Variations in human monocyte-derived macrophage antimicrobial activities and their associations with tuberculosis clinical manifestations

Macrophages play a significant role in preventing infection through antimicrobial activities, particularly acidification and proteolysis. Mycobacterium tuberculosis infection can lead to diverse outcomes, from latent asymptomatic infection to active disease involving multiple organs. Monocyte-derived macrophage is one of the main cell types accumulating in lungs following Mtb infection. The variation of intracellular activities of monocyte-derived macrophages in humans and the influence of these activities on the tuberculosis (TB) spectrum are not well understood. By exploiting ligand-specific bead-based assays, we investigated macrophage antimicrobial activities real-time in healthy volunteers (n=53) with 35 cases of latent TB (LTB), and those with active TB (ATB) and either pulmonary TB (PTB, n=70) or TB meningitis (TBM, n=77). We found wide person-to-person variations in acidification and proteolytic activities in response to both non-immunogenic IgG and pathogenic ligands comprising trehalose 6,6'-dimycolate (TDM) from Mycobacterium tuberculosis or beta-glucan from Saccharamyces cerevisiase. The variation in the macrophage activities remained similar regardless of stimuli; however, IgG induced stronger acidification activity than immunogenic ligands TDM (P=10e-5, 3x10e-5 and 0.01 at 30, 60 and 90 min) and beta-glucan (P=10e-4, 3x10e-4 and 0.04 at 30, 60 and 90 min). Variation in proteolysis activity was slightly higher in LTB than in ATB (CV=40% in LTB vs. 29% in ATB, P= 0.03). There was no difference in measured antimicrobial activities in response to TDM and bacterial killing in macrophages from LTB and ATB, or from PTB and TBM. Our results indicate that antimicrobial activities of monocyte-derived macrophages vary among individuals and show immunological dependence, but suggest these activities cannot be solely responsible for the control of bacterial replication or dissemination in TB.


Introduction 32
Macrophages play an important role in innate immunity to protect the human body from microbial 33 infection. After recognizing pathogens via surface receptors, macrophages phagocytose bacteria into 34 phagosomes. These phagosomes are key to macrophage antimicrobial activities. The phagosomes 35 undergo a maturation process in which they are acidified by recruiting V-ATPases (Levin,Grinstein,36 and Canton 2016) then sequentially fuse with multiple intra-vesicles including, lastly, lysosomes to 37 form phago-lysosomes. The environment inside these vesicles is highly acidic, around pH 4.5, and 38 enriched with an assortment of peptidases and hydrolases, which can kill the bacteria and process 39 antigens for presentation to T cells, leading on to secondary adaptive immune responses (Flannagan,40 Cosío Host genetics is associated with susceptibility to different clinical manifestations of TB infection but 62 little is known about the underlying mechanism(Berrington and Hawn 2007). For example, a 63 polymorphism in the receptor gene Toll-Like Receptor 2 is preferentially associated with TBM in 64 comparison to pulmonary TB (Thuong et al. 2007). Because Mtb recognition by receptors is upstream 65 of the phagocytosis process, the polymorphisms in these genes may influence the phagocytosis and 66 antimicrobial activities, resulting in different outcomes of Mtb infection. A previous study shows a 67 polymorphism in macrophage receptor with collagenous structure (MARCO), a receptor for trehalose 68 6,6′-dimycolate (TDM) from Mtb, is associated with impaired phagocytosis, which could lead to 69 defective antimicrobial activities and thus increases susceptibility to active disease (Thuong et  The beads were labeled with reporters for acidification or proteolytic activity, which allowed these 98 activities to be measured accurately in real-time and specifically in the bead-containing phagosomes. 99 Beads can also be coated with non-immunogenic or immunogenic ligands from the pathogen, allowing 100 us to study variation of macrophage antimicrobial activities in response to different ligands and to 101 dissect the specific pathways associated with such variation. This model of bead-treated macrophages 102 also enables us to handle multiple conditions with many experimental replications at the same time 103 using a microplate reader. 104 In this study, we assessed the variation in antimicrobial activities, primarily acidification and 105 proteolysis, of MDM from 53 Vietnamese healthy volunteers in response to immunogenic or non-106 immunogenic ligands, using our ligand-coated bead models. We then examined these activities in Bacteria were harvested, re-suspended in 10% glycerol, aliquoted and stored at -20 o C for later use. 148

