Scalable Enrichment of Immunomodulatory Human Acute Myeloid Leukemia Cell Line-Derived Extracellular Vesicles

Round 1

Reviewer 1 Report

The manuscript by Binder et al describes their experience with enrichment of extracellular vesicles (EVs) derived from human acute myeloid leukemia (AML) cell lines as well as how these EVs can influence immune responses and contribute to immune escape of AML cells. The manuscript addresses an interesting and important topic in the leukemia field, especially regarding mechanisms through which the bone marrow niche may be remodelled to favor proliferation of AML cells. While there is certainly some interesting data shown here, the lack of primary AML-derived EVs is disappointing. It is also not clear why the authors used PBMCs for the T and NK cell experiments instead of simply enriching T or NK cells, respectively, from PBMCs - these would be much cleaner experiments. The manuscript would be greatly improved if the authors could provide any data showing whether or not AML-derived EVs actually can remodel the bone marrow niche. 

In figure 2, the authors perform immunoblot analysis to show enrichment of EVs. Is it really fair to compare identical sample volumes via immunoblot in this instance? Isn´t the volume greatly reduced during the EV purification process? Wouldn´t it be more fair to compare the respective volumes of the earlier steps that correspond to the volume of the final concentrated EV preparation? For example, if the process results in 100-fold concentration, then one would compare 100 µL of unprocessed sample with 1 µL of final concentrated sample.

In figure 3, the authors should explain why multiple signals (bands) are observed for some proteins (e.g. CD81, CD63). Also, the protocol for the western blot analysis of EVs seems to be missing. As this is a key aspect of the manuscript that the authors use to demonstrate EV purity, this information is important to include. Figure 3E does not seem to be called out in the main text.

In figure 4, the authors should explain why the various EV doses were chosen. Which dose of EVs is comparable to which ratio of AML to PBMCs (e.g., is 15,000:1 EVs:PBMCs comparable to 1:1 AML:PBMCs)? The authors write n=3-8, but only 1-2 data points are visible in several graphs, so the authors should better display this data. Do the bars in the graphs represent means? If so, then they do not seem to correspond to the numbers given in the main text. The authors should double check this and modify as needed (and/or provide more explanation to clarify this).

Author Response

While there is certainly some interesting data shown here, the lack of primary AML-derived EVs is disappointing.

Thank you for this important point that was also raised by reviewer #3. We have performed additional experiments comparing plasma-derived EVs from three AML patients compared to healthy control plasma EVs. The results are now shown in Fig.A4. We found an inhibitory effect of AML-EVs from patient’s plasma. Two additional authors had to be added for providing AML plasma and enabling performance of these experiments.

It is also not clear why the authors used PBMCs for the T and NK cell experiments instead of simply enriching T or NK cells, respectively, from PBMCs - these would be much cleaner experiments.

We appreciate your suggestion and agree that this would be a more specific way to conduct these experiments. However, several considerations prompted us to conduct the experiments as presented. First, the primary focus of this work was to establish a process for scalable production and enrichment of AML cell line-derived EVs accompanied by basic verification of the functionality of these EVs as a proof-of-concept. Second, unsorted PBMCs resemble the composition of leukocytes found in peripheral blood more closely compared to sorted cells. Therefore, we selected PBMCs for our T cell proliferation assays. Results on sorted cells will refer to research questions encompassing mode and mechanism of action, which are intended to be included in future projects. Third, as for the cyotoxicity assays, the only other cell population with cytotoxic capacity besides NK cells are CD8⁺ cytotoxic T lymphocytes (CTL). The K-562 target cells used are devoid of MHC class I (and class II), which excludes activation of CTLs via allo-recognition. In contrast, NK cells are activated by the absence of MHC class I („missing self“), resulting in rapid cytotoxicity. Therefore, the cytotoxicity observed in these experiments was considered NK-specific. A corresponding statement was thus included in the Methods section, lines 161-162.

The manuscript would be greatly improved if the authors could provide any data showing whether or not AML-derived EVs actually can remodel the bone marrow niche.

We fully agree with this reviewer that this aspect surely is an important topic. Modulation of the bone marrow niche by AML-EVs has been shown in several studies (e.g. Huan et al., Cancer Res. 2013, 73(2):918-29; Kumar et al., Leukemia. 2018, 32(3)575-587; Doron et al., Leukemia. 2019, 33(4):918-930). Additional experiments reproducing these published data were outside the scope of our current study.

In figure 2, the authors perform immunoblot analysis to show enrichment of EVs. Is it really fair to compare identical sample volumes via immunoblot in this instance? Isn´t the volume greatly reduced during the EV purification process? Wouldn´t it be more fair to compare the respective volumes of the earlier steps that correspond to the volume of the final concentrated EV preparation? For example, if the process results in 100-fold concentration, then one would compare 100 µL of unprocessed sample with 1 µL of final concentrated sample.

