RR:C19 Evidence Scale rating by reviewer:
Potentially informative. The main claims made are not strongly justified by the methods and data, but may yield some insight. The results and conclusions of the study may resemble those from the hypothetical ideal study, but there is substantial room for doubt. Decision-makers should consider this evidence only with a thorough understanding of its weaknesses, alongside other evidence and theory. Decision-makers should not consider this actionable, unless the weaknesses are clearly understood and there is other theory and evidence to further support it.
***************************************
Review: This paper is by a great team of HIV research innovators who employ cutting edge methods. 9 NHPs are infected with 3 MT145 variant SCIVs, then analyzed for changes in MAb and Env sequences using NGS. The long-term goal appears to be based on the previous observation by some members of this group, that there are some cross-reactive epitopes between the chimp Env and HIV Envs, in particular at the V2 apex. Although I did not see it stated in the paper, the immunofocusing strategy would presumably be to prime with the CIV Env and then boost with V2-sensitive HIV Env(s), in the hope that cross-reactive V2 Abs would be amplified upon boosts, and so become immunodominant and broadly cross-reactive.
Following SCIV infection, the NHPs developed NAbs to the infecting viruses, but very little NAbs to HIV strains except for an N160K variant of CRF250. This appears to contrast SHIV infections, for which some of the same co-authors have intriguingly shown broadly neutralizing V2 specific and other canonical specificities that resemble the kinds of bNAbs that develop in natural HIV infections. The reason for the lack of a strong V2 response may in part be because of the immunodominant V5 loop. This target was deleted in a MT145 variant. Although V2 responses were still not immunodominant, sequence data appeared to show greater evidence for selection in the V2 of this mutant virus and enrichment of canonical V2 prototype specificities characterized by long anionic HCDR3s with DDY motifs/putative sulfation, D gene usage, etc. This difference did not register in the serum neutralization data with any changes in type specific responses to the “immunogen” strains or any of the HIV panel.
The above observations suggests that other epitopes are still immunodominant on chimp trimers, raising the question of possible CIV cross-reactive bNAbs that could be checked out of interest, given that HIV bNAbs develop quite frequently in SHIV infections, is this the same with SCIV infections? and what epitope(s) they target, or if there are glycan holes that may drain Ab responses. Also, do all 3 SCIVs including the deltaV5 retain a tier 2 phenotype? This could be checked: tier 1 profile trimers would be expected to show dramatically increased sensitivity to heterologous CIV sera. If so, any CIV variants with a “looser”/open trimer phenotype would not be expected to be ideal V2 immunogens.
The Ab sequence data describes V2 bNAbs’ special features such as long anionic HCDR3s and sulfation. Although this could indicate V2-specific Abs, they potentially could bind to some other recessed basic epitope elsewhere on Env or even on another viral protein. Only near the end of the paper is antigenic specificity mentioned and then N160-dependence. Of the MAbs with these features, some but not all were N160-dependent raising questions about whether these MAb features are a consistent/reliable indicators of V2 Abs - an assumption on which a large part of this paper rests. Fig 6A seems to show only some of the MAbs with the anionic HCDR3s are N160-dependent. What do the others bind?
Overall, the paper is interesting, but could be written more clearly. There is a quite understandable struggle by the authors to curate the avalanche of “big data” from this work, which is very respectable. The paper laboriously goes through caches of data showing various patterns. However, the importance of these observations is ultimately unclear when the link between HCDR3 features and V2 specificity needs to be made more concrete. Deep sequencing work can be antigen specific and can recover paired heavy and light sequences (e.g., PMID: 31787378). One strategy might be to adjust the paper’s focus a bit – perhaps just to understand Ab responses to CIV rather than focus on V2 responses since they are a minor part of the response. The data are the result of a fact-finding mission to understand the viruses and the Abs they induce. The vaccine angle is limited, so the focus could be shifted slightly. The world that SHIVs open up for analysis is certainly of huge significance. A comparison to infection with some other SHIV might have made a better contrast for all the MAB and virus NGS data to understand the interactions. It is hard to know if there is anything special here in finding canonical mAbs as the viruses are closely related (do other CIVs not result in such HCDR3s?).
Some comments/suggestions are as follows:
MAbs seem to be isolated in Fig. 3A cartoon but only sequences are examined. Are H+L sequences paired in NGS recovery? Did any of the N160-dependent MAb sequences from antigen selection (Fig. 5, 6) match those of the NGS? DDY motifs? As mentioned above, there seem to be Abs with desired motifs that don’t target the desired site, implying that it is possible the V2 canonical features are not stringent enough for us to assume V2 binding solely based on sequence, especially given the various other potential Ab targets during infection. In short, do the epitope mapped Abs match the putative ones from NGS and does this support the profiling done in all the other figures based on Ab sequence only?
I don’t understand Fig. 5C, D (is D even covered in the narrative?). What is the rep-seq1 and 16 (repertoire sequences at week zero and 16?). Are the sulfated MAbs with or without anionic or canonical D regions the same ones that are N160 dependent? Those would seem to be the only ones to focus/mention here?
Mark Lewis address is bioqual not Los Alamos. Check affiliations are correct.
Fig. 1A - Why do they collect samples only until week 16, but RNA measurements to week 88. Does fig 1A need adjusting? Blood samples were taken at monthly intervals for over a year.
Fig S1, S2 - The animals/weeks are not clearly demarked, so we can see when one NHP stops and new one begins. The labels in S2 and elsewhere are far too small. I can barely make them out at 300% magnification. Microscopic labels, assigning samples to particular infections/animals is not so clear in many figures, which is not helped by the fact that some figures pass rapidly by the reader in only one sentence - and in some cases are not mentioned at all. The story is hard going and not altogether clear.
