Identification of compounds that decrease numbers of Mycobacteria in human macrophages in the presence of serum amyloid P

Short abstract

Blocking the interaction between a constitutive signal and its receptor on macrophages potentiates anti‐tuberculosis M1 macrophages.

Abstract

Mϕs are a heterogeneous population of cells and include classically activated Mϕs (M1) and alternatively activated Mϕs (M2). Mϕs can change from M1 to M2 and vice versa in response to environmental stimuli. Serum amyloid P (SAP) is a constitutive plasma protein that polarizes Mϕs to an M2 phenotype, and part of this effect is mediated through FcγRI receptors. In an effort to find ways to alter Mϕs phenotypes, we screened for compounds that can block the SAP–FcγRI interaction. From a screen of 3000 compounds, we found 12 compounds that reduced the ability of fluorescently labeled human SAP to bind cells expressing human FcγRI. Based on cell surface marker expression, 8 of the compounds inhibited the effect of SAP on skewing human Mϕs to an M2 phenotype and in the presence of SAP polarized Mϕs to an M1 phenotype. In diseases, such as tuberculosis, M1s are more effective at killing bacteria than M2s. SAP potentiated the numbers of the mycobacterial strains Mycobacterium smegmatis and Mycobacterium tuberculosis in Mϕs. When added along with SAP, 2 of the compounds reduced intracellular Mycobacterium numbers. Together, these results indicate that the blocking of SAP effects on Mϕs can skew these cells toward an M1 phenotype, and this may be useful in treating diseases, such as tuberculosis.

Abbreviations

HEK293

human embryonic kidney cells 293

K_d

dissociation constant

M1

classically activated Mϕs

M2

alternatively activated Mϕs

M_reg

regulatory Mϕ

NCI

National Cancer Institute

NSC

National Service Center

OD600

OD at a wavelength of 600 nm

SAP

serum amyloid P

Introduction

Tuberculosis is caused by infection with the bacterium M. tuberculosis and causes 2 million deaths worldwide each year [1, 2]. Mϕs in the lung are a major line of defense against tuberculosis. However, after phagocytosis, M. tuberculosis can grow inside these Mϕs. Mϕs are heterogeneous and can be classified loosely as proinflammatory M1 or anti‐inflammatory M2 [3]. This heterogeneity also exists in tissues, such as the lung [4]. In response to signals, Mϕs can change from 1 phenotype to another [5]. For instance, activation of Mϕs to an M1 phenotype with IFN‐γ leads to strong microbicidal activity and production of proinflammatory cytokines. Conversely, treatment with IL‐4 or IL‐13 potentiates an M2a profibrotic phenotype involved in wound healing and immune regulation.

The ratio of M1 to M2 affects the outcome of a tuberculosis infection [6, 7]. M1 Mϕs activation inhibits M. tuberculosis infection [8]. Bacteria can survive inside Mϕs by inhibiting phagosome maturation, and this inhibition can be blocked with IFN‐γ treatment [9]. IFN‐γ also causes Mϕs to recruit nicotinamide adenine dinucleotide phosphate‐oxidase 2 to the phagosome to generate reactive oxygen species to kill the internalized bacteria [6, 10, 11]. In addition, IFN‐γ can induce autophagy, which is another way to kill intracellular bacteria [12]. In mice, M2 express high levels of arginase, which with the depletion of levels of the NO precursor arginine, decreases NO production, leading to impaired killing of bacteria [13]. M. tuberculosis use several strategies to evade killing by Mϕs. One is to block inflammatory cytokine production [14], and another is to drive M2 polarization [15, 16]. M. tuberculosis preferentially infects M2 and uses lipids to recruit the M2 cells [17]. Together, these results indicate that the discovery of ways to skew Mϕs to an M1 phenotype would be useful to fight tuberculosis.

SAP is a constitutive component of the circulating plasma [18]. One role of SAP is to bind material that should not be in the circulation, such as DNA, cell debris, and pathogens, and to help phagocytic cells engulf the material [19]. The phagocytic cells, such as monocytes, Mϕs, and neutrophils, use FcγR1 receptors to bind the SAP and the SAP‐bound material [20, 21]. SAP also signals through FcγR1 receptors to calm parts of the innate immune system [22, 23–24]. With respect to Mϕs, SAP potentiates the differentiation of monocytes into M2c M_regs and promotes the conversion of M1into M2c [23, 25]. In this report, we show that with the blocking of the interaction between the constitutive protein SAP and FcγR1 receptors, we can skew human Mϕs toward an M1 phenotype and as a result, decrease the ability of mycobacteria to thrive inside Mϕs.

