FcγR engagement reprograms neutrophils into antigen cross-presenting cells that elicit acquired anti-tumor immunity

Classical dendritic cells (cDC) are professional antigen-presenting cells (APC) that regulate immunity and tolerance. Neutrophil-derived cells with properties of DCs (nAPC) are observed in human diseases and after culture of neutrophils with cytokines. Here we show that FcγR-mediated endocytosis of antibody-antigen complexes or an anti-FcγRIIIB-antigen conjugate converts neutrophils into nAPCs that, in contrast to those generated with cytokines alone, activate T cells to levels observed with cDCs and elicit CD8+ T cell-dependent anti-tumor immunity in mice. Single cell transcript analyses and validation studies implicate the transcription factor PU.1 in neutrophil to nAPC conversion. In humans, blood nAPC frequency in lupus patients correlates with disease. Moreover, anti-FcγRIIIB-antigen conjugate treatment induces nAPCs that can activate autologous T cells when using neutrophils from individuals with myeloid neoplasms that harbor neoantigens or those vaccinated against bacterial toxins. Thus, anti-FcγRIIIB-antigen conjugate-induced conversion of neutrophils to immunogenic nAPCs may represent a possible immunotherapy for cancer and infectious diseases.

The other question is whether FcgRIIIB is uniquely important in this process compared to FcgRI and FcgRIIA. The statements on FcgRIIA in Fig 1g and Suppl Fig 1q are misleading as in these cases 3G8 is not the right antibody. Authors stated: "neutrophils from γ-/-mice expressing human FcγRIIIB acquired DC markers when treated with 3G8-conjugate and GM-CSF (Figure 1g), changes that were largely absent in γγ-/-neutrophils lacking FcγRIIIB or expressing only FcγRIIA" . Here, FcgRIIA specific antibody (IV.3) not 3G8 should be used to address if FcgRIIA also contributed to CD11c expression. Similarly, the statement of faster internalization of IIIB than IIA in 3G8-conjugate treated human neutrophils (Suppl. fig.1q) is misleading as IIA internalization should be measured in the presence of IV.3. In fact, the 3G8 treatment on human neutrophils should be also done using IV.3. As shown in Figure 1c, similar results (Fig 1g-h) should be obtained through FcgRIIA. Namely, if FcgRIIIB is not unique in driving neutrophil differentiation to nAPC, the significance of FcgRIIIB in therapy is less clear.
As the concept of neutrophils differentiation into APC to stimulate T cells has been demonstrated before, the manuscript needs to provide novel mechanistic understanding of the nAPC generation instead of simply broaden the GM-CSF treatment to IC+GMCSF. It is not clear if FcRg chain is involved as FcgRIIIB does not interact with signaling molecules. If FcRg is not necessary, how did Ova-IC signal the neutrophils ? In several places, the authors made statement of TLR is not required for Ova-IC mediated nAPC generation and their stimulation of T cells is out of context. TLR is understandably less relevant in IC stimulation in general as IC signaling involves src,erk pathway not MyD88 . To address if TLR signaling can lead to neutrophil differentiation to APC-like. The authors need to use LPS stimulation not Ova-IC. 3G8 (without Ova conjugation ) injection to FcgRIIIB transgenic animals is a better control than 3G8-Ova injection to g-/-animal as g-/-animals failed to generate nAPC due to lack of FcgRIIIB rather than lack of IC stimulation. The 3G8 control should be included in Fig 3a. Similarly, unconjugated 3G8 should be a control for Fig 1g. Reviewer #3 (Neutrophil biology, autoimmune disease.)(Remarks to the Author): This is an interesting manuscript from the Mayadas group that proposes that a subset of neutrophils have the ability to acquire APC capabilties in response to immune complex stimulation with a putative role in anti-tumor responses. Conceptually, this is an important advance but given that various observations are contrary to the current dogma of various aspects of neutrophil biology, ruling out certain confounders or alternative explanations to the observations is particularly important. There are also several areas that need better characterization of phenotype/function of this subset.
Major comments: 1. Overall, the manuscript is structured in a manner that makes it difficult to read and follow, with a lot of jargon and abbreviations that distract from the findings. Authors are encouraged to rewrite this in a more structured/clear manner with better detail on the justifications of their experiments/models and findings.
2. I have some concerns about the general in vitro protocol used. The authors take neutrophils from peripheral blood, stimulate for 2h, then culture the cells for 3 days. Very long time, especially since in some experiments (Figs 1e and 3a) the cells adopt the nAPC phenotype quicker and neutrophils tend to die within a few hours in culture. There is insufficient detail, or this reviewer found it hard to find the details in the paper, on what percentage of these neutrophils in culture die and what percentage survive and differentiate into "APC_like cells". Careful viability assays, ideally with some sort of tracking, would be ideal to get a better estimate of the percentage of initial neutrophils being plated that : a) survive over a 3 day period; b) acquire APC-like capabilities based on cell markers; c) any evidence of proliferation in these cells? Are these more immature neutrophils that retain plasticity or is this bona fide transdifferentiation from a terminally differentiated state of the neutrophil into other cell? This is not clear, despite the multiple and complex experimental approaches. If these cells indeed acquire a survival advantage, how does this happen? 2. In the work with human cells, the authors stimulate in whole blood, then purify the neutrophils.The rationale for this is not clear in the text. Are they attempting to look at indirect effects from other cells? In this regard, this needs to be clarified as other cells may interfere with experiment, even if it may be more reflective of the in vivo situation.