Line 220 - While SCIV_MT145 group showed significantly higher frequency at 166 than SCIV_MT145K.dV5 group. Is this in the logo plots? I don’t see in Fig 1G or S6
Line 221 text - “However, despite selection in V1V2 and gp41 regions, no changes were observed at the canonical V2-apex bnAb site for any animal infected with SCIV_MT145K and SCIV_MT145K.dV5 from EMPEM analysis (Fig. 1I & 2).” This is the first mention of gp41. What epitope is this? This sentence on EMPEM rapidly brushes past fig 2. If I understand this right, don’t see how the figure would or would not show “changes” at the canonical apex site since only 1 timepoint of Fabs is shown (week 12). Also, the trimers were presumably made of the initial infecting mutants, not those that might mutate in the apex over the course of the experiment. There are no apex binders in the MT145K group – does this deserve a comment also? This figure and the text could be written a bit more clearly.
Line 241 - It seems MT145K fraction of lineages was not maintained by week 16 (Fig. 3C), although the higher numbers were. I am not sure what to make of this difference. Does this mean more total lineages were produced in later samples? The fraction of lineages is a more conservative way to present this data?
Fig. 3F - Needs to be better explained in the text. It contains a lot of information. The legend is somewhat helpful, but given the complexity, it needs better explanation.
Line 287 - While the fraction of large, expanded lineages remained stable in SCIV_MT145K and SCIV_MT145K.dV5 infected animals, this fraction decreased slightly in SCIV_MT145 infected animals (Fig. S9A). I don’t understand fig S9, it talks about ighd3-9 not mentioned in the narrative. I don’t know what “large” refers to? not HCDR3, but this is the diversity of the oligoclonal Ab profile to the antigen that sets in after the initial diversity subsides? A lot of information is crammed into the narrative and requires referral to different figures, making it hard going. I suggest introducing concepts with a few more words for clarity.
Line 289 - All groups showed increases in the average HCDR3 length over time, with the SCIV_MT145K.dV5 group showing the most pronounced increase (Fig. S9B). The MT145K doesn’t really show increased cdrh3 length, except for the early transient increase.
The color coding for the different NHP groups could be helpful in Fig S1 and elsewhere for clarity.
Fig. S13 is cited before S12. Need to fix that.
Line 301 - “HCDR3 lineages (Fig. S11 and S12), although SCIV_MT145 infected animals used the IGHD3-9 gene less frequently (Fig. S11D)” there is no Fig S11D, 12D? Without anchoring specificity, it is not clear what this means. Can we assume these are v2 mabs? If so, an explanation of why needs to be added.
Line 320 - “Collectively, these results suggest that the germline-targeting SCIV that was modified to limit V5 directed off target responses generated an Ab response that most closely resembled the activation of V2-directed bnAb precursors.” This sentence saves this section, but the data and narrative preceding it could be improved. We don’t know the increased Abs with “desired” characteristics bind V2. So much data is analyzed, but it is hard to know what to make of it when the narrative shifts so rapidly without connecting sentences to understand where things are going.
Line 328 - “Given that IgM and IgG repertoires were significantly different between the 3 groups of SCIV infected RMs, we next asked whether these differences were also present in the antigen-specific repertoires. This sentence also “saves it”, but antigen specificity should come sooner in the narrative as by line 327 we still don’t know for sure if anything is V2 specific (what is the point of looking at changes that are not antigen specific?). A possible solution might be to make some of the NGS Abs if H+L pairs that were recovered (if pairs are available) and check their epitopes (V2 or not?).
Line 305 - This text refers to Fig S13 not S14. Check all subsequent figure labels.
I don’t understand S14 Fig comparison of NHP and human D segments. The last few supplemental figures are a bit hard to follow.
Fig. 4b - What does + mean? Are these individual NHPs and sorts. Which trimer was used to sort?
Line 117 - SV should be SIV.
Line 180 - Fig. 1F is neutralization data. Is the wrong figure being referenced in the text?
Fig. 1F - narrative is missing from that paragraph.
Line 213 - Fig. S5 not mentioned in text before S6.
Line 296 - Missing figure reference?
Line 317 - Could this be Fig. S17?
Line 319 - Could this be Fig. S18?
Line 276 - Overall
Line 278 - Kruskal-Wallis
Consistent naming of MT145K.dV5. Some has punctuation, some don’t. Same goes for graphs/figures. Suggest abbreviating the names to remove SCIV to save space/improve clarity/readability.
Line 383 - Supplementary Table 01 missing.
Line 392-400 - Narrative for Fig 5D missing.
Fig. 4B is not clear. What is the x-axis? What does each “cluster” of bar graph mean? Does each “cluster” of bar graph represent one macaque?
Line 1055 - What is Table X?
Fig. 1D and other logoplots. Cyan for Asn in potential N-linked glycans. However, the amino acid labeled with cyan is “O” (this may be known by bioinformatics people but could be explained here). Shouldn’t it be N? How about marking potential N-linked glycans with asterisk?
Line 1174 - MT145K
Line 1194-1196 - Possible typo? “Bottom row shows SCIV_MT145.dV% Envs bound to Fabs at the apex adjacent (green), C3/V5 (blue) and gp41 (yellow) sites?
Fig. 6A - What does the gray color in the first 3 columns represent? Is it “No binding”? Indicate that in the legend. In addition, are N160 and N160K trimers or gp120? For VH color, suggest avoiding using red (V4) and orange (V3) because these colors are used in the MT145 binding. Same goes with sY color.
Creative Commons Attribution 4.0 International License (CC-BY 4.0)
