Human Vγ9Vδ2 T cells exhibit antifungal activity against Aspergillus fumigatus and other filamentous fungi

ABSTRACT Invasive aspergillosis (IA) and mucormycosis are life-threatening diseases, especially among immunocompromised patients. Drug-resistant Aspergillus fumigatus strains have been isolated worldwide, which can pose a serious clinical problem. As IA mainly occurs in patients with compromised immune systems, the ideal therapeutic approach should aim to bolster the immune system. In this study, we focused on Vγ9Vδ2 T cells that exhibit immune effector functions and examined the possibility of harnessing this unconventional T cell subset as a novel therapeutic modality for IA. A potent antifungal effect was observed when A. fumigatus (Af293) hyphae were challenged by Vγ9Vδ2 T cells derived from peripheral blood. In addition, Vγ9Vδ2 T cells exhibited antifungal activity against hyphae of all Aspergillus spp., Cunninghamella bertholletiae, and Rhizopus microsporus but not against their conidia. Furthermore, Vγ9Vδ2 T cells also exhibited antifungal activity against azole-resistant A. fumigatus, indicating that Vγ9Vδ2 T cells could be used for treating drug-resistant A. fumigatus. The antifungal activity of Vγ9Vδ2 T cells depended on cell-to-cell contact with A. fumigatus hyphae, and degranulation characterized by CD107a mobilization seems essential for this activity against A. fumigatus. Vγ9Vδ2 T cells could be developed as a novel modality for treating IA or mucormycosis. IMPORTANCE Invasive aspergillosis (IA) and mucormycosis are often resistant to treatment with conventional antifungal agents and have a high mortality rate. Additionally, effective antifungal treatment is hindered by drug toxicity, given that both fungal and human cells are eukaryotic, and antifungal agents are also likely to act on human cells, resulting in adverse effects. Therefore, the development of novel therapeutic agents specifically targeting fungi is challenging. This study demonstrated the antifungal activity of Vγ9Vδ2 T cells against various Aspergillus spp. and several Mucorales in vitro and discussed the mechanism underlying their antifungal activity. We indicate that adoptive immunotherapy using Vγ9Vδ2 T cells may offer a new therapeutic approach to IA.

The manuscript clearly suggests that the authors' focus was the use of γδT cells as an immunotherapy, which also aligns with the use of an expanded γδT-cell formulation for their experiments.Nonetheless, the authors need to make a clearer point in the Discussion that the properties of primary, unexpanded γδT cells in vivo might be vastly different.Along these lines, given that the authors had obtained > 1 million γδT cells at baseline (before expansion), it would have been interesting to perform at least some selected experiments with unexpanded cells to gain insights into the natural role of these cells in antifungal immunity.In particular, it would have been beneficial to compare the expression of various PRRs and pro-apoptotic effectors (lines 126-134), and possibly also exhaustion markers (which might be induced during ex-vivo expansion) in unexpanded and expanded cells.
Given the focus on an "off-the-shelf" cell therapy, the use of cryopreserved cells for the authors' experiments makes sense.Nonetheless, cryopreservation might be another confounder and might alter cell functionality, which poses further limitations on the applicability of the findings to the actual in-vivo function of unexpanded cells.This should be discussed.
Overall, I think the issue of antifungal/azole-resistant strains and activity of γδT cells against such strains is overstated.The common mutations facilitating such resistance are unrelated to cell wall biochemistry and architecture; therefore, the baseline expectation would be that host recognition of cell wall structures (e.g., GAG or beta-glucans) remains unaffected by the isolates' azole resistance.Some assays use very high E/T ratios.While it is understood that factors like plate geometry might necessitate such E/T ratios, it would have been beneficial to include co-culture studies with a known fungicidal cell type (e.g., NK cells or even crude PBMCs) as a control to put the fungicidal activity of γδT cells into perspective.Although the overall immunotherapeutic impact of such cells would likely extend beyond direct fungal killing (e.g., via cytokine/chemokine release to attract other effector cells), the authors place a strong focus on fungal killing/fungal growth inhibition.Therefore, such an additional control would be essential and would have been feasible to include (even from the same donor), give the use of PBMCs as a starting point.Moreover, some of the figures (e.g., 1A and 1B) do not even indicate the E/T ratio, which needs to be addressed.