3. How physiological is the short 2h pulse? The experiments with SLE sera are likely to promote significant cell death at that time point and this is not clearly addressed.
4. The authors need to better address the finding that the nAPC phenotype correlates with a change in morphology. This is quite unclear in the text. Have these cells become bona fide DCs? Or is this change in phenotype just reflective of NET formation that has been shown by various groups to be induced by immune complexes? Some of the pictures are certainly reminiscent of this phenomenon, given the protrusions, potential evidence of extracellular debris and the /contraction/expansion of the nucleus at specific time points. If that's the case, could this be a late event when these "APC-like" markers become displayed during NET formation? Or is the change in nuclear morphology associated to a de-differentiation state? This is to me one of the most unclear aspects of the paper that needs to be better defined.
5. The scRNA is well done and interesting. One thing stands out that needs to be addressed. The authors use sorted cells, so the hAPCs have the appropriate surface markers. They conclude that hAPC2 cluster corresponds to the functional cells based on expression of APC genes and such. The PU.1 connection is also based on this cluster. But what is hAPC2? Do they express the same surface markers? Are they derived from neutrophil Nt.1? What drives the transition into hAPC2 and what would the functionality be? It appears to be of similar size to hAPC1. This needs to be better addressed and explained.
7. I was unclear on why the authors first gate on aCD11c HLA positive cells to then look at neutrophil markers in human blood. I would have expected to do it the opposite way? Gate on neutrophils first and then look at how many express APC markers to better understand the prevalence of these cells in circulation in immune-mediated diseases and healthy people. Since the paper functionality is mostly on cancer, what are the levels of these cells in various cancers and association with prognosis? While this may be beyond the scope of this paper, it would certainly enhance the implications. 8. Do these cells maintain neutrophil-related functions? This would be important to answer the question on whether they still represent bona fide neutrophils versus a transdifferentiation/dedifferentiation phenomenon. At the least the authors can look at degranulation and/or phagocytosis or ROS synthesis. 9. The Eruslanov group has published extensively on a subset of tumor-associated neutrophils with APC capabilities. This reviewer thinks it would be very important to compare this subset to the subset described in this paper. These neutrophils have been described as APC-like "hybrid neutrophils," which originate from CD11b(+)CD15(hi)CD10(-)CD16(low) immature progenitors, are able to cross-present antigens, as well as trigger and augment anti-tumor T cell responses. Are these the same cells? A thorough comparison and discussion is warranted here. Similarly, how does this APC-neutrophil fit into the LDG literature in autoimmunity and are there any overlaps here based on the gene expression data already published on these cells? NOTE: In the manuscript, all new experimental data is in red.

REVIEWER COMMENTS
Reviewer #1 (Fc receptor signaling, cancer therapy): The authors show that treatment of mouse neutrophils with immune complexes in the presence of GM-CSF results in induction of a DC phenotype exemplified by expression of CD11c, MHC-II, CD80, CD86 and CCR7. This effect was dependent on the γ chain subunit, which associates with activatory FcγRs, but not MyD88 or TRIF and thus independent of TLR signaling. Human FCGRIIA or FCGRIIIB were able to substitute for gamma chain deficiency. These CD11c and MHCII+ neutrophils were more frequent in SLE patients than in normal donors and furthermore there was a correlation of DC marker expression with increased disease severity scores. The authors provide evidence that immune complexed ovalbumin is internalised by neutrophils and presented to naïve TCR transgenic CD4 and CD8 T cells in vitro and in vivo and this approach could also mobilise endogenous T cells to elicit immunity against OVA-expressing B16 melanoma cells. The authors also provide a strategy using an anti-FCGRIIIB in a humanised mouse model, to deliver antigen to neutrophils and demonstrate that this can lead to successful priming of T cell responses. Interestingly, the authors provide preliminary data to suggest that neutrophils from 1 primary myelofibrosis and 2 AML patients can be stimulated with anti-FCGRIIIB conjugate to activate autologous T cells and attribute this to enhanced presentation of endogenous antigens. Using scRNAseq the authors show that the induced neutrophil cell population is heterogeneous comprising a subset that shares a number of genes with DCs. Given that GM-CSF and IgG immune complexes contribute to conversion of neutrophils to APCs, the authors tried to identify shared driver mechanisms. The authors identified PU.1 from RNAseq data as a potential transcription factor involved in the transition of neutrophils into DC-like APCs. Pharmacological PU.1 inhibitors as well as conditional deletion of PU.1 in neutrophils resulted in a reduction in the generation of DC-like neutrophils in vitro without affecting the induction of CD86.
1) The data in the first part of the manuscript describing conversion of neutrophils into professional APCs including the ability to cross-present Ag to CD8 T cells is largely known, although previous studies focused on the role of cytokines. In contrast, identification of PU.1 as a regulator of the generation of this population is novel.
We would like to address the question of novelty as this is a strong point of our manuscript. Each of the points below have been made clearer in the expanded manuscript. We would like to point out that, to date, whether nAPCs made with cytokines and soluble antigen can cross-present antigen to CD8 and whether mature rather than immature neutrophils can convert to nAPCs is still debated. A more thorough discussion of what is known and not known is now included in the Introduction section.

