Exosomes derived from HIV-1-infected cells promote growth and progression of cancer via HIV TAR RNA

People living with HIV/AIDS on antiretroviral therapy have increased risk of non-AIDS-defining cancers (NADCs). However, the underlying mechanism for development and progression of certain NADCs remains obscure. Here we show that exosomes released from HIV-infected T cells and those purified from blood of HIV-positive patients stimulate proliferation, migration and invasion of oral/oropharyngeal and lung cancer cells. The HIV transactivation response (TAR) element RNA in HIV-infected T-cell exosomes is responsible for promoting cancer cell proliferation and inducing expression of proto-oncogenes and Toll-like receptor 3 (TLR3)-inducible genes. These effects depend on the loop/bulge region of the molecule. HIV-infected T-cell exosomes rapidly enter recipient cells through epidermal growth factor receptor (EGFR) and stimulate ERK1/2 phosphorylation via the EGFR/TLR3 axis. Thus, our findings indicate that TAR RNA-containing exosomes from HIV-infected T cells promote growth and progression of particular NADCs through activation of the ERK cascade in an EGFR/TLR3-dependent manner.

cell exosomes grew faster and exhibited significantly higher tumor weight compared to those mixed with Jurkat exosomes. There are many explanations possible for these results and technical issues require that these data are interpreted with caution, -The authors do not show that the effects seen in vivo are dependent on TAR and/or TLR3/ERK cascade.
-Mixing tumor cells with exosomes upon inoculation do not recapitulate in vivo 'priming' of the tumor cells.
-Using Immune compromised mice, neglects the role of immune (stroma) cells.
-Ultracentrifugation of exosomes, leads to many artifacts through co-isolation of unwanted particles.
-To substantiate the claim that TLR3 is involved, can the authors inhibit this in vivo? Does this reduce tumor growth.
-It is unclear whether the amount of ctrl exosomes is the same, quantitation of exosomes remains an issue.
-C an the authors show that TAR is present in tumors of HIV+ patients?
Apart from the in vivo data, i feel the in vitro data is more convincing and better developed but as said, I'm not sure what the real mechanistic advance is here compared to prior publications on this topic. The finding that serum exosomes from IV patients have similar effects as purified in vitro TAR+ exosomes is very interesting but what i miss here is quantitative data. How much Tar is in the exosomes, how much is transferred, what is the entry mechanism of the exosomes in the tumor cells, is this tumor cell specific, do HIV proteins play a role (what i could imagine)? C an the authors reveal something about the stoichiometry? Such data would set it more apart from prior studies.
Reviewer #3: Remarks to the Author: The study suggests that exosomes from HIV-infected cells contain TAR RNA fragments to drive growth and cancer progression. This story builds in part on a wealth of previous studies e.g. on HIV-1 Tat as Kaposi's sarcoma inducer. All studies in artificial experimental systems should be carefully checked as the major reason of HIV-related cancers is simply the virus-induced immune suppression! I therefore did not like the first 2 sentences of the abstract. It is not really known at all if "residual and persistent HIV replication" is needed for such cancer induction. Antiviral drugs stop this cancer route, but they also resolve the immune suppression.
The study is presented in a confusing manner, e.g. the C EM control cell is presented on page 4, but only shows up in Fig 1B. Panel 1A lacks any of these controls. Panel 1B lacks an internal control. Later AC hE is launched as control, but this should be explained before panel 1B. And why do we need to see two AC hE titration in panels 1C and 1D. Only 1C is mentioned in the text (which probably should be 1D?). Anyhow, all very confusing.
One sometimes uses J1.1 cells with or without TNF treatment. Isn't a latent HIV infection mean that the integrated provirus is transcriptionally silent? One should then not expect any TAR transcript to be present.
Anyhow, it may be important to analyze some of the other J-LAT clones to beef up these findings. Do they all do the same thing. Detailed proteomic and lipodomic analyses were done, but the results are not worked out or discussed.
The study should provide more information on items like exosome-free FBS and the quality of the differential centrifugation method.