How did the authors induce or confirm the presence of "biofilms" in Fig. 2? What was the difference in the experimental design between Fig. 1 and Fig. 2? Technically, although A. fumigatus can form biofilms, a hyphal layer at the bottom of the plate (without any additional components or extracellular matrix) might not constitute a biofilm.I cannot follow the authors' conclusions drawn from Fig. 3.While the use of different assays is reasonable due to planktonic growth of Δuge3, the impact of γδT cells on both Δuge3 and Δuge5 was minimal and non-significant.Therefore, the authors cannot reasonably conclude that GAG but not GM is involved in recognition by γδT cells.Moreover, the experiment shown does not suffice to conclusively demonstrate that γδT cells recognize GAG, given that the mutant has significantly altered growth patterns and substantially altered biochemical and structural properties.To make such an argument, the authors would have needed to stimulate the cells with purified fungal cell wall components and measured surrogates of activation (e.g., cytokine release) as a 2nd line of evidence.In addition, Fig. 1B suggests activity of γδT cells against some Mucorales (e.g., Cunninghamella), although Mucorales do not produce GAG.Therefore, it is unlikely that GAG is the sole or primary driver of the γδT-cell response.I suggest deleting this part or at least amending and substantially toning down the conclusions and adding more nuanced discussion.I do not like the term "Th1 cytokines" for readouts shown in Fig. 4, i.e., an experiment not using conventional Th cells.IFN-γ is secreted by various cell types and is not solely a Th1 cytokine.The same applies to similar references to Th1 cytokines in the Discussion, e.g., in lines 277 and 340.
The experiment shown in Fig. 4 should have been analyzed with a multiplex panel or at least several ELISAs.For instance, why did the authors only cite that TNF-α may be produced but not actually confirm this in their experiment?
The sentence in Line 207ff.(Notably,...) should be deleted or re-written, as the corresponding experiment (Fig. 4B) did not assess fungal killing.
All figures except Fig. 3 are purely descriptive and lack indicators of statistical significance.Significance testing has been described but not added to the panels.Fig. 5 would need panels (e.g., columns + error bars) that show the effect of both direct co-culture and supernatants on CD107a expression (in addition to the flow panels from a representative dataset).It is certainly not ideal to show such a panel for a technical/mechanistic control (Fig. 5B) while not showing the actual biological effect and its reproducibility.

Minor Comments:
The Introduction is somewhat unfocused.First, the long first paragraph goes back and forth between Aspergillus and Mucorales.Second, given that this is a predominantly immunologic/mechanistic paper, there is no need to spend 21 lines on the ineffectiveness of antifungal therapies and need for adjunct immunotherapies.This point can be made much more concisely, as emerging resistance and limitations of antifungal therapy are well known to the readership of Microbiology Spectrum.Third, the rationale for antifungal immunotherapy is not only based on the limitations of antifungal therapy but on the fact that outcomes of invasive mold infections in severely immunocompromised patients are predominantly host driven (i.e., remission of underlying hematological malignancies and recovery of cytopenia).There is abundant literature on this phenomenon, especially from U.S. cancer centers.Forth, the properties and known functions of γδT cells, especially in the context of infectious diseases, should be introduced more comprehensively.To that end, I would suggest putting the content of lines 266-284 in the Introduction.
Several statements in the Discussion, including references to the authors' own prior work, require references, e.g., in lines 266-273 or lines 285-295.
The Materials & Methods section is overly detailed, especially with regard to standard methodology (e.g., isolation of PBMCs).These sections might be shortened or relegated to a Supplementary Methods section.
The manuscript is overall well-written but some editing would be needed to correct minor grammar mistakes and stylistic issues.

January 3, 2024
Dear Editor, The authors thank the Reviewers for their valuable comments.We have now revised our manuscript to address these comments with point-by-point responses.
Reviewer 2's comment 1: The manuscript clearly suggests that the authors' focus was the use of γδT cells as an immunotherapy, which also aligns with the use of an expanded γδT-cell formulation for their experiments.Nonetheless, the authors need to make a clearer point in the Discussion that the properties of primary, unexpanded γδT cells in vivo might be vastly different.Along these lines, given that the authors had obtained > 1 million γδT cells at baseline (before expansion), it would have been interesting to perform at least some selected experiments with unexpanded cells to gain insights into the natural role of these cells in antifungal immunity.In particular, it would have been beneficial to compare the expression of various PRRs and pro-apoptotic effectors (lines 126-134), and possibly also exhaustion markers (which might be induced during ex-vivo expansion) in unexpanded and expanded cells.Reviewer 2's comment 4: Some assays use very high E/T ratios.While it is understood that factors like plate geometry might necessitate such E/T ratios, it would have been beneficial to include co-culture studies with a known fungicidal cell type (e.g., NK cells or even crude PBMCs) as a control to put the fungicidal activity of γδ T cells into perspective.Although the overall immunotherapeutic impact of such cells would likely extend beyond direct fungal killing (e.g., via cytokine/chemokine release to attract other effector cells), the authors place a strong focus on fungal killing/fungal growth inhibition.Therefore, such an additional control would be essential and would have been feasible to include (even from the same donor), give the use of PBMCs as a starting point.