1) Immune complex binding to FcγRs represents a distinct stimulus for neutrophil to APC conversion in vitro (a) and in vivo (b): a)
In vitro: SLE immune complexes alone in the absence of GM-CSF drive neutrophil conversion to nAPC in a FcgR dependent manner (Figure 1b,c). b) In vivo: Anti-FcgRIIIB-antigen conjugate (in the absence of GM-CSF) leads to the generation of nAPCs in blood, secondary lymphoid organs, liver and lung. In new data, we show that i.v. injected GM-CSF does not lead to appreciable conversion and when injected with the conjugate does not enhance conversion compared to conjugate alone (Figure 4b).
2) Mature blood neutrophils undergo conversion to nAPC in vitro and in vivo (in mice) upon FcgR engagement. This is unlike previous studies, which largely reported on the acquisition of DC-like features by immature, immediate precursors of end stage neutrophils (e.g. band neutrophils).
3) Functionally, nAPCs generated with antibody-antigen complexes far exceed nAPCs generated with uncomplexed antigen in their ability to activate naïve CD4 and cross-present antigen to naïve CD8 T cells, and are functionally equivalent to classical DCs (cDCs) in vitro. To date, among APCs, only cDCs, robustly activate immunologically naïve CD4 and CD8 T cells (whose activation threshold significantly exceeds that for memory T cells) and cross-present antigen to CD8, a process critical for anti-tumor immunity. It is in this context that our findings are very novel as detailed below. Note: nAPC were generated with Ova-anti-Ova (Ova-IC-nAPC), Ova alone (Ova-nAPC) or no antigen. All were generated in the presence of GM-CSF. a) In vitro, Ova-IC-nAPCs stimulated the proliferation of greater than 80% of naïve CD4 T cells (Figure 2a). This was 5-10 fold higher than that observed for CD4 T cells incubated with Ova-nAPC, the condition used in previously published studies. Likewise, Ova-IC-nAPC stimulated the proliferation of greater than 80% of CD8 T cells (Figure 2b), a canonical read-out of crosspresentation. In contrast, Ova-nAPC failed to stimulate CD8 T cells: The low T cell proliferation observed with Ova-nAPC was similar to that observed with nAPCs that had no antigen ( Figure  2b). Importantly, the level of CD4 and CD8 T cell proliferation induced by Ova-IC-nAPCs was comparable to that observed for IL-4/GM-CSF monocyte derived cDCs and splenic cDCs ( Figure  2c,d). b) Ova-IC-nAPCs secrete several T cell immunomodulatory cytokines (e.g. IL-1b, IL-15, IL-23) at levels that were significantly higher than for Ova-nAPCs and, in fact, cDCs (Figure 2e).

4) In vivo, only nAPCs generated with antibody-antigen complexes activate naïve CD4 and crosspresent to CD8 T cells.
First, in a side-by-side comparison, we show that adoptively transferred Ova-IC-nAPCs but not Ova-nAPCs provide CD8 dependent anti-tumor immunity and induce the generation of endogenous, antigen specific CD8 T cells (Figure 2h). Second, the i.v. injection of an anti-FcγRIIIB-Ova conjugate (3G8-fOva) in the absence of GM-CSF generates nAPCs that promote CD4 (Figure 4c) and CD8 (Figure 4d) T cell proliferation, activate endogenous naïve CD8 T cells for robust target cell killing (Figure 4e) and inhibit growth of Ova-expressing melanoma ( Figure  5a) with associated generation of endogenous Ova-specific T cells (Figure 5b).

5) The pioneer transcription factor PU.1 promotes neutrophil to nAPC conversion.
As the reviewer acknowledges, another novel aspect of our studies is our finding that the pioneer transcription factor PU.1 is required for neutrophil to nAPC conversion ( Figure 6).
2) It remains unclear if these neutrophils express the full array of costimulatory ligands which are associated with DC activation. Similarly their ability to secret IL-12 and induce Th1 differentiation is not known. A side by side comparison with resting and activated DCs using a titration of antigenic peptides/proteins would have been useful as this could provide a direct comparison of the T cell stimulatory effects on a cell per cell basis.
We did not query the full array of co-stimulatory ligands on nAPCs associated with DC activation but rather focused on nAPC functionality and compared it to cDCs as detailed in point 1). We show that the levels of CD4 and CD8 T cell proliferation by Ova-IC-nAPC were comparable to splenic cDCs (Flt3DC) and IL-4/GM-CSF activated monocyte derived cDC (Figure 2a-d). These results infer that the IC-induced nAPCs have the full array of co-stimulatory molecules. DC derived cytokines represent an important signal for promoting T cell survival, differentiation and function. We performed ELISAs on nAPC supernatants to show that Ova-IC-nAPC released ng quantities of T cell immunomodulatory cytokines including IL-1b, TNF and IL-15 that are 5 to 3,500 fold higher than those for nAPCs generated with Ova/GM-CSF and, in many cases, higher than those observed for cDCs (Figure 2e).
3) Another drawback is the lack of a model to demonstrate the role of the conversion process during a natural immune response, for example during infection, or if this is a phenomenon that is primarily associated with immune pathology?
It is not possible currently to determine if nAPCs play a significant role in natural protective immune responses since there is currently no method to selectively deplete nAPCs. However, we do show that they are part of the natural history of SLE, a prototypical IC-mediated disease, and that they can enhance responses to vaccine antigens (in vaccinated human volunteers) and tumor antigens (in B16F10 melanoma model and patients with myeloid neoplasia). These data indicate their potential relevance to immune diseases and to therapeutic vaccinations. 4) Finally, the reader is left wondering about the mechanism through which FCGRIIIB, a GPI-anchored receptor, mediates these effects, as well as the role of GM-CSF in this context.