1
Authors' Response NCOMM-17-01488B Response to Editor's comments: We would expect any revised manuscript to provide further evidence that TAR in exosomes from patients under cART affects cancer development and to provide additional data to support its pathophysiological relevance as pointed by all three reviewers.
Authors' Response: We thank the Editor for pointing out the major concerns from all three reviewers. We have provided further experimental data and analyses to address this concern. We collected plasma samples from six HIV-positive patients under cART, six normal control individuals and five HIV-positive head and neck squamous cell carcinoma (HNSCC) patients treated with cART. We tested those plasma exosomes and found that all HIV-positive exosomes significantly promoted proliferation of cancer cells compared with those from normal controls and that all HIV-positive exosome samples contained TAR RNA. Taking together, our data indicate the pathophysiolocal relevance of TAR RNA-bearing exosomes with promotion of cancer growth and progression. We have made substantial revision on the issue by adding a new section (page 9) and a new figure ( Figure 4) with new data to address this concern. Patient information, including genders, ages, cART treatment, viral load, and CD4 T cell counts was listed in Supplementary Table 1 and 2.
Response to reviewers' comments: We thank reviewers for their constructive criticisms and comments. We have added further experimental data and analysis and substantially revised the manuscript to address each of reviewers' comments. Following is our response to the comments.
Reviewer #1, Expertise: HIV related cancers/viral related cancers (Remarks to the Author): This is an intriguing manuscript that presents a potentially important mechanism by which HIVinfected cells can promote tumorigenesis through HIV-TAR found in exosomes. The results can help explain the increased incidence of certain tumors (especially non-AIDS defining tumors) in patients with HIV infection. I did, however, have some general and specific comments and concerns: Authors' Response: We greatly appreciate the recognition of the potential importance of our discoveries by reviewer 1 and find the comments both insightful and helpful. Below is our pointby-point response to the reviewer's critiques. Figure 2a), use a high concentration (4x109) exosomes. The experiments performed with exosomes from HIV-infected patients (like Figure  2e) used a similarly high concentration of exosomes. The question must be raised if these effects are only seen with an extremely high concentration of exosomes, or whether they could occur in patients. The authors should clearly address this and address the uncertainty regarding whether this effect can apply in patients. Also, a dose-response curve should ideally be shown which would provide a sense of the concentration-dependence of the effect.

The experiments with exosomes (such as
Authors Response: 1) We thought about this good question as well. Concentrations of exosomes used in our experiments were quantified using the EXOCET exosome quantification assay (System Biol Inc.), which converted acetylcolineaseasterase (AChE) activity to numbers of exosomes using the manufacture's calibration standards. We quantified total exosomal proteins at various concentrations of exosomes and found that concentrations of exosomes used in our report was in the range of 60-180 µg/ml total exosomal proteins and reported the data in Supplementary Fig. 1c. Others have reported using exosomes with total exosomal proteins from 50-250 µg/ml (Lee J, et al. PLoS One, V.8(12);2013;Higginbotham JN, et al. Curr Biol, 21(9);2011;Zhang B, et al. Stem Cell Int. 2016: 1929536). Therefore, the concentrations that we used are in the regular range for studies involving exosomes. The amount of exosomes at 4x10 9 /ml in Fig 2 was equivalent to ~80 µg/ml of total exosomal protein. Therefore, the protumor effect of HIV-infected T-cell exosomes was not due to extremely high concentrations of exosomes.
2) The concentration of exosomes in patients can vary both spatially and temporarily. In this revised report, we typically obtained 30~40 x 10 9 exosomes from 2ml of patient plasma, which was about 15-20 x10 9 exosomes/ml plasma, indicating that we were able to readily isolate significant amount of exosomes from HIV patient plasma. In the new results that we obtained since the initial submission, we were also able to detect TAR RNA in the FFPE tissues of HIVpositive oral squamous cell carcinoma patients, the sources of which can be either circulation, or local infiltrating T cells, or both. Therefore, it is very likely that tumor cells can encounter the level of exosomes that we used in this study.

Much of the increase in head and neck squamous cell carcinoma (HNSCC) in HIV patients is from HPV-associated tumors, and much can be accounted for by an increase in smoking in this population. Also, epidemiologic evidence for a direct role of HIV in lung cancer is somewhat controversial. The findings here would be bolstered by looking at another tumor in which it is clearer that HIV plays a direct role in pathogenesis, such as B cell non-Hodgkin lymphoma.
Authors Response: 1) The reviewer's point is valid. Indeed, infection with high-risk HPV and smoking are major causal factors for HNSCC in both HIV-positive and HIV-negative populations. However, a most recent epidemiological study reported by Mahale et al. indicated that standardized incidence ratios (SIRs) significantly increased for HNSCC (SIR=1.66) and lung cancer (SIR=1.71) in HIV-infected subjects. Excess absolute risks (EARs) increased with age for lung cancer and HNSCC, among other cancers (Mahale et al., Clin Infect Dis, 2018. PMID: 29325033). In addition, Shiels et al. provided epidemiological evidence showing that lung cancer, anal cancer, HNSCC and myeloma were diagnosed at modestly younger ages, possibly reflecting accelerated cancer progression (Shiels, et al., Clin Infect Dis, 64:468-, 2017. PMID: 27940936). Therefore, our data reported in this manuscript were supported by these epidemiological evidence.