MATERIALS AND METHODS

SAP binding to FcγRI

The binding of SAP labeled with DyLight 650 NHS Ester (Thermo Fisher Scientific, Waltham, MA, USA) to HEK293 cells transfected with plasmids driving expression of FcγRI–GFP and Fc∊ common γ‐chain was done as described [26], with the exception that cells were transfected using a 4D‐Nucleofector (Lonza, Allendale, NJ, USA) with the preset parameters for HEK293 cells. The SAP binding was measured as median fluorescence intensity. The fluorescence intensity from SAP binding was normalized as 100, and all compound groups were compared with SAP. Libraries containing 10 or 1 mM compounds were obtained from the NCI (Diversity and Mechanistic Diversity Sets). To reduce assay size, compounds were initially pooled into groups of 10 compounds immediately before doing the assays. The NCI compound collections are sent out in 96‐well plates, with 10 compounds per row. Pools 1A–9H corresponded to rows from compound plates 4762/62–4776/29 from the Diversity Set IV Library, pools 17A–23H corresponded to compound plates 4784/29–4790/18 from the Diversity Set IV Library, and pools 24A–32H corresponded to compound plates 4791/18–4800/18 from the Mechanistic Diversity Set II Library. The drug pool (1 µl) containing 100 µM of each drug in 100% DMSO (VWR International, Radnor, PA, USA) was mixed with 50 µl binding buffer containing labeled SAP, and this was then mixed with 50 µl cells. For follow‐up of active pools, the 12 compounds that inhibited SAP binding were ordered as specific compounds from the NCI repository, and these were used for subsequent studies, including the rescreening of these compounds for their ability to inhibit SAP binding. None of the 12 compounds has been used in clinical trials.

Macrophage polarization

Human blood was collected from adult volunteers who gave written consent and with specific approval from the Texas A&M University Institutional Review Board for human subjects. The only criterion for donor selection was that they described themselves as healthy. Blood collection and isolation of PBMCs were done as described previously [27]. The PBMCs were resuspended to 1 × 10⁶ cells/ml in RPMI 1640 (Lonza), and 250 µl was placed in each well of a type 89626 µ‐Plate 96‐well plate (Ibidi, Madison, WI, USA). After 2 h at 37°C in a humidified 5% CO₂ incubator, gentle pipetting was used to remove nonadherent cells [28, 29]. The medium was replaced with 250 µl serum medium [10% FCS (VWR Life Science Seradigm; VWR International) and 100 U/ml penicillin/streptomycin and 2 mM glutamine (Lonza) in RPMI]. At day 7, the medium was replaced with serum‐free medium [10 mM HEPES (Lonza); 2 mM glutamine, 100 U/ml penicillin/streptomycin, and 1 × ITS+3 (Sigma‐Aldrich, St. Louis, MO, USA); and 1 mM sodium pyruvate and 1 × nonessential amino acids (Lonza) in RPMI]. Human SAP (Millipore Sigma, Billerica, MA, USA) was buffer exchanged to 20 mM sodium phosphate, pH 7.4, using a 0.5 ml, 10 kDa cutoff Amicon Ultra centrifugal concentrator (Millipore Sigma). Compounds (stored in 100% DMSO) were diluted in RPMI to 100 µM. Where indicated, SAP and/or compounds were added at day 7 to the cells to 5 µg/ml and 1 µM final concentration, respectively. After an additional 3 d (day 10 after isolating the PBMC), cells were air dried, fixed, and stained by immunocytochemistry for CD54 or CD206, as described previously [30]. Approximately 100–200 Mϕs were examined, and the number of Mϕs and the number of the Mϕs that showed staining were recorded. All experiments used PBMCs from at least 3 different blood donors and 2 technical replicates for each donor. The average of the 2 technical replicates was then calculated, and the results are expressed as the means ± sem of the 3 or more averages, with each average from a different donor. We never used the same donor twice for a given experiment.