To explore how the GPI-linked FcgRIIIB may signal neutrophil to DC conversion, we first evaluated whether monomeric 3G8 alone is sufficient or whether 3G8-fOva, which is a heterogeneous species containing 3G8 and fOva at different ratios (unpublished data), is required. We found that only 3G8-fOva but not 3G8 treatment of FcgRIIIB/g -/neutrophils promotes receptor internalization (Figure 3b) and subsequent neutrophil to nAPC conversion to (Figure 3c) compared to isotype control. The endocytic machinery is well recognized to promote receptor signaling and gene transcription 1-4 and our previous work demonstrated that FcgRIIIB and FcgRIIA internalize ICs via a lipid-raft, actin and cdc42 regulated pathway. Cytochalasin D and MbCD that disrupt the actin cytoskeleton and lipid rafts, respectively, prevented 3G8-fOva induced FcgRIIIB internalization (Figure 3b) and reduced subsequent nAPC generation to levels seen with 3G8 or isotype control (Figure 3c) without effecting binding of 3G8-fOva to the surface (Suppl. Figure 3b). Similar results were obtained when using Ova-IC or SLE-ICs, (Figure 3d-e). nAPCs generated with GM-CSF alone (in presence of Ova control) were unaffected by the inhibitors (Figure 3e). Thus, endocytosis is an analogous proximal step driving 3G8-fOva and IC induced neutrophil to nAPC conversion. However, we cannot rule out the possibility that the pharmacological inhibitors of endocytosis effect other pathways necessary for neutrophil differentiation to nAPCs.
We also assessed the potential role of GM-CSF in FcgRIIIB induced nAPC conversion in vivo. By 72hr after 3G8-fOva injection, nAPCs generated in vivo accumulated in lymph nodes, spleen, and the nonlymphoid organs, liver and lung (Figure 4b), similar to the reported tropisms of blood cDCs. In contrast, 3G8 alone did not induce significant accumulation of nAPCs in any organs. Notably, GM-CSF treatment of mice given 3G8-fOva did not further increase the frequency of nAPCs. However, it did result in a small but significant increase in the frequency of nAPCs in some tissues of mice given 3G8 alone (Figure 4b) albeit this was significantly lower compared to mice given 3G8-fOva alone.
5) The authors suggest that their anti-FCGRIIIB-FITC-OVA conjugate comprises 1 IgG:2 FITC-OVA molecules, but characterisation of the protein is not shown. If this claim is correct then this conjugate is likely to be different from an immune complex and the implication, therefore, is that the bivalent engagement of FCGRIIIB by the conjugate is sufficient for the conversion of neutrophils into APCs.
We postulate that the conjugate has its effects by bivalent engagement of FcgRIIIB as anti-FcgRIIIB (3G8) alone does not promote conversion (Point 4) and western blot analysis of 3G8-fOva reveals a heterogenous mixture containing 3G8 and Ova at different ratios (unpublished data). Notably, five independent preparations of the antibody-antigen conjugate gave reproducible results. As the ICs (Ova-anti-Ova and SLE-IgG-RNP, which also remain uncharacterized) and the conjugate induced neutrophil to nAPC conversion that was associated with FcgRIIIB internalization and prevented by inhibitors of the actin cytoskeleton and lipid rafts, we can infer that FcgRIIIB-antigen conjugate and IC modalities trigger analogous, proximal mechanisms for neutrophil to nAPC conversion..
The degree or nature of cross-linking required to signal neutrophil to nAPC conversion is an interesting area for future investigation and would require careful fractionation of the 3G8-fOva and IC preparations followed by evaluation of the activity of each fraction.
6) Another pertinent question is whether induction of DC-like features by anti-FCGRIIIB can be dissociated from antigen delivery.
We presume that the reviewer is asking whether anti-FcgRIIIB alone can induce nAPC. Our study shows that anti-FcgRIIIB (3G8) alone is unable to induce neutrophil to nAPC conversion (see response to point 5). This suggests that bivalent receptor cross-linking, as occurs with soluble ICs and likely the anti-FcgRIIIB-Ova conjugate is required to signal neutrophil transition to nAPCs. 2) It is possible IC treatment resulted in cytokine production that further enhanced GM-CSF in neutrophil differentiation.
To address this possibility, we exploited neutrophils expressing CD45.1 or CD45.2, which can be distinguished by antibodies that recognize each isoform. We co-cultured Ova or anti-Ova pretreated CD45.2 + neutrophils with Ova-IC-pretreated CD45.1 + neutrophils in media supplemented with GM-CSF followed by the analysis of nAPC markers in each group. A significantly greater percent of CD45.1 + Ova-IC-versus CD45.2 + Ova-or CD45.2 + anti-Ova treated neutrophils acquired nAPC markers (Supp. Fig 1j). Thus, ICs enhance the frequency of neutrophil differentiation to nAPCs compared to GM-CSF alone in a cell autonomous manner.
3) That is IC amplified the conversion and resulting in more nAPCs, but GM-CSF induced phenotypic conversion of neutrophils. The current results are consistent with GM-CSF being the primary differentiation cytokine, but augmented with IC treatment.
See response to 1) and 2) 4) For example, the authors showed Ova-IC + 3 days GM-CSF resulted in higher expression of CD11c than anti-Ova or Ova treatment. This could very well be a kinetic effect as Matsushima et al showed 6 days GM-CSF culture resulted in higher CD11c expression than 4 days.
