigated as well, to prove that the observed effects are not specific to 7PA2 produced Abeta.

Response:
We thank the reviewer for this pertinent comment. In accordance with the Reviewer's comment, we examined the Aβ degrading activity of CCF-STTG1 cells using the conditioned medium of BE(2)-C cells, which is derived from Human Caucasian neuroblastoma and secrete the substantial amount of endogenous Aβ. We confirmed that human Aβ from BE(2)-C cells was also degraded by the conditioned medium of CCF-STTG1 cells in a similar manner to that from 7PA2 cells ( Supplementary Fig. 1C).
Comment 2: Is the 50% reduction of KLK7 mRNA in the Japanese AD brains also reflected on protein levels? Response: We validated several commercially available KLK7 antibodies as well as custom-made antibodies using lysates from Klk7 -/mouse. However, no antibody specifically detected an endogenous KLK7 protein in the brain lysates Klk7 wt, hets and KO mouse so far (see left). We have been trying to generate antibodies that detect the endogenous KLK7 in brain, and will report the characters of these antibodies elsewhere in future.

Comment 3: What means "Mock" in Fig. 3? Is this just a buffer control or an isotype control (what would be required)? Is the inhibitory activity of the antibody dose dependent?
Response: We appreciate the reviewer for critical reading. We used PBS (i.e., a buffer control) as Mock in Figure 3A and 4B. To address the concern by the reviewer, we examined the effect of anti-V5 Tag (R960-25, Thermo Fisher Scientific) and anti-LR11 (#611860, BD Biosciences) monoclonal antibodies as an isotype control of MAB2624 (mouse IgG2a) and observed the same results as a buffer control in both the conditioned medium of CCF-STTG1 and primary astrocytes culture (Supplementary Fig. 3A and 4B). In addition, we also confirmed the dose dependence of MAB2624 in the conditioned medium of CCF-STTG1 (Fig. 3B). Fig. 3D numerous bands are observed. How can the authors be sure that they detect indeed the indicated dimers and trimers? Response: We understand the reviewer's concern. We tested the effects of Congo Red and γsecretase inhibitor DAPT on the secreted Aβ species from 7PA2 cells (See right). We utilized 6E10 antibody, which reacts with human Aβ species starting at Asp1, but not with Aβx-peptides in this assay. We found that the bands between 2 to 10 kDa were completely diminished by γsecretase inhibitor DAPT. As Aη is generated independently of γ-secretase, these bands represent an Aβ-related protein including Aβ monomer at 4 kDa. Congo Red decreased the bands between 5 to 10 kDa, suggesting that these bands represent oligomeric form of Aβ. However, as we have not analyzed the biochemical properties of these bands, we referred them as "oligomer" Aβ.

Comment 4: In
Comment 5: Why are the KLK7 driven effects in Fig. 3B so weak (hard to see). Was there a significant overexpression of KLK7? Response: We agree with the reviewer. One possibility is that the protein concentration of overexpressed KLK7 used in this assay was low. In fact, relatively high protein concentration (i.e., µg/mL order) is required for complete degradation of Aβ by recombinant KLK7 ( Supplementary  Fig. 3). In addition, we shortened the incubation time for 3 hours in this experiment to make the difference clearer. Fig. 5D must be shown in a better resolution. In the overview I hardly see the plaques. Response: We thank the reviewer for the comment. We have changed immunohistochemical image of the Figure 5D. Fig. 5E. What do the authors want to conclude from that Figure? Astrocytes surrounding amyloid plaques should occur in the presence and absence of KLK7. Is there overall a higher astrogliosis in KLK7 ko mice? Response: We appreciate the reviewer for critical reading. We observed the significant increase of GFAP-positive activated astrocytes around plaques in App NL-G-F/NL-G-F ; Klk7 -/mouse, presumably accelerated maturation of amyloid plaques at this age. We have replaced Figure 5E to indicate that astrocyte was not so activated, even around Thioflavin S positive plaques in App NL-G-F/NL-G-F ; Klk7 +/+ mouse. We appreciate the precise comment by the reviewer. Importantly, KLK7 degrades not only Aβ monomer, but oligomer and fibrils (Fig. 3). Thus, we expected that the clearance of aggregated Aβ species as senile plaques is also facilitated by the activation of KLK7. But as we have not examined the kinetics of amyloid plaques in memantine-treated nor Klk7 knockout mouse brain, we cannot hypothesize the beneficial effect on the individuals with senile plaques. Also, we agree with the potential of memantine in the combination therapy with the immunotherapy. We deleted the indicated sentences, added phrases regarding possible combination effect of memantine in the discussion section. Comment 3: Figure 6D: LPS effect does not reach significance but it should be mentioned in the text that despite the current lack of significance a role in regulation of Klk7 cannot be excluded. Response: We appreciate the reviewer for critical reading. We added following sentence in the result section; "Although we are unable to exclude the possibility that lipopolysaccharide has some regulatory role in the Klk7,"