Bacterial cultures

M. smegmatis Mc² 155 [31] (a gift from Jim Sacchettini, BioBio, Texas A&M University, College Station, TX, USA) was grown at 37°C on 7H10 (Becton Dickinson, Franklin Lakes, NJ, USA) plates, supplemented with 0.2% glucose and 0.083% NaCl, or in shaking culture in 3 ml 7H9 (Becton Dickinson) medium, supplemented with 0.2% glucose, 0.083% NaCl, and 0.07% Tween 80 (Sigma‐Aldrich) [32]. mCherry‐labeled M. tuberculosis Mc² 7000 transformed with an mCherry hygroR plasmid [33, 34] (a gift from J. Sacchettini, BioBio, Texas A&M University) was grown as described previously [35].

M. tuberculosis infection assay; M. smegmatis infection assay

PBMCs were isolated and plated, and nonadherent cells were removed, as described above. The enriched monocytes in each well were cultured in 200 µl serum medium, supplemented with 10 ng/ml GM‐CSF (BioLegend, San Diego, CA, USA) for 7 d. The medium was then changed to serum‐free medium, with or without 5 µg/ml SAP and/or 1 µM compound. Three days later (day 10 after isolating PBMCs), cells from an overnight M. smegmatis culture at an OD600 of ∼1 were vortexed with ColiRollers Plating beads (Millipore Sigma) at high for 10 min with an analog vortex mixer (VWR International). After allowing the beads to settle out, the supernatant was decanted and sonicated at power 6 for 30 s, 6 times, using a Sonic Dismembrator F60 (Thermo Fisher Scientific). Clumped bacteria were removed by centrifugation at 50 g for 6 min at room temperature. The OD600 of the supernatant containing nonclumped bacteria was measured to estimate the bacterial concentration, with an OD600 of 1, equivalent to 3.13 × 10⁷ CFU/ml [36]. Before infection, cells in 1 well were washed with PBS and treated with Trypsin‐Versene mixture (Lonza) at 37°C for 5 min to detach cells, and cells were counted. M. smegmatis was diluted in RPMI/10% FCS/2 mM glutamine (and no antibiotics) and incubated with Mϕs for 4 h using a multiplicity of infection of 10, with respect to the cell count described above. After incubation, cells were gently washed with 37°C PBS, 3 times, and cultured with RPMI/10% FCS/2 mM glutamine/20 µg/ml gentamicin (Thermo Fisher Scientific) to kill bacteria outside of the cells [37, 38]. At 48 h after the infection, the cells were lysed with ice‐cold 0.5% Tween 20 in water. Serial dilutions of the cell lysates were plated on 7H10 plates, supplemented as described above, and colonies were counted after 3 d at 37°C.

mCherry expressing M. tuberculosis infection was done, as described above, with the exception that the M. tuberculosis was incubated with Mϕs for 4 h in RPMI/10% FCS/25 µg/ml pantothenic acid (Sigma‐Aldrich). After infection, the Mϕs were washed as above and cultured in RPMI/10% FCS/20 µg/ml gentamicin/25 µg/ml pantothenic acid. After 4 d, cells were imaged using an IN Cell Analyzer 2000 (GE Healthcare Life Sciences, Fairfield, CT, USA) in 6 randomly selected fields of view from each well, with a 20× objective in the Texas Red channel and a 0.05 s exposure. Phase images were also taken. Pairs of images were analyzed using IN Cell Developer Toolbox 1.8 (GE Healthcare Life Sciences), with the segmentation and sieve parameters adjusted to count Mϕs detected in the phase images with internalized fluorescent bacteria. Phagocytosis of bacteria was measured as the amount of bacteria inside Mϕs at 2 h after the cells were changed back into gentamicin medium.

Macrophage viability

Mϕ viability assays were done, as described for the infection assay, except that at 48 h after M. smegmatis infection or 4 d after M. tuberculosis infection, Mϕs were washed with HBSS (Sigma‐Aldrich) and then incubated for 15 min at 37°C with 165 nM SYTOX Green (Thermo Fisher Scientific) in HBSS. The medium was then removed and replaced with HBSS, and cells were imaged with a 20× objective on an IN Cell Analyzer 2000 in the FITC channel with an exposure time of 0.05 s.