Isolation of PBMC and human hMDM 149
Peripheral blood monocyte cells (PBMCs) were separated from 20 mL heparinized whole blood by 150 Lymphoprep (Axis-Shield) gradient centrifugation in accordance with the manufacturer's protocol and 151 previous studies (Thuong et al. 2016). PBMCs were plated in cell-culture treated 60 mm × 15 mm petri 152 dishes (Corning) with 6 -8 ×10 6 cells per dish in serum-free media and non-adhered cells were washed 153 off. On the following day, for each sample, some cells were seeded in 96-well plates (8 ×10 4 cells per 154 well) and incubated at 5% CO2, 37 o C in complete media (RPMI 1640 (Sigma) supplemented with 10% 155 heat-inactivated fetal bovine serum (FBS, Sigma), 2mM L-glutamine (Sigma), 100 units of penicillin 156 (Sigma)) containing 10 ng/mL human macrophage colony-stimulating factor (m-CSF, R&D system). 157 Cells left over were cryopreserved in FBS + 10% DMSO for further infection experiments. To derive 158 monocytes, adhered cells were incubated for 5 -7 days. The complete media was changed at day 4, 159 and the assays were performed at day 7. 160

Preparation of acidification and proteolytic beads 161
Beads coated with ligands (IgG, TDM or β-glucan) to measure acidification and proteolysis activities 162 were prepared as previously described ( hours with agitation. Coated beads were washed; for proteolysis, beads were then re-suspended in 10 169 μL of the 5 mg/mL stock of the calibration Alexa Fluor 594 SE, and agitated for 1 hour. For 170 acidification activity, beads were labelled with the pH-sensitive reporter carboxyfluorescein-SE (10 171 µL of the 5 mg/mL stock, Molecular Probes) and agitated for 2 hours. Beads were then washed and re-172 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 15, 2020. . https://doi.org/10.1101/2020.09.14.20194043 doi: medRxiv preprint suspended in 1 mL PBS with 0.02% sodium azide and stored at 4°C. The concentration of beads 173 (number of beads per ml) was counted by FACS (BD). 174 Acidification and proteolysis assays 175 The assays were performed as described previously (Tram et al. 2019). Briefly, macrophages were 176 incubated in 100 µL pre-warmed assay buffer (1 mM CaCl2, 2.7 mM KCl, 0.5 mM MgCl2, 5 mM 177 dextrose, 10 mM HEPES, 10% FBS in PBS). Beads were added into each well at a concentration 178 sufficient to achieve an average of 4 -5 beads internalized per macrophage as previously 179 reported (Podinovskaia et al. 2013). This ratio of beads allows measurement of the change in fluorescent 180 intensity in macrophages that corresponds to a drop of pH from 7.5 to 5.5, which signifies phagosomal 181 maturation of macrophages at resting stage (Podinovskaia et  of macrophages at 30, 60, and 90 min was calculated by the ratio of the RFU at 10 min over that at 30,192 60, or 90 min, respectively. Likewise, the proteolysis activity index at 60, 120, 180 and 210 min was 193 calculated by dividing the RFU at these time-points by that at 10 min. 194