Comment 6:
Comment 4: Fourth, Introduction: Change your misleading statement regarding "activation of remaining neurons to eliminate Aβ deposition" to what I would suggest: "remaining neurons by removal of Aβ deposits might not be suitable". Response: We appreciate the reviewer for critical reading. We changed the text in the Introduction.
Comment 5: Fifth, use more appropriate terminology for the age of mice -replace "mice at 3month-old"by "mice 3-months-old" or "mice 3 months of age". Response: In accordance with the Reviewer's comment, we unified the writing from "month-old" to "months of age". Comment 2: Figure 3 in the 7PA2 cells it is not clear that all of the "oligomeric bands" are truly oligomers but represent novel ABeta immunoreactive cleavage products reported by Hass and colleagues. Response: We appreciate the comment by the reviewer. Please see our response to the comment 4 by the reviewer 1.

Comment 3:
The data on fibril degradation by KLK7 is important but buried in supplemental information. I think this should be in the main part of the manuscript. Response: We appreciate kind suggestion by the reviewer. We added the data regarding fibril degradation in Fig. 3 from the supplementary data.
Comment 4: Figure 5. The mouse data looks convincing in terms of effect size but is poorly documented. An n of 3-4 is simply insufficient to draw firm conclusion. This data needs to be repeated with a larger group size. Response: We understand the reviewer's concern. In this revision, we were able to add samples in the analysis of tau phosphorylation (see the answer for comment 8). However, due to the limitation of the animal facility and time, because we could not prepare additional App NL-G-F/NL-G-F ; klk7 -/mice with appropriate age in this study. Nevertheless, we are planning to analyze congenic mice at older age and report in the future.

Comment 5: A t-test is not appropriate as multiple comparison are being made the data needs to be adjusted for multiple comparisons.
Response: We appreciate the reviewer for critical reading. We conducted the statistical analysis of Figure 6C, Figure 7D and Figure 7E using Tukey's test.

Comment 6: Both the human KLK7 expression data and the mouse data could be greatly improved by mining public data sets now available at SAGE AD-AMP portal. This data is freely accessible and contains hundreds of samples form AD and control brain and longitudinal data from APP and tau mice.
Response: We appreciate kind suggestion by the reviewer. We analyzed the expression data of human brain samples using both Mayo RNAseq and Mount Sinai Brain Bank (MSBB) AD cohorts. In Mayo RNAseq database, we observed that the KLK7 expression was significantly reduced in the temporal cortex of AD group compared with that of control (q <0.05). In MBSS RNA-seq database, we found that the expression levels of KLK7 mRNA were significantly reduced in the BM22 (Superior temporal gyrus) and BM36 (Parahippocampal gyrus) with the progression of Braak NFT stage as well as plaque load. Unfortunately the expression levels of Klk7 mRNA in the database of MAPT_P301L, rTG4510, APPPS1 and TgCRND8 of AD model mice were quite low. Thus, we were unable to perform appropriate analyses. Nevertheless, we have obtained similar results from different cohorts, supporting our notion that KLK7 mRNA expression is significantly reduced in the brains of AD patients.