M. smegmatis viability

M. smegmatis were grown in liquid culture to an OD600 of ∼1 and were then diluted in 7H9 medium, supplemented as described above to an OD600 of 0.005. In a well of an 80040LE 0910 clear, flat‐bottom, 96‐well plate (Thermo Fisher Scientific), 196 µl diluted bacteria were mixed with 4 µl compound (diluted in DMSO) and incubated at 37°C. After 24 h, 20 µl Deep Blue Cell Viability compound (Resazurin; BioLegend) was added to each well and incubated at 37°C for 2 h. Fluorescence was measured on a Synergy Mx plate reader (BioTek, Winooski, VT, USA), following the manufacturer's directions.

Statistics

Statistical analysis with t tests or 1‐way ANOVA with appropriate post‐test and curve fits were done using GraphPad Prism 4 (GraphPad, San Diego, CA, USA). Fisher's least significant difference test was used for the ANOVA post‐test during the initial screening stage to improve the sensitivity of the screening. Subsequently, the hit compounds were tested again to rule out potential false positives, and comparisons were done using Dunnett's post‐test. Significance was defined as P < 0.05.

RESULTS

Identification of compounds that decrease SAP binding to FcγR1

To identify compounds that can block SAP binding to FcγR1, we used Fluor 647 to label SAP and transfected HEK293 cells with GFP‐tagged FcγR1. The binding of labeled SAP to the FcγR1 expressed on HEK293 cells was measured by flow cytometry. With the use of this method, we previously observed a K_d of 4.6 nM for SAP binding to FcγR1 [26]. As a result of transfection efficiency, only some cells expressed FcγR1. As the FcγR1 was tagged with GFP, we were able to identify FcγR1⁺ and FcγR1⁻ cells by flow cytometry. The binding of SAP to the FcγR1⁻ cells was used to measure nonspecific binding, which was subtracted from the SAP binding to FcγR1⁺ cells to obtain SAP binding to FcγR1. Pools of 10 compounds, containing 1 µM final concentration of compound in the assay, were assayed for their ability to reduce the binding of 80 nM SAP (10 µg/ml) to FcγR1. From a screen of 200 pools, 9 pools decreased SAP binding, and 11 pools increased SAP binding ( Fig. 1 ). We retested the 9 compound pools that significantly reduced the binding of SAP to FcγR1 and additionally retested 15 pools that appeared to decrease SAP binding, although without achieving statistical significance ( Fig. 2 ). Six of the pools again showed an inhibition of SAP binding. The 60 individual compounds from these 6 pools were then assayed for inhibition of SAP binding at a final compound concentration of 1 µM, and 12 compounds decreased SAP binding ( Fig. 3 ). When retested at a variety of compound concentrations, all 12 compounds significantly decreased SAP binding ( Fig. 4 ). The structures of the compounds are shown in Table 1 .

Figure 1.

Identification of compound pools that decrease SAP binding to HEK293 cells expressing FcγR1.

HEK293 cells expressing FcγR1 were incubated with 10 µg/ml fluorescently labeled SAP and the indicated compound pools. Flow cytometry was then used to measure SAP binding to the cells. The binding was measured as median fluorescence intensity. The binding of the labeled SAP to FcγR1‐negative cells was used to estimate nonspecific binding, and this was subtracted from total binding. Values are means ± sem; n = 3 independent experiments. *P < 0.05; **P < 0.01 (1‐way ANOVA, Fisher's test compared with SAP and no compound).

Figure 2.

Second round of screening.

Compound pools that significantly decreased SAP binding, along with pools that potentially decreased binding, were rescreened as in Fig. 1. Values are means ± sem; n = 3 independent experiments. *P < 0.05 (1‐way ANOVA, Dunnett's test compared with SAP and no compound).

Figure 3.

Screening of individual compounds for ability to decrease SAP binding.

Individual compounds from the pools that significantly decreased SAP binding in Fig. 2 were screened as in Fig. 1. Values are means ± sem; n = 3 independent experiments. *P < 0.05 (1‐way ANOVA, Dunnett's test compared with SAP and no compound).

Figure 4.

The reduction of SAP binding as a function of compound concentration.

Compounds that significantly decreased SAP binding in Fig. 3 were tested at the indicated concentrations for their ability to decrease the binding of fluorescently labeled SAP to HEK293 cells expressing FcγR1 as in Fig. 1. Values are means ± sem; n = 3 independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 (1‐way ANOVA, Dunnett's test compared with SAP and no compound). Lines are nonlinear regression fits to a 1‐site competitive binding curve, using a SAP concentration of 80 and 4.6 nM for the K_d of SAP binding to FcγR1.