As noted in response to point 1), a major difference is that Matsushima et al primarily used immature neutrophils for the majority of their study while we used mature blood and bone marrow derived neutrophils. More importantly, functionally, we show that Ova-IC generated nAPCs were far superior to nAPCs generated with uncomplexed antigen, Ova (conditions of previous studies) in promoting CD4 T cell proliferation and was the only condition under which CD8 T cell proliferation was observed.
The result of 3G8-fOva alone had no effect on the growth of B16F10-Ova melanoma, but 3G8-fOva+GM-CSF reduced tumor growth is also consistent with GM-CSF is the main neutrophil differentiation factor ( Figure 3g). If IC is more important than GM-CSF in suppressing tumor growth, why not to use 3G8-Ova to boost instead of GM-CSF. It is also not clear if all control animals received GM-CSF injection.
In the B16F10 model, all control animals received GM-CSF, yet only FcγRIIIB expressing mice given 3G8-fOva (AAC) had reduced tumor growth. As noted in Reviewer 1 (point 1), in new studies, we show that GM-CSF does not increase 3G8-fOva induced nAPC generation in vivo. Also, Ova-specific cytotoxic T cells (CD62L low CD44 hi CD8 + /Tetramer + ) were generated in B16F10-Ova bearing FcgR humanized mice versus wild-type mice (control) preimmunized with 3G8-fOva in the absence of GM-CSF (2A3B/g -/-, 0.75 +/-0.1%; WT, 0.11+/-0.02%) albeit it was lower compared to 3G8-fOva immunized mice given GM-CSF (2A3B/g-/-: 2.12+/-0.50%). This indicates that GM-CSF does not directly affect the ability to generate nAPCs per se but we cannot rule out that it regulates the survival or function of nAPCs and/or T cells. It is possible that consecutive injections of 3G8-fOva, rather than the single injection administered may overcome the need for GM-CSF, a possibility that will be addressed in future experiments.
Similarly, Ova-nAPC showed partial tumor protection compared to Ova-IC nAPC when transferred to B16F10 tumor expressing mice (Figure 2f).
We apologize if the figure is misleading, but there is actually no statistically significant difference between Ova-nAPC treated and untreated mice (Figure 2h). The asterisks represent the statistical significance between Ova-IC versus Ova and untreated animals. Consistent with this finding, the number of endogenous T cells recognizing Ova (as detected with an MHC-I tetramer recognizing Ova-SIINFEKL peptide specific CD8+ T cells) in Ova-nAPC treated animals is 10-20 fold lower than for Ova-IC-nAPC treated mice ( Figure  2h).

5) Similar morphological changes from polymorphonuclear to mononuclear neutrophils (shown in Fig 1) was shown by Matsushima et al (Blood 2013) in GM-CSF treated cells.
Matsushima et al 2013 and Oehler et al 1998 did report morphological changes but both used immature immediate precursors of end-stage neutrophils treated with GM-CSF over 6-9 days and relied on static images. We used mature peripheral blood neutrophils treated with SLE-ICs (without GM-CSF) and conducted live cell imaging of CD11c-YFP neutrophils, which revealed the rapid kinetics of neutrophil to nAPC conversion (within 11hrs). The dramatic conversion of a nucleus that is polymorphonuclear to one that is mononuclear in appearance suggests that the neutrophil transdifferentiates into nAPCs and demonstrates that SLE-IC alone in the absence of GM-CSF can induce neutrophil to nAPC conversion. In addition, the live cell imaging answered potential criticisms of data relying solely on static images as the latter cannot rule out that progenitors contaminating neutrophil preparations expand over a 3 day culture period or that neutrophils preferentially die leaving behind "contaminating" mononuclear cells.
Likewise, the ability of differentiated neutrophils to present antigen to CD4 and CD8 T cells and stimulate their proliferation were also reported by Matsushima et al.
Please note that Matsushima et al, stated in the manuscript that GM-CSF treated immature neutrophils only supported "modest CD8 T cell proliferation". In our study, a head-to-head comparison of nAPC generated from mature neutrophils treated with Ova+GM-CSF or Ova-ICs+GM-CSF clearly showed that naïve CD4 T cell proliferation in the latter was markedly higher than the former (Figure 2a) and only Ova-ICs+GM-CSF induced nAPC cross-presented antigen to CD8 T cells (Figure 2b). This demonstrates that nAPCs generated with antigen-antibody complexes that engage FcgRs and not uncomplexed antigen generate fully functional nAPCs. The Ova-IC nAPCs induced CD4 and CD8 T cell proliferation that was comparable to classical, monocytes derived DCs and splenic cDCs (Figure 2c,d).
6) Regarding to the role of FcgRIIIB in neutrophil function, it has been shown that FcgRIIIB plays important role in neutrophil response to IC, particularly in IC-mediated neutrophil secretion of reactive oxidants (Fossati et al. Arthritis & Rheumatism 2002). The other question is whether FcgRIIIB is uniquely important in this process compared to FcgRI and FcgRIIA.
Neutrophil functions have been assigned to this receptor albeit most reports have not been able to delineate whether the FcgRIIIB is sufficient for the queried function because of the presence of FcgRIIA. Our humanized mice, expressing FcgRIIIB and/or FcgRIIA on neutrophils in the absence of any other FcgRs allows us to assign a role specifically to FcgRIIIB in the process being studied. We do not know which of the effector functions ascribed to FcgRIIIB potentially play a role in neutrophil to nAPC conversion. In new studies, we show that FcgRIIIB endocytosis plays an important role in neutrophil to nAPC conversion (see Reviewer 1, point 4-6).