Comment 7:
Memantine effects. What is the rational for using 30 uM? A dose response should be shown on primary astrocytes. The expression data in vivo in figure 7c is based on too few of an n number, and again is not adjusted for multiple comparisons. The effect size is marginal. I don't think this data is robust enough to publish in its current form. Toxicity controls are lacking and it is not clear that the mechanism here is worked out. Again mining public data for memantine effects on Abeta levels in CSF would be a logical implication of this data that could be rapidly done through ADNI. Unfortunately the most relevant preclinical experiment would be to look at whether memantine lowers Abeta in an APP mouse crossed into a KLK7 background. Response: We understand the reviewer's concerns. First, we examined the cytotoxicity of memantine against primary astrocytes using alamar blue test and decided to use memantine up to 30 µM. Next, we examined the Aβ degrading activity of the conditioned medium of primary astrocytes treated with memantine, and confirmed that the Aβ degrading activity was augmented in a dosedependent manner ( Supplementary Fig. 10).
Regarding ADNI sample, we expect that there is no difference in the level of Aβ level in cerebrospinal fluid between memantine recipients and non-recipients of US-ADNI subjects by following reason. In general, Aβ42 level in CSF of AD patients is significantly decreased (Shaw et al., Ann Neurol 2009). Thus, it would be difficult to discriminate whether the additional decrease of CSF Aβ42 by upregulation of KLK7 with memantine or the progression of AD pathology. In addition, CSF Aβ40 level has not examined in ADNI study. Nevertheless, we are willing to examine the CSF Aβ42 level in cerebrospinal fluid from the patients with or without memantine in the future.

Comment 8: The data on tau again shows a modest effect size and is based onto low of n number.
Response: We understand the reviewer's concern. Here we added the number of samples to 4-5 and analyzed by Tukey's test (Supplementary Figure 8B). Again, we confirmed the increase of phosphorylated tau in App NL-G-F/NL-G-F ; Klk7 -/-. We also observed the relationship between KLK7 expression and the NFT stage from analyses by MSBB RNA-seq. Thus, we believe the loss of KLK7 expression accelerates the amyloid deposition, thereby increasing the tau phosphorylation at this age. However, to further clarify the role of KLK7 in tau pathology, additional studies using aged mice would be needed. We will report this issue in the future. Thank you for the submission of your revised manuscript to EMBO Molecular Medicine. We have now received the enclosed reports from the referees that were asked to re-assess it. As you will see the reviewers are now globally supportive and I am pleased to inform you that we will be able to accept your manuscript pending the following final amendments: 1) Please address the changes commented and recommended by the referees. Please provide a letter INCLUDING the reviewer's reports and your detailed responses to their comments (as Word file).
***** Reviewer's comments ***** Referee #1 (Remarks for Author): Abeta produced from another source (human cells or even better primary neurons from APPPS1 mice) may be investigated as well, to prove that the observed effects are not specific to 7PA2 produced Abeta. The authors have addressed this point appropriately by adding the requested new data.
Is the 50% reduction of KLK7 mRNA in the Japanese AD brains also reflected on protein levels?
The authors tried, but there is no antibody available, which recognizes endogenous KLK7. Thus, the authors have addressed this point appropriately.
What means "Mock" in Fig. 3? Is this just a buffer control or an isotype control (what would be required)? Is the inhibitory activity of the antibody dose dependent?
The authors have addressed this point appropriately by adding the requested new data.
In Fig. 3D numerous bands are observed. How can the authors be sure that they detect indeed the indicated dimers and trimers? I am still not very happy with this figure. Now the bands are labeled as oligomers (previously dimer and trimer). Furthermore in the figure added to comment 4, Congo Red reduced Abeta generation. This is surprising as the original paper (Podlisny et al., JBC 1995) showed a dramatic increase. I understand that the authors want to prove that KLK7 also degrades oligomers , which are believed to be toxic. However, at the end the authors would need to check if KLK7 is capable to block the LTP lowering activity of purified and characterized oligomers. I would therefore recommend to remove Fig. 3F.
Why are the KLK7 driven effects in Fig. 3B so weak (hard to see). Was there a significant overexpression of KLK7? This has been addressed as good as possible. I do not understand Fig. 5E. What do the authors want to conclude from that Figure? Astrocytes surrounding amyloid plaques should occur in the presence and absence of KLK7. Is there overall a higher astrogliosis in KLK7 ko mice? I think the authors want to say that the increased amyloid burden in the absence of KLK7 increases neuritic and astrocytic pathology. Why not making a separate figure, which includes the data from the Supplementary Fig. 8 and an appropriate description? This would allow the authors to draw a nice conclusion! What about KLK7 protein levels in Fig. 7C? This has been described as described in the response to comment 2.
Referee #3 (Remarks for Author): The manuscript is improved. One area of lingering concern is the kinetic data is still a gap in the current manuscript. Some discussion of this should be included as to the relative effect this protease has compered to others. Referee #1 (Remarks for Author): Abeta produced from another source (human cells or even better primary neurons from APPPS1 mice) may be investigated as well, to prove that the observed effects are not specific to 7PA2 produced Abeta.