Table 1.

Compounds and their effects on Mϕ phenotype and mycobacteria infection

Compound code	NSC number	Structure	CD54 expression	CD206 expression	M. smegmatis infection	M. tuberculosis infection
4D7	88882
19E3	7419		↑		↓
19E8	91356		↑
21G2	5856
21G3	57608
21G8	15910
27H2	622690		↑	↓	↓
27H8	643162		↑		↓	↓
28F2	327697		↑	↓	↓	↓
28F5	77021			↓	↓
28F6	262665			↓	↓	↓
29C2	2186			↓		↓

Some compounds that decrease SAP binding to FcγR1 decrease SAP effects on macrophages

SAP potentiates the polarization of Mϕs to an anti‐inflammatory M_reg (M2c) phenotype [23, 25]. To determine if the compounds that reduce SAP binding to FcγR1 can reduce the ability of SAP to potentiate M2c, we examined the effects of the compounds in the presence or absence of SAP on the Mϕ phenotype. Human PBMCs were cultured for 6 d to allow the monocytes to differentiate into Mϕs, and the cells were then washed and cultured in the presence or absence of SAP and compounds for 3 d. The cells were then fixed and stained, and the percent of the Mϕs that stained for a given marker was measured. Representative images for the control group are shown in Supplemental Fig. 1. ICAM‐1 (CD54) is a proinflammatory M1marker [39, 40–41]. In the absence of SAP, 4D7, 21G2, 27H2, and 28F6 significantly decreased the percent of Mϕs showing ICAM‐1 staining ( Fig. 5A ). SAP also decreased the percent of ICAM‐1⁺ Mϕs, and compared with SAP alone, 19E3, 19E8, 27H2, 27H8, and 28F2 in the presence of SAP significantly increased the percent of ICAM‐1⁺ Mϕs (Fig. 5A). CD206 is an M_reg (M2c) marker [42]. In the absence of SAP, 4D7, 21G3, 28F5, and 28F6 significantly increased the percent of Mϕs that were CD206⁺ (Fig. 5B). SAP also increased the percent of CD206⁺ Mϕs, and compared with SAP alone, 27H2, 28F2, 28F5, 28F6, and 29C2 in the presence of SAP significantly decreased CD206⁺ Mϕs (Fig. 5B).

Figure 5.

Some compounds that decrease SAP binding affect Mϕ marker expression.

Human PBMCs were cultured for 6 d to allow Mϕ differentiation and were then cultured for an additional 3 d in the presence or absence of SAP and the indicated compounds. (A) Cells were fixed and stained for the M1 marker ICAM‐1 (CD54) or (B) the M2 marker CD206. The number of morphologically identifiable Mϕs that stained for the indicated marker was then counted and expressed as a percentage of Mϕs. Values are means ± sem; n = 3 different donors. #P < 0.05 (t test compared with control); *P < 0.05 (1‐way ANOVA, Dunnett's test comparing compounds in the absence of SAP with the control and comparing compounds incubated with SAP with SAP alone).

SAP‐polarized macrophages become M. tuberculosis tolerant

M2 appear to be more associated with tuberculosis than M1 [43, 44, 45, 46–47]. In agreement with this, we observed that compared with unpolarized human Mϕs, human Mϕs polarized to an M1 phenotype with IFN‐γ [48] showed decreased numbers of intracellular M. smegmatis at 48 h after infection and a decreased percentage of cells infected with M. tuberculosis at 5 d after infection ( Fig. 6A ). In addition to M1 and M_reg, Mϕs can also be polarized to a profibrotic M2a phenotype with IL‐4 [49, 50], and we observed that compared with unpolarized Mϕs, polarization of Mϕs to an M2a phenotype with IL‐4 resulted in increased numbers of intracellular M. smegmatis and an increased percentage of cells infected with M. tuberculosis (Fig. 6A). The survival of mycobacteria within Mϕs over infection was measured by comparing the 0 h time point after infection with 48 h (M. smegmatis) or 5 d (M. tuberculosis) after infection. In the absence of added factors, the number of M. smegmatis cells decreased, whereas the number of M. tuberculosis cells increased (Supplemental Fig. 2).

Figure 6.

SAP potentiates the number of intracellular mycobacteria.