We agree with this reviewer that concentration factor normalization of the volumes used would result in reduction of the concomitant increase in marker abundance as well. The Western blot data in Figure 2C and quantification in Figure 2D were intended to validate the enrichment of vesicles (using EV-relevant markers like CD63 and CD81) beyond particle measurement. Therefore, comparison of equal volumes of the preparations was selected to demonstrate concentration of vesicles.

In figure 3, the authors should explain why multiple signals (bands) are observed for some proteins (e.g. CD81, CD63).

CD63 is subject to extensive post transcriptional modifications (glycosylation, ubiquitination) and thus commonly shows a variable size of 30 – 60 kDa in Western blots. CD81 has two isoforms, isoform 1 with 236aa and 26 kDa (NCBI RefSeq: NP_004347.1) and isoform 2 with 165aa and 18 kDa (NCBI RefSeq: NP_001284578.1), which correspond to the bands in the presented Western blots. Thus, for both markers the observed staining patterns can be considered typical. Following this reviewer’s criticism, this comment was included in the revised manuscript (lines 338-339).

Also, the protocol for the western blot analysis of EVs seems to be missing. As this is a key aspect of the manuscript that the authors use to demonstrate EV purity, this information is important to include.

Thank you! The protocol section referring to Western blot analysis was accidentally merged with the preceding section. We corrected this in the revised manuscript (see 2.3 Transmission electron microscopy (TEM), cryo-TEM and Western blot (line 137).

Figure 3E does not seem to be called out in the main text.

Thank you for drawing our attention to this. Indeed, the figure was described but not called out. This has been corrected in the revised manuscript (line 341).

In figure 4, the authors should explain why the various EV doses were chosen. Which dose of EVs is comparable to which ratio of AML to PBMCs (e.g., is 15,000:1 EVs:PBMCs comparable to 1:1 AML:PBMCs)?

Thank you for asking these excellent questions. The different EVs (from HL-60, KG-1, OCI-AML3 and MOLM-14) were chosen to represent the different AML cell lines analyzed in this study. The soluble factor fraction and the cells secreting both, soluble factors and EVs, were tested for comparison. This is stated in the revised manuscript (lines 366-368). We also refer to our previous work giving additional explanations (Ref. 32+33). We also agree with this reviewer, that selection of EV doses is always arbitrary. Many researchers define EV doses based on protein amount. True EV concentration will thus depend on particle/protein ratio. Both data sets are shown in this study in Figure 2A. We selected particle count and cell number ranges based on previous titration (Ref. 15, 32, 33). A single cell can secrete >1,500 EVs per 24h under steady state conditions and increases EV secretion under inflammatory and/or immune response conditions (Ref. 30). We can just speculate that 15,000 EVs per target cell represent a realistic ratio corresponding to an increase, by one order of magnitude, under stress conditions. This is now included in the revised manuscript (page 4, lines 153-155)

The authors write n=3-8, but only 1-2 data points are visible in several graphs, so the authors should better display this data.

Thank you for highlighting this point. We improved visibility by increasing data point size in the revised manuscript.

Do the bars in the graphs represent means? If so, then they do not seem to correspond to the numbers given in the main text.

The data in Figure 4 are presented as median with range to facilitate interpretation of the variation between the individual experiments. However, the numbers given in the old text in fact did belong to a previous incomplete version and were corrected accordingly. Thank you!

Reviewer 2 Report

Well-written and comprehensive review of the effect of AML-derived EV. The impact of AML-EVs on immune surveillance seems than can be very important. The authors show various technologies to purify EVs. Both T cell- and NK cell-based functional assays were performed.  EVs showed significant enrichment of immune response and leukemia-related pathways in tandem mass-tag proteomics and a significant dose-dependent inhibition of T cell proliferation, which was not observed with AML cells or their soluble factors. This is of particular interest since different AML-EVs dosing can reduce NK cell activities and can serve as a new model of testing new targeted therapy in vitro.

Author Response

Thank you!

Reviewer 3 Report

The authors propose a scalable workflow for the purification of extracellular vesicles in acute myeloid leukemia (AML). To do so they used commercially available leukemia cells. The final message of the paper is to emphasize the role of AML extracellular vesicles in immune escape of leukemia cells.

The manuscript is interesting and it is clear in the presentation. My comments are below:

1. The authors should incorporate data of primary specimens.
2. The discussion should incorporate a couple of paragraphs indicating the clinical impact of extracellular vesicles.

Author Response

The authors should incorporate data of primary specimens.

Thank you for this important point that was also raised by reviewer #1. We have performed additional experiments comparing plasma-derived EVs from three AML patients compared to healthy control plasma EVs. The results are now shown in Fig.A4. We found an inhibitory effect of AML-EVs from patient’s plasma. Two additional authors had to be added for providing AML plasma and enabling performance of these experiments.

The discussion should incorporate a couple of paragraphs indicating the clinical impact of extracellular vesicles.

The discussion was amended accordingly by adding paragraphs on clinical impact of EV. We also added 20 new references.  

Round 2

Reviewer 1 Report

The authors have satisfactorily addressed my concerns.