More should be said about the HIV-infected patients from whom the exosomes were purified. Were they on cART? What was their viral load? Can the authors study exosomes from patients with both high and low viral loads?
Authors' Response: Supplementary Table 1 and 2 listed HIV-infected patient information, including genders, ages, viral load, treatment, and CD4 counts when patients were recruited. Our results included exosomes from patients with both high and low viral loads; exosomes purified from all HIV-positive patients were able to induce HNSCC and lung cancer cell proliferation. We have made substantial revision on the issue by adding a new section and a new figure (Fig. 4) with new data to address this concern. Furthermore, our results showed that exosomes purified from latent or TNFalpha HIV-positive J1.1 cells promoted cancer cell proliferation and migration at the same level, suggesting that viral loads were not the determining factor for exosomes to promote tumor growth and progression. Figure 3j, the authors conclude that scrambled R06 aptamer does not suppress DEFB103 mRNA induction. However, there is a trend down (p=0.08), and one can question whether this represents a non-effect, especially when using a Student's t test with 3 replicates (not a truly proper use of Student's t test). This finding should be studied further, or the results reinterpreted with regard to the above comment.

In
Authors Response: We have repeated the experiments at least three times with modification in transfection and found that transfection of R06 to exosomes from HIV-infected T cells significantly reduced ability of HIV+ exosomes to enhance expression of the DEFB103 gene. We agreed with the reviewer that the Student's t test was not proper for statistical analysis in this case. In this revision we presented data (means ± SD, n = 3; one experiment out of three repeats) with F-test, which fits for replicates fewer than 5. figure 4, lane 1 should be labeled. Also, more detailed explanation of this essentially negative figure should be provided.

Minor comment: In Supplement
Authors Response: Lane 1 of the figure has been labeled. Supplementary Fig. 4 has been moved to new Fig. 4. More detailed explanation has been provided in Result Section (page 9 and 10).

In this study, the authors claim that exosomes from both latently HIV-1-infected T cells or those from sera of HIV-positive subjects significantly promote proliferation and progression of HNSCC and lung cancer cells and that this effect is mediated by activation of the ERK cascade via HIV TAR RNA in the exosomes that upon transfer promotes tumor growth.
While in principle this is an interesting hypothesis, the in vivo data shown is not sufficiently conclusive to support the major claim. My largest concern, reducing enthusiasm for publication at this stage in Nat Comm., is the physiological relevance of the findings. The in vivo data (Fig.  2d) are the most important advance in their study but unfortunately not sufficiently developed or controlled. Moreover, prior reports, that were not cited (Boelens et al., Cell 2014;Baglio et al., PNAS 2016;Nabet et al., 2017)

, have shown that defined small RNA species in exosomes from virus and tumor cells activate PRRs in recipient cells, making the mechanistic advance of this study somewhat limited.
Authors' Response: We greatly appreciate the reviewer's valuable comments. To fully address the reviewer's largest concern, multi-pronged efforts were undertaken to demonstrate the physiological relevance of TAR RNA in human patients.

1) We have doubled HIV+ patient plasma exosome samples and detected TAR RNA in all of them.
2) In addition, since the last submission, we have obtained plasma from HIV+ HNSCC patients and detected TAR RNA in all of them as well.
3) Moreover, we have acquired FFPE sections from HIV+ HNSCC biopsies. Remarkably, we were able to detect TAR RNA from RNA extracted from these biopsy tumor sections, which not only provided support for disease relevance of our study, but also pointed to the potential of using this technology for future identification of patients who may benefit from novel therapies targeting TAR RNA. Our results show a novel mechanism by which the exosomal HIV TAR RNA not only represents exogenous pathogen-associated molecular pattern (PAMP) signals to potentially trigger the pro-inflammatory TLR3 pathway, but also directly enhance proliferation and progression of cancer cells through ERK1/2 phosphorylation in an EGFR-dependent manner. Interestingly, EGFR mediates rapid entry of exosomes into recipient cells, together with TLR3 and the TAR RNA, to stimulate activation of the ERK1/2 cascade and expression of ISGs.