Macrophage infection 195
Mtb reporter strain was cultured from -20 o C stock in 7H9T containing 30 µg/mL kanamycin at 37 o C 196 with shaking for about 15 -20 days until OD600 reached 1 -2. The bacteria was then sub-cultured. 197 When OD600 reached 0.5 -1 the culture was used for macrophage infection. 198 Frozen monocytes from LTB, PTB and TBM were rapidly thawed at 37 o C on the same day. Cells were 199 then seeded in black flat bottomed 96-well plates at 8×10 4 cells per well and differentiated to hMDM 200 as described above. At day 7 after thawing, macrophages were infected with the fluorescent Mtb 201 reporter at a multiplicity of infection (MOI) of 1, and incubated at 37 o C, 5 % CO2. Four hours after 202 infection, macrophages were washed with a pre-warmed medium without antibiotics. The intracellular 203 growth of the Mtb reporter strain was assessed over 5 days using a fluorescent microplate reader. This 204 measured the intensity of mCherry protein fluorescence at 620 nm after excitation at 575 nm. The 205 viability of macrophages was observed under light microscopy. 206 Statistical analysis 207 The donor-to-donor variation in phagosomal activities of macrophages from groups of healthy subjects 208 or active TB patients was expressed by the percentage of coefficient of variation (% CV) calculated by 209 the standard deviation (SD) divided by the mean.  when the phagosomes were completely acidified ( Table 2). The activity varied among individuals, with 240 ranges shown in Table 2. No significant difference was seen in the extent of variation of macrophage 241 activity from these 53 subjects across IgG, TDM and β-glucan beads (Table 2) or from subgroups of 242 LTB and uninfected individuals (Table S1 and S2). We then compared the macrophage acidification 243 kinetics for the three ligands ( Figure 2C). After 30, 60 and 90 min, IgG coated beads induced more 244 acidification compared to other ligand coated beads (IgG vs. TDM beads: P=10 -5 , 3×10 -5 and 0.01 at 245 30, 60 and 90 min respectively; IgG vs. β-glucan: P=10 -4 , 3×10 -4 and 0.04 at 30, 60 and 90 min), but 246 there was no statistical difference in activity between the beads coated with the two immunogenic 247 ligands, TDM and β-glucan. The higher acidification activity of macrophages in response to IgG was 248 also observed in both subgroups of LTB and uninfected persons categorized by T.SPOT ( Figure S1). 249 Proteolytic activity, measured over 210 min, increased with time in a similar way for macrophages 250 treated with IgG, TDM or β-glucan ligand-coated beads ( Figure 2B). We used the normalized activity 251 index at time points of 60, 120 and 180 min to examine the inter-donor and inter-ligand variation in 252 activity. The macrophages from the 53 healthy individuals showed wide ranges of proteolytic activity 253 in response to different ligand beads (Table 2). For example, after 180 min, the ranges of activity were 254 2.7-29.3, 3.3-24.6, 3.8-29.3 for IgG, TDM or β-glucan beads respectively. This variation between 255 individuals was further indicated by large CVs of 45% when highest activity was reached, which was 256 much greater than that of acidification (CVs of 5%). As for acidification, the variation in proteolytic 257 activity from donor to donor was similar for all ligands (Table 2, S1 and S2). Comparison of the median 258 values of the activity index for IgG, TDM and β-glucan beads showed no significant difference between 259 non-immunogenic and immunogenic ligand coated beads at any time point ( Figure 2D). 260 In summary, our results showed variations in macrophage antimicrobial activities among healthy 261 individuals, particularly for proteolysis. The extent of these variations was highly consistent across 262 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 15, 2020. . https://doi.org/10.1101/2020.09.14.20194043 doi: medRxiv preprint stimulation with different ligands. Meanwhile human macrophage acidification activity was higher in 263 response to beads coated with non-immunogenic IgG than with pathogenic TDM or β-glucan. 264