(A) Human PBMCs were cultured in 10% FCS/RPMI with 10 µg/ml GM‐CSF for 8 d and treated with IFN‐γ or IL‐4 for an additional 2 d. Cells were then infected with M. smegmatis or recombinant M. tuberculosis (Mtb) expressing a fluorescent protein. M. smegmatis‐infected cells were lysed, bacteria were plated at 48 h postinfection, and colonies were counted. M. tuberculosis fluorescence was imaged at 5 d postinfection. (B) PBMCs were similarly cultured with GM‐CSF for 7 d and then cultured for an additional 3 d in the presence or absence of SAP. Mycobacteria were measured as in A. Values are means ± sem; n = 3 different donors. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 (t test).

Mycobacterium infection appears to potentiate M2 [51]. SAP also potentiates M2 and inhibits M1 [23, 52]. To test the prediction that SAP potentiates Mycobacterium infection of Mϕs, we added Mycobacterium bacteria to Mϕs in the presence or absence of SAP. SAP potentiated the number of intracellular M. smegmatis at 48 h after infection and potentiated the percentage of Mϕs infected with M. tuberculosis at 5 d after infection (Fig. 6B).

Some SAP–FcγR1 inhibitors decrease mycobacteria numbers in infected macrophages

As some of the compounds that decrease SAP binding to FcγR1 also inhibit the effect of SAP on Mϕs, we determined whether any of these compounds might affect the number of Mycobacterium in Mϕs. Human PBMCs were cultured for 6 d to allow the monocytes to differentiate into Mϕs, and the cells were then washed and cultured in the presence or absence of SAP and compounds for 3 d. The medium was then removed, and medium containing Mycobacterium was then added. After 4 h, the Mϕs were washed to remove noninternalized bacteria and cultured for 2 or 4 d. The level of internal Mycobacterium was then measured. In the absence of SAP, 27H2 increased the number of intracellular M. smegmatis at 48 h after infection ( Fig. 7A ). SAP also increased the number of intracellular M. smegmatis, and compared with SAP alone, in the presence of SAP, 19E3, 27H2, 27H8, 28F2, 28F5, and 28F6 decreased the number of intracellular M. smegmatis (Fig. 7A). In the absence of SAP, 21G8 and 28F5 increased the percentage of Mϕs infected with M. tuberculosis at 4 d after infection (Fig. 7B), and compared with SAP alone, in the presence of SAP, 27H8, 28F2, 28F6, and 29C2 decreased, whereas 21G8 increased the percentage of Mϕs infected with M. tuberculosis (Fig. 7B).

Figure 7.

Some compounds inhibit the number of intracellular mycobacteria in the presence of SAP.

PBMCs were cultured as in Fig. 6B, with the exception that the indicated compounds were added in the presence or absence of SAP for the 3 d before bacterial infection. (A) M. smegmatis were measured as in Fig. 6. (B) The percent of Mϕs infected with M. tuberculosis was measured as in Fig. 6. Values are means ± sem; n = 3 different donors. #P < 0.05 (t test compared with control); *P < 0.05; **P < 0.01; ***P < 0.001 (1‐way ANOVA, Dunnett's test).

Confirmation of lead compounds’ activity on macrophage infection

We found that in the presence of SAP, 7 compounds inhibited either M. smegmatis or M. tuberculosis numbers inside Mϕs. To determine if these compounds are directly inhibiting Mycobacterium growth and/or proliferation, we added these compounds to M. smegmatis for 24 h. We then quantified the amounts of live bacteria with the fluorescent dye Resazurin. Compound 27H2 killed bacteria at concentrations of 4 µM and above, similar to the killing ability of the antibiotic Rifampicin ( Fig. 8 ). However, in the mycobacteria infection assay, compound 27H2 did not kill bacteria directly because the concentration used was 1 µM, and the compounds were washed off before adding in bacteria. The other 6 compounds did not significantly affect M. smegmatis viability (Fig. 8). These results indicate that 6 of the compounds that we identified improve the response of SAP‐exposed Mϕs to mycobacteria infection.

Figure 8.

The direct effect of compounds on M. smegmatis proliferation and growth.

M. smegmatis was grown for 24 h in the presence of the indicated concentrations of compounds and then incubated with Resazurin for 2 h. Growth was measured as the fluorescence intensity of reduced Resazurin. Values are means ± sem; n = 3 independent experiments. *P < 0.05; **P < 0.01 (t test compared with DMSO control).