To evaluate the effect of HIV-1-infected T-cell exosomes on tumorigenesis in vivo, the authors inoculated nude mice subcutaneously with HNSCC cells mixed with the same concentration of exosomes from HIV-positive (HIV+) or HIV-negative (HIV-) control Jurkat cells, respectively.
Tumors inoculated together with the J1.1 cell exosomes grew faster and exhibited significantly higher tumor weight compared to those mixed with Jurkat exosomes. There are many explanations possible for these results and technical issues require that these data are interpreted with caution.
-The authors do not show that the effects seen in vivo are dependent on TAR and/or TLR3/ERK cascade.
Authors' Response: We generated TLR3-knockout HNSCC cells and inoculated together with exosomes from HIV-infected or control T cells into nude mice. While HIV+ T cell exosomes significantly promoted TLR3 wild type HNSCC xenograft tumor growth, the exosomes failed to induce growth of xenograft tumors inoculated with TLR3-KO cancer cells, indicating that HIVinfected T-cell exosomes enhanced tumor growth in vivo was TLR3-dependent (now in Fig. 6e).

-Mixing tumor cells with exosomes upon inoculation do not recapitulate in vivo 'priming' of the tumor cells.
Authors' Response: The reviewer raised a valid point. However, in the absence of better model this approach was as close as a simulation in vivo, as to do systemic treatment will require far more much HIV+ exosomes to be feasible. In addition, our results suggested that HIV-infected T-cell exosomes were risk factors for growth and progression of existing tumors.
-Using Immune compromised mice, neglects the role of immune (stroma) cells.
Authors' Response: Although it will be better to use immune competent mice as the reviewer suggested, with human cancer cell xenograft model, we had to use immune deficient mice.

-Ultracentrifugation of exosomes, leads to many artifacts through co-isolation of unwanted particles.
Authors Response: We have perfect control of NON-HIV exosomes prepared by the same method. In addition, cells were maintained in media supplemented with exosome-depleted FBS. The differential ultracentrifugation method removed some unwanted particles, such as apoptotic bodies, at high speed.

-To substantiate the claim that TLR3 is involved, can the authors inhibit this in vivo? Does this reduce tumor growth?
Authors Response: Yes, we generated TLR3-null HNSCC cells and inoculated together with exosomes from HIV-infected or control T cells into nude mice. We found that HIV+ T cell exosomes did not promote TLR3-null tumor growth (in revised Fig. 6e).

-It is unclear whether the amount of ctrl exosomes is the same, quantitation of exosomes remains an issue.
Authors' Response: We did control the amount of exosomes to be the same between HIV+ and HIV-preps for all experiments using the method shown in Fig. 1.

-Can the authors show that TAR is present in tumors of HIV+ patients?
Authors' Response: Thanks for the suggestion by the reviewer. Yes, we have detected TAR in 1) formalin-fixed paraffin-embedded (FFPE) biopsy specimens of HIV-positive HNSCC patients, 2) plasma exosomes of HIV-positive HNSCC patients. These results have been present in new Fig.  4d, e, f, relative patient information in supplementary table 1 and 2.

Apart from the in vivo data, i feel the in vitro data is more convincing and better developed but as said, I'm not sure what the real mechanistic advance is here compared to prior publications on this topic. The finding that serum exosomes from HIV patients have similar effects as purified in vitro TAR+ exosomes is very interesting but what i miss here is quantitative data.
How much Tar is in the exosomes, how much is transferred.
Authors' Response: We quantified TAR RNA in exosomes purified from plasma of HIV-infected subjects and compared it with exosomal TAR RNA from 8E5 cells, which contained a single copy of HIV genome. Our results showed that HIV-patient plasma exosomes contained 7~80fold higher TAR RNA than that in 8E5 exosomes (in revised Fig. 4d, e).

What is the entry mechanism of the exosomes in the tumor cells?
Authors' Response: We found that entry of exosomes into tumor cells was EGFR dependent, because treatment of cancer cells with the humanized monoclonal antibody to EGFR blocked entry of fluorescently labeled exosomes into recipient cells. These data are now presented as Supplementary Movie 1 and Supplementary Fig. 5 in the revised manuscript.