Variation of macrophage antimicrobial activities in LTB versus ATB 265
We next compared the macrophage antimicrobial activities in participants with LTB and ATB ( Figure  266 1). Beads coated with TDM from Mtb were used to measure the activities of hMDM from 101 ATB 267 and 35 LTB participants. Macrophages from both groups showed similar patterns of phagosomal 268 acidification and proteolysis ( Figure S2). The variation in acidification was not significantly different 269 for these two groups (Table 3). Meanwhile for proteolytic activity the variation was greater in LTB 270 than in ATB and the difference became significantly with time (at 120 min, P=0.08; at 180 min CV = 271 40% in LTB, 29% in ATB, P=0.03). The macrophages from TB uninfected individuals showed more 272 variation than those from LTB or ATB (Table S3 and S4). 273 We also compared median values of acidification and proteolytic activities in LTB versus ATB and 274 found no difference at the time points measured ( Figure 3A). These activities, particularly proteolysis, 275 were significantly higher in either LTB or ATB in comparison to 14 healthy subjects who were negative 276 with T.SPOT ( Figure S3). Ability to control bacterial growth was assessed for macrophages from 18 277 LTB and 55 ATB patients both by measuring mCherry intensity, indicative of Mtb survival ( Figure  278 S4) and by observing the viability of infected macrophages. For both groups, bacterial survival 279 gradually increased with time ( Figure 3B) while macrophage viability gradually decreased ( Figure 3C). 280 Comparisons between LTB and ATB groups showed no difference in either mCherry intensity or cell 281 lysis during 5 days of infection. 282

Relationship between macrophage antimicrobial activities and disseminated tuberculosis 283
Acidification and proteolytic activities and ability to control bacteria of macrophages from patients 284 with TBM (n= 64) were compared with those from PTB patients (n=60) (Figure 1). The variation in 285 macrophage acidification and proteolytic activities in response to the TDM-coated beads looked similar 286 for TBM patients as for PTB patients (Table 4, Figure 4A). At all time points in acidification and 287 proteolysis, there was no significant difference in the activity levels between the two groups of patients. 288 Mtb survival and viability of infected macrophages were also assessed for macrophages from 52 PTB 289 and 53 TBM patients. For both PTB and TBM we observed an increased mCherry signal in infected 290 macrophages during 5 days of infection ( Figure 4B). All infected macrophages remained adherent and 291 intact until day 3 when some of the cells became lysed ( Figure 4C). The hMDM from two groups of 292 patients showed no difference in either mCherry intensity or cell lysis during the 5 days. people who were healthy, had latent TB or active TB, either PTB or TBM, to extend understanding of 298 the impact of macrophage activities on TB disease protection and progression. 299 We explored person-to-person variations of acidification and proteolysis activities of macrophages 300 stimulated with immunogenic or non-immunogenic ligands. These variations were highly preserved 301 among different stimuli or among cells originating from different hosts, ranging from healthy to having 302 different forms of TB. We found acidification activity of healthy donors' macrophages was greater in 303 response to non-immunogenic IgG than to pathogenic ligands TDM or β-glucan. We did not find any 304 differences in observed activities between LTB and ATB, or PTB and TBM. 305 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review) preprint
The copyright holder for this this version posted September 15, 2020. . https://doi.org/10.1101/2020.09.14.20194043 doi: medRxiv preprint The adapted bead-based assays allowed us to investigate ligand-specific activities of macrophages. 306 Acidification responses to ligand stimulation occurred rapidly and reduced pH in the phagosomal 307 environment from neutral to below 5.5, which then facilitates later hydrolytic activity such as 308 proteolysis. In the healthy population, we found the variation of proteolysis (CV 45%) to be much 309 greater than that of acidification (CV 5%). However, for both activities, the variations remained 310 unchanged when macrophages were activated with either non-immunologic or pathogenic ligands, 311 suggesting a stable diversity of macrophage activities among different individuals in response to 312 various stimuli. Although inter-donor variations of these activities were preserved for different ligands, 313 acidification activities were much stronger for macrophages stimulated with IgG than with either TDM 314 or β-glucan. These findings, together with those of Podinovskaia et al, ( Mtb ligands could also influence results by interacting with different host pattern recognition receptors, 334 all these results together suggest similar patterns in antimicrobial activities of macrophages in response 335 to beads coated with TDM and live Mtb. 336 In both LTB and ATB groups, individuals showed heterogeneity in their macrophage antimicrobial 337 activities. The variation of macrophage proteolysis in LTB was slightly greater than the variation in 338 ATB, indicating wider capacity of this activity and thus its potential involvement in protecting healthy 339 subjects from developing active TB. Although the mean values of measured activities were not 340 significantly different in LTB and ATB, such diversity in proteolysis and acidification among 341 individuals allows for further study for underlying factors such as genetic variants in the phagocytic 342 genes which initiate the phagocytosis and phagosomal activities of macrophages (Thuong et al. 2016). 343 Antimicrobial activities of macrophages are important for controlling bacterial replication and could 344 be involved in bacterial dissemination from one locality to other organs. In this study, macrophages 345 from PTB and TBM showed similarity in their levels of antimicrobial activity as well as in their ability 346 to control Mtb, suggesting these activities are not associated with Mtb spreading in patients. 347 The ex vivo human MDM model used in this study allowed us to examine primary macrophages from 348 many subjects, from those without infection to those with different TB phenotypes, but also has some 349 limitations. Firstly, our model could not reflect the heterogeneity of lung macrophage populations 350 during the course of Mtb infection, which may influence the outcome of infections. Secondly, using a 351 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review) preprint
The copyright holder for this this version posted September 15, 2020. However, our results suggest that these activities cannot be solely responsible for the control of 359 bacterial replication or dissemination in TB. 360