Another possible reason that the compounds might appear to decrease the numbers of bacteria inside Mϕs is that the compounds decrease Mϕ viability. To test this possibility, we did compound treatments and Mycobacterium infections, as described for Fig. 7, and measured Mϕ viability using SYTOX Green to stain for dead Mϕs. In the M. smegmatis model, we observed that ∼10% of the Mϕs were dead at 48 h, after adding bacteria to Mϕs, and that SAP did not significantly affect this percentage ( Fig. 9A–F ). Of the 6 compounds, which in the presence of SAP, decreased the number of intracellular M. smegmatis in Mϕs, only 5 µM 28F6 increased the percentage of dead Mϕs (Fig. 9F). At 4 d after M. tuberculosis infection, SAP did not significantly increase the number of dead Mϕs (Fig. 9G). Of the 4 compounds, which in the presence of SAP, decreased the percentage of Mϕs infected with M. tuberculosis at 4 d after infection, 29C2 increased the percentage of dead Mϕs in the absence of SAP, but in the presence of SAP, none of the 4 compounds had a significant effect on Mϕ viability (Fig. 9G). Together, these results suggest that 5 µM 28F6 decreases Mϕ viability in the presence of SAP, whereas the other compounds appear to potentiate the response of Mϕs to Mycobacterium infection in the presence of SAP rather than inhibiting Mϕ viability.

Figure 9.

The effect of compounds on Mϕ viability.

Mϕs were cultured and infected as in Fig. 6. (A–F) Mϕ viability was measured by staining with SYTOX Green at 48 h after M. smegmatis infection, and the percent of dead cells in images was calculated. (F) *P < 0.05 for 5 µg/ml SAP (t test compared with control). (G) Mϕ viability was measured at 5 d after M. tuberculosis infection, as in (A–F). Values in A–G are means ± sem; n = 3 different donors. *P < 0.05 compared with DMSO control (1‐way ANOVA, Dunnett's test).

SAP and the compounds could potentially alter the ability of Mϕs to bind to and/or internalize bacteria. To determine if SAP and/or the identified compounds affect these processes, we infected Mϕs with bacteria, as described for Fig. 7, and then lysed Mϕs immediately after washing off bacteria that had not been internalized. Neither SAP nor the 6 compounds that decreased M. smegmatis affected the phagocytosis of M. smegmatis ( Fig. 10A ), and neither SAP nor the compounds that decreased M. tuberculosis affected the phagocytosis of M. tuberculosis (Fig. 10B).

Figure 10.

The effect of compounds on phagocytosis of bacteria by Mϕs.

Mϕs were cultured and infected as in Fig. 6. After infection, Mϕs were treated with gentamycin for 2 h to remove extracellular bacteria. (A) M. smegmatis‐infected cells were then lysed immediately after the 2 h gentamycin treatment, bacteria were plated, and colonies were counted. (B) M. tuberculosis fluorescence was imaged immediately after the 2 h gentamycin treatment. Values are means ± sem; n = 3 different donors. No values were significantly different from the DMSO controls (1‐way ANOVA, Dunnett's test).

DISCUSSION

SAP is a constitutive component of plasma [18], and SAP also appears to be present in the extracellular space of tissues [54]. We [26] and others [20] found that SAP regulates monocyte and Mϕ phenotypes, at least in part, through its binding to FcγRI. We screened a small compound library and found 12 compounds that inhibit human SAP binding to cells expressing human FcγRI. Eight of the compounds blocked the ability of human SAP to increase the expression of an M2 marker or decrease the expression of an M1 marker. As predicted, pre‐exposure of human Mϕs to human SAP potentiated the number of mycobacteria in the Mϕs. Some, but not all, of the compounds reduced this effect of SAP. Compounds 27H8, 28F2, and 28F6 effectively ameliorate both M. tuberculosis and M. smegmatis infection, although 28F6 may have a moderate cytotoxic effect at high concentration. Together, these results indicate that the blocking of the human SAP–human FcγRI could be a potential way to decrease the ability of mycobacteria to proliferate in human Mϕs. This is in contrast to the effect of mouse SAP on mouse alveolar Mϕs, where mouse SAP decreases the uptake and number of M. tuberculosis in mouse alveolar Mϕs [55]. This difference may be a result of either differences between mice and humans or different responses between circulating monocyte‐derived Mϕs and alveolar Mϕs.