Engagement of FcgRIIIB alone is sufficient although it is not uniquely important. Like FcgRIIIB/g-/neutrophils, FcgRIIA/g-/-neutrophils also convert to nAPCs after engaging ICs. With respect to FcgRI it is not constitutively expressed on human neutrophils and in our humanized FcgR mice, IIA and IIIB transgenes are expressed in mice lacking the g-chain and therefore lacking all murine FcgRs including FcgRI. We don't have a humanized mouse model for human FcgRI and therefore cannot evaluate the role of this receptor in neutrophil to nAPC conversion. In the experiment referred to by the reviewer was not conducted to conclude that FcgRIIA is not important. Rather, neutrophils from the FcgRIIA expressing mice were used as a negative control to show specificity of the 3G8-conjugate. We anticipate engaging FcgRIIA with an anti-IV.3-Ova conjugate would promote neutrophil to nAPC conversion given that, in vitro, IC treatment of FcgRIIA/g -/neutrophils promotes neutrophil to nAPC conversion in the absence of FcgRIIIB.
One of the goals of our studies was to potentially generate a therapeutic, as appreciated by the reviewer. For this, FcgRIIIB is a superior target than FcgRIIA as the former is abundantly and selectively expressed on neutrophils while the latter is also expressed on human platelets and other leukocyte subsets including NK cells. That said, if this platform were to be advanced for therapeutic purposes, a FcgRIIIB selective antibody will need to be generated, as 3G8 recognizes not only recognizes FcgRIIIB but also FcgRIIIA, which is expressed on multiple cell types.
8) Similarly, the statement of faster internalization of IIIB than IIA in 3G8-conjugate treated human neutrophils (Suppl. fig.1q) is misleading as IIA internalization should be measured in the presence of IV.3. In fact, the 3G8 treatment on human neutrophils should be also done using IV.3. As shown in Figure 1c, similar results (Fig 1g-h) should be obtained through FcgRIIA. Namely, if FcgRIIIB is not unique in driving neutrophil differentiation to nAPC, the significance of FcgRIIIB in therapy is less clear.
Please see the response above. 9) As the concept of neutrophils differentiation into APC to stimulate T cells has been demonstrated before, the manuscript needs to provide novel mechanistic understanding of the nAPC generation instead of simply broaden the GM-CSF treatment to IC+GMCSF.
The novelty of our studies is detailed in Reviewer 1, Point 1.
10) It is not clear if FcRg chain is involved as FcgRIIIB does not interact with signaling molecules. If FcRg is not necessary, how did Ova-IC signal the neutrophils ?
In our humanized mouse model, FcRg-chain is not involved in signaling as our mice are on a g -/background. Please see response to Reviewer 1 (point 4-6) for a description of new experiments that suggest a potential mechanism by which FcgRIIIB mediates neutrophil to nAPC conversion. 11) In several places, the authors made statement of TLR is not required for Ova-IC mediated nAPC generation and their stimulation of T cells is out of context. TLR is understandably less relevant in IC stimulation in general as IC signaling involves src,erk pathway not MyD88 . To address if TLR signaling can lead to neutrophil differentiation to APC-like. The authors need to use LPS stimulation not Ova-IC.
The rationale for evaluating MyD88/TLR deficient mice was not to interrogate the contribution of TLR signaling in neutrophil to nAPC conversion. Rather, it was to rule out the contribution of TLR engaged by potential LPS/endotoxin contamination of our reagents in FcgR dependent neutrophil to nAPC conversion and nAPC immunogenicity as TLR agonists are known to be required to induce cDC immunogenicity.
3G8 (without Ova conjugation ) injection to FcgRIIIB transgenic animals is a better control than 3G8-Ova injection to g-/-animal as g-/-animals failed to generate nAPC due to lack of FcgRIIIB rather than lack of IC stimulation. The 3G8 control should be included in Fig 3a. Similarly, unconjugated 3G8 should be a control for Fig 1g. We apologize that unconjugated 38G was not included in original Figure

Reviewer #3
1. Overall, the manuscript is structured in a manner that makes it difficult to read and follow, with a lot of jargon and abbreviations that distract from the findings. Authors are encouraged to rewrite this in a more structured/clear manner with better detail on the justifications of their experiments/models and findings.
The paper has been re-written and separated into sections with headings to improve readability.
2. I have some concerns about the general in vitro protocol used.
a) The authors take neutrophils from peripheral blood, stimulate for 2h, then culture the cells for 3 days. Very long time, especially since in some experiments (Figs 1e and 3a) the cells adopt the nAPC phenotype quicker and neutrophils tend to die within a few hours in culture. There is insufficient detail, or this reviewer found it hard to find the details in the paper, on what percentage of these neutrophils in culture die and what percentage survive and differentiate into "APC_like cells".
We show that only 10-15% of cells die (this information is noted as % survival above the majority of the graphs). Normal neutrophils can survive in vivo for several days, depending on the study, (Hidalgo et al Trends Immunol 2019) while in vitro, their survival is significantly increased in the presence of serum and appropriate survival factors. The method of isolation that minimizes activation is also key to retaining viability of isolated neutrophils in vitro.
b) Careful viability assays, ideally with some sort of tracking, would be ideal to get a better estimate of the percentage of initial neutrophils being plated that: a) survive over a 3 day period; b) acquire APC-like capabilities based on cell markers; c) any evidence of proliferation in these cells?