required)? Is the inhibitory activity of the antibody dose dependent? The authors have addressed this point appropriately by adding the requested new data.
We appreciate the reviewer's comment.
We understand the reviewer's concern. We have deleted the result regarding oligomeric Abeta (Fig.  3F) and related discussions.
Why are the KLK7 driven effects in Fig. 3B so weak (hard to see). Was there a significant overexpression of KLK7? This has been addressed as good as possible.
We appreciate the reviewer's comment. We appreciate the reviewer's comment.
I do not understand Fig. 5E. What do the authors want to conclude from that Figure? Astrocytes surrounding amyloid plaques should occur in the presence and absence of KLK7. Is there overall a higher astrogliosis in KLK7 ko mice? I think the authors want to say that the increased amyloid burden in the absence of KLK7 increases neuritic and astrocytic pathology. Why not making a separate figure, which includes the data from the Supplementary Fig. 8 and an appropriate description? This would allow the authors to draw a nice conclusion!
We thank the reviewer for the comment. We moved the Fig. 5E to Appendix Fig. S8D. Also, we have separated the description regarding amyloid-induced pathology (i.e., phospho tau, neuritic change and astrogliosis) in the result section.
What about KLK7 protein levels in Fig. 7C? This has been described as described in the response to comment 2.
We appreciate the reviewer's comment.

Referee #3 (Remarks for Author):
The manuscript is improved. One area of lingering concern is the kinetic data is still a gap in the current manuscript. Some discussion of this should be included as to the relative effect this protease has compered to others.
We understand the reviewer's concern. We added the comment on the effects of known Abeta degrading enzymes on Abeta metabolism in mouse brain at discussion section. Any descriptions too long for the figure legend should be included in the methods section and/or with the source data.
In the pink boxes below, please ensure that the answers to the following questions are reported in the manuscript itself. Every question should be answered. If the question is not relevant to your research, please write NA (non applicable). We encourage you to include a specific subsection in the methods section for statistics, reagents, animal models and human subjects.

definitions of statistical methods and measures:
a description of the sample collection allowing the reader to understand whether the samples represent technical or biological replicates (including how many animals, litters, cultures, etc.).
Please fill out these boxes ê (Do not worry if you cannot see all your text once you press return) a specification of the experimental system investigated (eg cell line, species name).

B--Statistics and general methods
the assay(s) and method(s) used to carry out the reported observations and measurements an explicit mention of the biological and chemical entity(ies) that are being measured. an explicit mention of the biological and chemical entity(ies) that are altered/varied/perturbed in a controlled manner.

Data
the data were obtained and processed according to the field's best practice and are presented to reflect the results of the experiments in an accurate and unbiased manner. figure panels include only data points, measurements or observations that can be compared to each other in a scientifically meaningful way. graphs include clearly labeled error bars for independent experiments and sample sizes. Unless justified, error bars should not be shown for technical replicates. if n< 5, the individual data points from each experiment should be plotted and any statistical test employed should be justified the exact sample size (n) for each experimental group/condition, given as a number, not a range; Each figure caption should contain the following information, for each panel where they are relevant:

Captions
The data shown in figures should satisfy the following conditions: Source Data should be included to report the data underlying graphs. Please follow the guidelines set out in the author ship guidelines on Data Presentation.

YOU MUST COMPLETE ALL CELLS WITH A PINK BACKGROUND ê
We have described statistical method in the materials and methods section. No statistical methods were used to predetermine sample sizes.
No statistical methods were used to predetermine sample sizes, but our sample sizes are similar to those previously reported in several papers.

E--Human Subjects
We confirm the compliance. G--Dual use research of concern N.A.
We provided all data in the supplementary information.

Yes
We described all information of antibodies in the Materials and Method section.
We described the source of cell line in the Materials and Method section. We have routinely checked the mycoplasma contamination by DAPI staining (indicated in the materials and methods section).
We described the information of animals in the Materials and Method section. Animals were group-housed one to four per cage in a temperature-and humidity-controlled, pathogen-free vivarium, with ad libitum access to food and water. Lighting was on a 12 h/12 h cycle.