Predestined neutrophil heterogeneity in homeostasis varies in transcriptional and phenotypic response to Candida

Once perceived to be homogenous effector cells, neutrophils have since been shown to exhibit population heterogeneity. Here, we established an experimental model of clonal neutrophil heterogeneity using conditionally immortalized clonal granulocyte monocyte progenitors (GMPs) and their mature neutrophil progeny. Transcriptional and epigenetic profiling showed conserved genome-wide signatures of transcription and chromatin accessibility that were specific to individual GMP clones and their paired neutrophil progeny, suggesting that clone specificity is established as early as the GMP stage. Clone-specific genes in vital regulatory pathways were pre-programmed and exhibited delayed expression in the mature neutrophil stage. The clone-specific gene expression in the mature neutrophils paired to enhancer activation in their parental GMPs. To determine whether transcriptional heterogeneity predicted the response to fungal pathogens, neutrophil clones were functionally profiled. Clones demonstrated heterogeneous responses to fungal pathogens in vitro and revealed neutrophil subsets with evidence for tailored functional responses to Candida spp. as well as specific transcriptional and epigenetic patterns that may explain these differences. Together, this work establishes that heterogenous GMP and neutrophil compartments exist under homeostatic conditions and that these represent predefined clusters that are uniquely adapted to control invasive fungal pathogens. Short Summary Clonal neutrophil progenitors demonstrate heterogeneity in transcription and chromatin accessibility which may inform response to later fungal challenges.


61
Unlike heterogeneity within the adaptive immune cell compartment, evidence for 62 heterogeneity in innate cells has been challenging to define. Within the innate immune

141
To demonstrate that the GMP clones mature similarly to primary neutrophils, we utilized

169
To illustrate potential relationships between clones, we clustered the eight clones based  Fig. 4a), with a total of 860 inter-cluster DEGs, and at the neutrophil level (Suppl. 174 Fig. 4b), with a total of 523 inter-cluster DEGs. Notable similarities between clones using 175 transcriptional clustering are seen in clones 7 and 8 (blue), 11 and 13 (purple), and 14 and 16 176 (green). These clones appear to be more closely related at the mature neutrophil state than at 177 the GMP state (Figure 2d) neutrophil. This points towards the second hypothesis proposed in Figure 1a in which GMPs are 192 pre-programmed, thereby resulting in mature neutrophil heterogeneity.

195
To assess the chromatin state for clone-specific epigenetic differences, we performed  circles to closed purple circles), as was found with transcriptional profiling (Figure 2b), however,

216
other clones appear to be less correlated (e.g., open green circles to closed green circles).

217
Promoter chromatin accessibility showed a distinct yet modest correlation with gene 218 expression in the same clone (Figure 3, Suppl. Fig. 5a-b). However, the comparison of promoter

226
Together, these epigenetic signatures suggest that GMPs have enhancer activation that precedes 227 neutrophil maturation which may alter the transcriptional and phenotypic response of the mature 228 neutrophil.

230
Clones follow the same differentiation trajectory while maintaining specific expression 231 differences

232
We sought to understand the extent to which the signatures of expression differ between 233 individual clones at the GMP stage translated into expression differences at the neutrophil stage.

234
The heatmap in Figure 4a shows the extent of overlap between clone-specific DEGs observed 235 for individual clone pairs at the GMP stage and clone-specific DEGs observed for the same clone 236 pairs at the neutrophil stage. This overlap is shown as a fraction relative to GMP DEGs (upper part of the heatmap) and to neutrophil DEGs (lower part of the heatmap). Although this overlap 238 was relatively modest (typically 10-30%), it was highly statistically significant (Suppl . Table 1), 239 especially in the context of strong transcriptional changes that accompany differentiation.

240
Remarkably, all clones shared a strong similarity of their differentiation gene signatures ( Figure   241 4b, Suppl. Table 2). Approximately ~80% of genes differentially expressed between GMPs and 242 neutrophils were shared between all clones (Figure 4b). When we analyzed the subset of these 243 shared DEGs that have a DAR in either a nearby promoter or enhancer region, we found that a 244 large fraction of these genes showed a strong correlation between the levels of chromatin 245 accessibility and gene expression.

254
Together, this suggests that the crosstalk between clone-specific regulation of gene expression 255 and chromatin remodeling may play a role in maintaining clonal integrity across the maturation 256 transition.

257
Next, we analyzed the temporal progression of clone-specific patterns of chromatin 258 accessibility and gene expression between GMPs and neutrophils. Notably, we found that several 259 genes had a temporal delay between the onset of clone-specific chromatin accessibility of a 260 nearby enhancer and clone-specific expression of the gene. Some genes had clone-specific 261 differences in the ATAC-seq signal at a nearby enhancer at the GMP stage that did not manifest 262 in a significant clone-specific difference in the corresponding gene expression. However, this   clone-specific chromatin regulation preceded clone-specific gene expression, as well as the 288 sequence motifs at the promoters of these genes, demonstrated the enrichment of transcription factor binding sites (Figure 4h). Notably, Spi-1 was found within both enhancers and promoters. 332 and can be utilized as a "fingerprint" to functionally identify each clone.

333
Neutrophils control pathogens through phagocytosis and the production of ROS within the 334 phagosome to achieve microbial elimination. We, therefore, tested the ability of clonal neutrophils 335 to phagocytose and produce ROS against three Candida spp. Using a flow cytometry-based 336 assay with neutrophils marked by CD11b expression (based on successful neutrophil labeling 337 shown in Figure 1d) and heat-killed fungi labeled with AlexaFluor647, we found that clones 1, 2, 338 and 5 were best able to phagocytose and produce ROS consistently between all species of 339 Candida while clones 12, 13, 15, and 16 were poorly suited to these tasks (Suppl. Fig. 6a-d). However, the clone's ability to phagocytose Candida spp. was not necessarily correlated with the 341 ability to produce ROS when examined by cluster groups (Suppl. Fig. 6e). Additionally, clones 342 overall only minimally phagocytosed or produced ROS in response to heat-killed C. glabrata 343 which is consistent with previous work using a single wild type clonal GMP cell line(Negoro et al., .

345
Clones were also tested for their ability to kill a wild-type C. albicans strain (SC5314-iRFP) and live Candida spp., but when they did phagocytose live C. albicans they were better able to 354 deliver fungicidal activity.

355
To determine if clonal neutrophils differed in their ability to release NETs, we co-incubated 356 neutrophils with C. albicans yeast or hyphae and measured extracellular DNA using Sytox (Suppl.

396
Each clone also differed in the robustness of their swarming response, quantified as the 397 area of the swarm over time. Using the same cluster groups, swarming responses for each cluster 398 were quantified (Figure 7d). Here, cluster 2 was found to have a more robust swarm than either

430
To explore and define heterogeneity beyond functional profiling, we performed 431 transcriptional and epigenetic analyses. The data link functional clustering of neutrophil clones to 432 their transcriptional and epigenetic grouping. RNA-seq based clusters define basic neutrophil 433 heterogeneity under homeostatic conditions, further investigation will be required to map the 434 evolution of neutrophil phenotypes in response to Candida spp. Additionally, more work will be 435 required to define changes in functional clusters following inflammatory stimuli or infection. We

436
identified several gene pathways that exhibit delayed expression, but which also demonstrated 437 the priming of specific genes as evidenced by enhancer chromatin accessibility at the GMP stage.

438
Recently it was demonstrated that trained immunity could be passed down from mother to 439 offspring and that these progeny exhibit altered progenitor cell development at the transcriptional