Survival and percent of cells acquiring APC-like markers are presented in each graph. Also, in all our flow cytometric analyses, only viable cells are evaluated for APC-like markers. We evaluated Ki67 expression, a proliferation marker, and saw no proliferation in the cultures through day 3, which is when we harvest our cells for all our phenotypic and functional analyses. This is included in the text of the result section.
c) Are these more immature neutrophils that retain plasticity or is this bona fide transdifferentiation from a terminally differentiated state of the neutrophil into other cell? This is not clear, despite the multiple and complex experimental approaches. If these cells indeed acquire a survival advantage, how does this happen?
Our live cell imaging of peripheral blood neutrophils provides the clear evidence that segmented, mature neutrophils directly convert to mononuclear, CD11c+ nAPCs, which suggests transdifferentitation. We show that this is a remarkably rapid process that occurs within 11hrs of stimulation with SLE-IC (with not GM-CSF present).
2. In the work with human cells, the authors stimulate in whole blood, then purify the neutrophils. The rationale for this is not clear in the text. Are they attempting to look at indirect effects from other cells? In this regard, this needs to be clarified as other cells may interfere with experiment, even if it may be more reflective of the in vivo situation.
Our rationale was to mimic a physiological relevant treatment strategy as we envision that the conjugate would be injected intravenously. This regimen also minimizes activation of human neutrophils, which are more susceptible to activation than murine neutrophils. That is, the addition of anti-FcgRIIIB-antigen to purified human neutrophils, followed by wash steps, resulted in neutrophil clumping and loss of viability, something that was not observed with murine neutrophils. Treating human neutrophils in whole blood before their isolation avoided this problem.
3. How physiological is the short 2h pulse?
We believe it is physiological as antigen is normally presented for short duration and in limited concentration. The 2hr pulse was also designed to restrict antigen uptake and processing to neutrophils before their conversion to nAPC as we were interested in answering the question of whether neutrophils can process antigen and load peptide on MHC following endocytosis of soluble protein, while still being a neutrophil as this would suggest a pathway for antigen processing in neutrophils that is distinct from its degradative routes. This could not have been appreciated in previous studies; where reported, antigen was present during the entire time of neutrophil conversion to APCs.
The experiments with SLE sera are likely to promote significant cell death at that time point and this is not clearly addressed.
We do not observe significant cell death following treatment with SLE sera or SLE-ICs compared to untreated neutrophils. The percent survival is noted above each graph that evaluates nAPC conversion. The relative lack of cell death is likely because cells are only pulsed for 2hrs with SLE sera or SLE-ICs before being placed in culture.
4. The authors need to better address the finding that the nAPC phenotype correlates with a change in morphology. This is quite unclear in the text. Have these cells become bona fide DCs? Or is this change in phenotype just reflective of NET formation that has been shown by various groups to be induced by immune complexes? Some of the pictures are certainly reminiscent of this phenomenon, given the protrusions, potential evidence of extracellular debris and the /contraction/expansion of the nucleus at specific time points. If that's the case, could this be a late event when these "APC-like" markers become displayed during NET formation? Or is the change in nuclear morphology associated to a de-differentiation state? This is to me one of the most unclear aspects of the paper that needs to be better defined.
Our live cell imaging demonstrates that mature neutrophils stimulated with SLE-ICs start to acquire CD11c and, at the same time, a mononuclear morphology. The morphological change is also seen by confocal microscopy and, in human neutrophils, by immunohistochemistry. NETosis would result in SYTO DNA dye extruded from the cells. By live cell imaging, extracellular DNA was observed in less than 5% of cells that were not associated with neutrophils that had acquired CD11c, a bonafide DC marker. As with our other in vitro studies, the SLE-ICs are washed away after a 2hr incubation in the microfluidic device, which avoids persistent stimulation and wide-spread activation of neutrophils.
5. The scRNA is well done and interesting. One thing stands out that needs to be addressed. The authors use sorted cells, so the hAPCs have the appropriate surface markers. They conclude that hAPC2 cluster corresponds to the functional cells based on expression of APC genes and such. The PU.1 connection is also based on this cluster. But what is hAPC2? Do they express the same surface markers?
Throughout the manuscript, nAPCs are regularly identified via surface expression of CD11c and MHC II surface markers using flow cytometry. In the single-cell experiment, a sorted population of CD11c+MHCII+ cells make up the cells in both nAPC1 and nAPC2 clusters and thus correspond to the cells explored throughout Figure 6.