Moreover, some of the figures (e.g., 1A and 1B) do not even indicate the E/T ratio, which needs to be addressed.
, and a T cell exhaustion maker (PD-1), on γδ T cells on day 0, day 3, and day 11 after stimulation of PBMC with PTA/IL-2.The aforementioned PRRs were selected because of their association with Aspergillus spp.One was Dectin-1, one of the most well characterized PRRs, was chosen, and the others (TLR2 and TLR4) were selected based on previous reports indicating that γδ T cells express the TLRs.The staining profiles were included in Supplementary Fig. 4, and the following sentences were added to the text.Line 143-153 in Results: In addition, we examined several cell surface makers expressed on Vγ9Vδ2 T cells derived from peripheral blood stimulated with PTA/IL-, and TRAIL were low at all the time points.In contrast, a high level of DNAM-1 expression was observed at all time points.The expression of NKG2D was initially high but decreased on day 3 after stimulation and subsequently resumed to an initial level on day 11.Conversely, CD25 was not expressed in the steady state, but was expressed to a markedly high level on day 3 after stimulation, and the expression gradually decreased thereafter.Programmed death-1 (PD-1, CD279) was slightly expressed on day 0, and the expression level increased on day 3 and returned to the stationary state on day 11.Line 322-332 in Discussion: In addition, we examined Vγ9Vδ2 T cells for the expression of several cell surface markers, including PRRs (Dectin-1, TLR2 and TLR4), which have been reported to be associated with Aspergillus infections (50), NK cell-related markers, NKG2D and DNAM-1, cell death receptor ligands (FasL and TRAIL), CD25, an activation marker of Vγ9Vδ2 T cells, and PD-1, a T cell exhaustion marker, on days 0, 3, and 11 after stimulation with PTA/IL-2.Dectin-1 is one of the most well-characterized PRRs recognizing b-D-glucans (50, 51), whereas its expression on Vγ9Vδ2 T cells had not been reported.It was demonstrated that TLR2 and TLR4 were upregulated on Vγ9Vδ2 T cells after stimulation with TLR4 ligands (52), leading to induced antibacterial responses (53).In this study, PRRs were not expressed on Vγ9Vδ2 T cells under a condition we employed.In addition, the expression levels of other cell surface markers remained noticeably unchanged before and after the expanded culture, except for NKG2D, CD25 and PD-1.Reviewer 2's comment 2: Given the focus on an "off-the-shelf" cell therapy, the use of cryopreserved cells for the authors' experiments makes sense.Nonetheless, cryopreservation might be another confounder and might alter cell functionality, which poses further limitations on the applicability of the findings to the actual in-vivo function of unexpanded cells.This should be discussed.Authors' responses: As pointed out by Reviewer 2, we compared the antifungal activity of freshly prepared and cryopreserved γδ T cells to examine the effect of cryopreservation.Although this finding was not included in the original manuscript, we have added it to Figure 1A (the rightmost bar).Additionally, the following sentences have been incorporated into Materials and Methods: In addition, the antifungal activity exhibited by Vγ9Vδ2 T cells before cryopreservation was evaluated under the same protocol.Line 753-754 in the legend of Fig. 1A: and cryopreservation did not significantly alter the antifungal activity of Vγ9Vδ2 T cells.Reviewer 2's comment 3: Overall, I think the issue of antifungal/azole-resistant strains and activity of γδ T cells against such strains is overstated.The common mutations facilitating such resistance are unrelated to cell wall biochemistry and architecture; therefore, the baseline expectation would be that host recognition of cell wall structures (e.g., GAG or beta-glucans) remains unaffected by the isolates' azole resistance.Authors' responses: We agree with Reviewer 2 regarding the statement on relationship between drug resistance and cell-wall constitution.In the original manuscript, we aimed to highlight the effect of γδ T cells against azole-resistant Aspergillus fumigatus, which is resistant to the treatment with conventional antifungal agents, and to demonstrate that γδ T cells exhibit antifungal activities distinctively from conventional antifungal agents.However, as pointed out by Reviewer 2, we overstated the relationship between drug-resistant strains and antifungal activity of γδ T cells.Therefore, we have eliminated or shortened the statement of azole-resistant A. fumigatus as follows: Lines 310-311 in Discussion: The existence of azole-resistant A. fumigatus has been reported worldwide (46), posing a clinically significant problem in future (7, 47).And we added the following statement in the revised manuscript.Lines 305-308 in Discussion: Since common mutations in A. fumigatus responsible for drug resistance are generally unrelated to the structure and components of cell walls that are highly antigenic for the host immune system, the antifungal activity of Vγ9Vδ2 T cells might have nothing to do with such azole-resistant mechanisms.