The 12 compounds identified in the initial screen were only able to inhibit SAP binding to cells expressing FcγRI by ∼20%, indicating that the identified compounds are far from optimal binding inhibitors. How the compounds inhibit SAP binding is unknown, as they could directly interfere at the SAP‐FcγRI binding site, allosterically alter the extracellular conformation of either of the 2 proteins, or act inside cells to alter allosterically the conformation of FcγRI. If a compound does bind to FcγRI in such a way that it interferes with the ability of FcγRI to bind the Fc domain of an antibody that has bound to a pathogen, such a compound would possibly have the undesired side effect of decreasing the ability of cells to uptake antibody‐bound pathogens. Although 27H2, 27H8, and 28F2 all have dimethylamines on an aromatic; 27H8 and 28F6 have α, β unsaturated carbonyl groups; and 19E3, 27H2, 27H8, 28F2, 28F5, and 28F6 all have electrophilic centers, there does not appear to be any consistent structural feature in the 12 compounds or the compounds that have a specific effect on a marker or numbers of Mycobacterium in Mϕs (Table 1). This diversity supports the idea that the compounds could affect Mϕs at a variety of binding sites.

Some compounds that inhibited SAP binding had no significant effect on the ability of SAP to decrease expression of the M1 marker CD54 and increase expression of the M2 marker CD206, suggesting that they may have partially activated FcγRI while inhibiting SAP binding. Several of the compounds mimicked the effect of SAP on CD54 and/or CD206 expression (Fig. 5), indicating that they may interact directly with FcγRI. Compound 27H2 decreased expression of the M1 marker CD54 in the absence of SAP but increased CD54 expression in the presence of SAP, whereas compounds 28F5 and 28F6 increased expression of the M2 marker CD206 in the absence of SAP but decreased CD206 expression in the presence of SAP. This indicates that these compounds can act as SAP agonists or SAP antagonists depending on whether SAP is present.

Under some conditions, such as SAP with 27H8, the number of cells expressing CD54 increased, and the number of cells expressing CD206 also increased, suggesting that some cells may have been expressing both markers. Under other conditions, the percent of CD54⁺ cells and the percent of CD206⁺ cells did not add up to 100, indicating that some cells may have expressed neither marker. Both results suggest that as has been observed by others [56, 57], in the presence of some of the compounds, some Mϕs develop a phenotype that is not perfectly M1 or perfectly M2.

In the presence or absence of SAP, compounds 27H8 and 28F2 did not affect the initial uptake of mycobacteria by Mϕs, the viability of mycobacteria in the absence of Mϕs, or the viability of Mϕs. However, in the presence of SAP, these 2 compounds reduced the number of M. smegmatis bacteria in Mϕs at 2 d after infection and the percent of Mϕs containing M. tuberculosis at 5 d after infection. Together, this indicates that in the presence of SAP, 27H8 and 28F2 may have induced some Mϕs to kill their ingested bacteria. Comparison of Figs. 10B with 7B indicates that over 4 d of M. tuberculosis infection, the percent of infected Mϕs increases. An alternative possibility is thus that in the presence of SAP, 27H8 and 28F2 may have inhibited the spread of M. tuberculosis from one Mϕ to another.

In addition to helping Mϕs kill internalized mycobacteria, the altering of the Mϕ phenotype may be useful for the treatment of diseases, such as cancer and obesity [58, 59]. Together with our results, this suggests that the blocking of SAP effects on Mϕs may be useful for a variety of diseases. However, SAP also appears to help prevent the development of fibrosis [60], so this approach would need to be used with some caution.

AUTHORSHIP

All authors contributed to the design and execution of experiments, analysis of data, and writing of the manuscript.

DISCLOSURES

W.X. has no existing or potential conflicts of interest. N.C. and R.H.G. are inventors on patent applications for the use of small‐molecule SAP mimetics for the treatment of fibrosing diseases and neutrophil‐driven diseases. R.H.G. is an inventor on patents for the use of SAP as a therapeutic for fibrosing diseases and is a cofounder of, and has equity in, Promedior, a company developing SAP as a therapeutic for fibrosing diseases. However, note that this manuscript addresses blocking the effects of SAP rather than mimicking or augmenting the effect of SAP.

Supporting information

Supplementary Material Files

Supplementary Material Files

Supplementary Material Files