Is this tumor cell specific, do HIV proteins play a role (what i could imagine)? Can the authors reveal something about the stoichiometry? Such data would set it more apart from prior studies.
Authors' Response: Yes. Specifically, it is carcinoma cell specific and requires EGFR of cancer cells. HIV proteins did not play a role because exosomes from HIV+ C22G T cells, which contain a defective HIV Tat and HIV Nef-null, could still induce cancer cell proliferation and migration. To address the concern about stoichiometry, we did proliferation assays of cancer cells in response to different concentrations of HIV-infected T-cell exosomes. The data indicated that cancer cells responded to HIV-infected T-cell exosomes in a dose-dependent fashion ( Supplementary Fig. 2c).

Reviewer #3, Expertise : HIV/TAR infection (Remarks to the Author):
The study suggests that exosomes from HIV-infected cells contain TAR RNA fragments to drive growth and cancer progression. This story builds in part on a wealth of previous studies e.g. on HIV-1 Tat as Kaposi's sarcoma inducer. All studies in artificial experimental systems should be carefully checked as the major reason of HIV-related cancers is simply the virus-induced immune suppression! I therefore did not like the first 2 sentences of the abstract. It is not really known at all if "residual and persistent HIV replication" is needed for such cancer induction. Antiviral drugs stop this cancer route, but they also resolve the immune suppression.
Authors' Response: The reviewer was correct in pointing out that HIV-related cancers is simply the virus-induced immune suppression, which is a causal factor for AIDS-defining cancers, including Kaposi's sarcoma, cervical cancer and Non-Hodgkin lymphoma.
However, the central goal of this manuscript is to elucidate the mechanisms of none-AIDSdefining cancers among HIV-infected people under cART, who do become susceptible to the development of non-AIDS-defining cancers, including HNSCC and lung cancer.
We agree with the reviewer's comments on "residual and persistent HIV replication". The statement has been changed to "human immunodeficiency virus pathogenesis plays key roles…" Even cART dramatically decreased AIDS-defining cancers and resolved the immune suppression, exosomes derived HIV-infected cells can still promote tumor growth and progression in non-AIDS-defining cancer patients.
The study is presented in a confusing manner, e.g. the CEM control cell is presented on page 4, but only shows up in Fig 1B.  Anyhow, it may be important to analyze some of the other J-LAT clones to beef up these findings. Do they all do the same thing?
Authors' Response: We used three HIV-infected Jurkat clones and an HIV-infected CEM cell line. They all express TAR in exosomes and promote cancer cell proliferation and progression.
Detailed proteomic and lipodomic analyses were done, but the results are not worked out or discussed.
Authors' Response: The full characterization of the proteins and lipids is beyond the scope of this study. The data have been removed.
The study should provide more information on items like exosome-free FBS and the quality of the differential centrifugation method.
Authors' Response: We thank the reviewer for pointing this out. We have provided more information about methodology of exosome-free FBS and quality of the differential ultracentrifugation method in revised manuscript (page 5 and 19).

Page 5: viral taxonomy?
Authors' Response: Since proteomics and lipidomics data have been removed from the revised manuscript, viral taxonomy is no longer needed.
The studies with patient samples is also not convincing at all. Very small patient numbers are used (3 versus 2) and the 2 control exosome preparations were mixed, but that was not done for the 3 patients. This is not correct as mixing will change the actual composition for differentially expressed items! Authors' Response: We agree with the reviewer. We have provided further experimental data and analyses to address this concern. We collected plasma samples from six HIV-positive patients under cART, six normal control individuals and five HIV-positive HNSCC patients treated with cART. We tested each of those plasma exosomes and found that all HIV-positive exosomes significantly promoted proliferation of cancer cells compared with those from normal controls and all HIV-positive exosome samples contained TAR RNA. We have made substantial revision on the issue by adding a new section (page 9) and a new figure (Fig. 4) with new data to address this concern. Patient information, including genders, ages, cART treatment, viral load, and CD4 T cell counts was listed in Supplementary Table 1 and 2. I understand the switch to synthetic TAR studies, although it is quantitatively difficult to compare synthetic RNA transfection with exosome delivery and one should always realize that the former method will likely test unnatural TAR amounts. Aptamers are used to demonstrate that TAR is also the critical component in exosomes, but the results did not convince me. The aptamer effect in Fig 3i is quite small and about the same as that of the mutated control aptamer.