In the original manuscript, we explored differences between nAPC1 and nAPC2 subsets by conducting a KEGG pathway analysis on the differentially expressed genes (DEGs) between the two clusters (Supp. Fig.6f). These DEGs were determined using a wilcoxon rank sum test and the top 200 genes were selected by AUC. In the revised manuscript, to more robustly define the differences between nAPC1 to nAPC2, we grouped all nAPC1 and nAPC2 single cells together into pseudo-bulk profiles and conducted differential gene expression analysis between the two (Suppl . Table S4). KEGG pathway analysis of the top 200 DEGs by p-value for each subset of nAPCs revealed upregulation of many cytokine/chemokine-related pathways among nAPC1 cells such as TNF signaling, IL-17 signaling, and NFkB signaling. nAPC2 cells on the other hand, show upregulation of pathways potentially directly related to nAPC conversion and function such as FcγR mediated phagocytosis, protein processing and antigen processing and presentation (Supp. Fig.6g). These pathways are in agreement with the original results from the KEGG pathway analysis presented in the manuscript. Yet, we posit that the pseudo-bulk analysis is more robust and resistant to incorrectly identifying differential expression among genes that are globally highly expressed. We also applied gene set enrichment analysis (GSEA) to the full list of 4374 variable genes between nAPC1 and nAPC2 cells, which further supports the KEGG pathway analysis results (Supp. Fig.6g). Altogether these results suggest key differences between the functionality of nAPC1 and nAPC2 subsets and predicts that nAPC2s represent more fully functional antigen presenting cells. These analyses have been added to the supplement and the following sentences were added to the manuscript: "Direct comparison of nAPC1 and nAPC2 subsets revealed upregulation of genes within the antigen processing and presentation pathway, as well as pathways related to protein processing, and endocytosis among nAPC2 cells (Suppl. fig.6g, Suppl Table S4)." Are they derived from neutrophil Nt.1? What drives the transition into hAPC2 We hypothesize that Nt.1 cells are progenitors of nAPC1 based on transcriptional similarity between these two clusters. As we note in the manuscript, the nAPC1 subsets show greater transcriptional similarity to Nt.1 and Nt.5 cells than to nAPC2 subsets (Supp. Fig.4c). Meanwhile, nAPC2 subsets show greater similarity to Nt.4 cells than to other Nt subsets or to nAPC1 subsets (Supp. Fig.4c). Transcriptional similarity alone does not prove that one subset is derived from the other. However, given the nature of our in vitro experiment, which captures cell states along a time course (day 0-day 3), these transcriptional similarities suggest that Nt.1 cells are the likely progenitors of nAPC1s, while Nt.4 cells are likely progenitors of nAPC2s as reflected in our trajectory (Supp Fig. 4c).
To definitively answer the reviewer's question regarding the origin of nAPC1 cells, further experiments are required. For example, once a sorting strategy is established for separating Nt.1 and Nt.4 subsets, their ability to generate nAPC1 versus nAPC2 could be evaluated and the transcription factors driving these processes for could be identified.
......and what would the functionality be? It appears to be of similar size to hAPC1. This needs to be better addressed and explained.
To directly explore the functional differences between nAPC1 and nAPC2, further experiments are required. First, a sorting strategy must be established by identifying a unique set of surface markers to separate the two subsets. This is challenging as it may not just be the absence or presence, but rather the surface expression level of a given receptor. Then, functional assays are needed to compare the ability of these subsets to promote T cell activation. We plan to pursue these interesting studies in the future but, in our view, they are beyond the scope of this paper, which establishes a new pathway of nAPC generation and demonstrates their biological relevance. 7. I was unclear on why the authors first gate on aCD11c HLA positive cells to then look at neutrophil markers in human blood. I would have expected to do it the opposite way? Gate on neutrophils first and then look at how many express APC markers to better understand the prevalence of these cells in circulation in immune-mediated diseases and healthy people.
We wanted to address the abundance of nAPCs within the total DC population so we gated first on DCs. In fact, we found that the frequency of CD11c + MHCII + DCs was similar in normal and SLE patient samples, while the CD11c + MHCI + expressing neutrophil markers differed. We also reasoned than since nAPCs are much less frequent than neutrophils in human peripheral blood, gating first on a restricted cell population would yield more accurate results.
Since the paper functionality is mostly on cancer, what are the levels of these cells in various cancers and association with prognosis? While this may be beyond the scope of this paper, it would certainly enhance the implications. Singhal et al., 2016 (Cancer Cell, 30:120-135) reported neutrophils with antigen-presenting cell features in tumor tissue taken from early-stage human lung cancer patients. This study was referenced in our discussion. However, there was no correlation with prognosis. To our knowledge, there are no published studies that correlate nAPC generation with prognosis in cancer or any other diseases. Our study in SLE patient blood is novel in that we do show that the frequency of mature nAPCs in the blood of patients with SLE, an immune complex mediated disease, correlates with clinical disease scores (Figure 1h).
8. Do these cells maintain neutrophil-related functions? This would be important to answer the question on whether they still represent bona fide neutrophils versus a transdifferentiation/dedifferentiation phenomenon. At the least the authors can look at degranulation and/or phagocytosis or ROS synthesis.
In new experiment Figure 1f-g, we observed similar levels of phagocytosis of IgG opsonized beads and E.coli by Ova-IC-nAPCs, SLE-IC-nAPCs and freshly isolated neutrophils. Reactive oxygen species generation following E.coli or zymosan treatment was also similar.
9. The Eruslanov group has published extensively on a subset of tumor-associated neutrophils with APC capabilities. This reviewer thinks it would be very important to compare this subset to the subset described in this paper. These neutrophils have been described as APC-like "hybrid neutrophils," which originate from CD11b(+)CD15(hi)CD10(-)CD16(low) immature progenitors, are able to cross-present antigens, as well as trigger and augment anti-tumor T cell responses. Are these the same cells? A thorough comparison and discussion is warranted here. Similarly, how does this APC-neutrophil fit into the LDG literature in autoimmunity and are there any overlaps here based on the gene expression data already published on these cells?
Our human neutrophils are CD15 and CD10 high, express CD66b and are CD16 (FcgRIIIB) high and are isolated from peripheral blood. Therefore, these are fully differentiated cells rather than immature neutrophil precursors described by Singhal et al. Our scRNAseq analysis was performed in murine neutrophils thus precluding a direct comparison with human LDGs. In future studies, we plan to conduct scRNAseq on human nAPCs and compare their profiles with LDGs and other neutrophil subsets.