Conflict of Interest 361
The authors declare that the research was conducted in the absence of any commercial or financial 362 relationships that could be construed as a potential conflict of interest.

Acknowledgments 373
We thank the participants in this study, and the clinicians and nurses at HTD and DTU 4 and 8 who 374 helped to perform this study. We would like to thank David Russell   . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 15, 2020. . https://doi.org/10.1101/2020.09.14.20194043 doi: medRxiv preprint  and 95% confidence interval. In (C, D) box plots represent the interquartile range (IQR) and median. 531 Vertical lines above and below each box extend to the most extreme data point that is within 1.5x IQR. 532 Each dot represents the activity index of hMDM from each subject. P values were determined by Mann-533 Whitney U test. 534 535 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 15, 2020. . https://doi.org/10.1101/2020.09.14.20194043 doi: medRxiv preprint 536 Figure 3. Antimicrobial activities and bacterial control of macrophages from LTB and ATB 537 participants. 538 (A) hMDM from LTB (n=35) and ATB (n=101) participants at day 7 were treated with beads coated 539 with TDM to measure acidification activity after 30, 60 and 90 min, or proteolytic activity after 60, 540 120 and 180 min. (B, C) Macrophages derived from cryopreserved monocytes from LTB (n=18) or 541 ATB (n=55) at day 7 were infected with a Mtb reporter strain at MOI 1. The bacterial intracellular 542 growth during 5 days of infection was assessed by the mCherry intensity readout (B) and the viability 543 of infected macrophages (C). Box plots represent the interquartile range (IQR) and median. Vertical 544 lines above and below each box extend to the most extreme data point that is within 1.5x IQR. Each 545 dot represents the activity index of hMDM from each subject. 546 547 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. the viability of infected macrophages (C). Box plots represent the interquartile range (IQR) and median. 556 Vertical lines above and below each box extend to the most extreme data point that is within 1.5x IQR. 557 Each dot represents the activity index of hMDM from each subject. 558 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 15, 2020. . (32-50) White blood cells -K/µl 5.9 (5.0-6.9) 5.9 (5.2-6.9) 6.2 (4.9-6.9) 9.4 (7.7-11.7) 9.7 (8.1 -11.7) 9.4 (7.2-12.0) Neutrophil -K/µl 3.4 (2.7-4.0) 3.4 (3-3.9) 3.4 (2.6-4.1) 6.9 (5.1-9.6) 6.6 (5.3 -8.3) 7.3 (4.8 -9.9) Lymphocyte -K/µl 1.9 (1.6-2.      Range: min-max; Mean±SD: Mean±standard deviation; CV: coefficient variation, calculated by percentage of SD divided by mean 610 P: P value for comparison of coefficient variation in either acidification or proteolytic activity among two groups 611