Variability in the virulence of specific Mycobacterium tuberculosis clinical isolates alters the capacity of human dendritic cells to signal for T cells (2024)

Abstract

MATERIALS AND METHODS

Human blood peripheral mononuclear cells, MDDCs and T lymphocytes -Mononuclear cells were isolated from buffy coats provided by the State BloodTransfusion Center of Zacatecas, SSA (Guadalupe, Zac., Mexico) by density gradientusing Ficoll-Paque PLUS (GE Healthcare, UK). The donors were healthy individualscomplying with the following criteria: 18-55 years old, minimum weight of 50 kg,fasting conditions for at least 8 h, and with no history of Hepatitis type B or C,HIV/AIDS, syphilis, organ transplants, epilepsy, tuberculosis, cardiovasculardisease, or cancer. Exclusion criteria included recreational drug use, mentaldisease, women who were pregnant or lactating, tattoos or skin perforations (12months previous to the donation), history of diabetes, surgery, mononucleosis,toxoplasmosis or meningitis (in the last six months), having received vaccines (inthe last 28 days), alcohol or narcotics use (in the last 12 h). Different sets ofindividuals were used for the distinct parameters evaluated in the present protocol.The number of individuals evaluated in each assay is annotated in the correspondingfigure legend. Whole blood from individuals positive for tuberculin was used onlyfor T-cell proliferation assays.

In order to generate MDDCs, mononuclear cells (5 × 10⁶) were seeded in a75 cm² flask using RPMI-1640 medium supplemented with 10% foetal calfserum (FCS) (Gibco Life Technologies, USA), 1% pyruvate (Sigma Aldrich, USA), 0.1%2-beta-mercaptoethanol (Gibco Life Technologies, USA), recombinant IL-4 (15 ng/mL)(Peprotech Inc., USA), and granulocyte macrophage colony stimulating factor (GM-CSF,200 ng/mL) (GRAMAL, Probiomed Lab, Mexico), and cultured at 37ºC in a humidifiedatmosphere at 5% CO₂. At days 3 and 5, the cells wereadditionally supplemented with IL-4 and GM-CSF. At day 7, MDDCs were collected bycentrifugation and used for functional analyses. Alternatively, mononuclear cellsisolated from healthy volunteers with positive tuberculin tests (range of 15 to 25mm) were used to generate MDDCs and to purify CD4⁺ T cells by a magneticbead approach using the CD4⁺ negative selection kit (Miltenyi, USA)according to manufacturer’s instructions.

Mtb strains, culture and storage - All manipulations with the Mtbstrains were performed in a dedicated BSL-3 laboratory. The Mtb strain H37Rv wasused in order to establish an experimental reference for the comparative analysesdone with the clinical isolate strains Ph1 (highly virulent, with no induction ofprotective immune response in mice) and Ph4 (less virulent, provoking a protectiveadaptive immune response).12 The strain Ph1, for spoligotyping, belongs to the family EAI and it issusceptible to all the first-line antibiotics used for TB treatment, while strainPh4 belongs to the Haarlem family and is streptomycin-monoresistant. All strainswere grown in Middlebrook 7H9 medium (BD-Diagnostic Systems, USA) supplemented with10% Middlebrook Oleic Albumin Dextrose Catalase Growth Supplement (OADC)(BD-Diagnostic Systems, USA). Strains were grown to reach exponential phase, andculture concentrations were measured by measuring optical density at 600 nm.Bacterial aliquots were stored at -80ºC until their use. Bacillary viability wastested using the colony forming unit (CFU) assay, growing serial dilutions in 7H10agar plates (BD-Diagnostic Systems, USA) supplemented with 10% OADC (BD-DiagnosticSystems, USA) for 14 and 21 days.

Internalisation of Mtb and infection of MDDCs - Bacterial aliquots(H37Rv, Ph1, and Ph4) were thawed at room temperature, and the bacterial aggregatedeclumping was achieved by vortexing with borosilicate beads for 5 min andcentrifuging at 2040 × g for 5 min. MDDC infection was performed ata multiplicity of infection (MOI) of 5 (5 bacteria to 1 cell). At 2 h post-infection(hpi), MDDCs were washed with RPMI-1640 medium to remove non-internalised bacteria.At 24 hpi, MDDCs were recollected by centrifugation and fixed with 4%paraformaldehyde (PFA) for 30 min at room temperature. In order to evaluateinternalisation of Mtb strains, the fixed MDDCs were centrifuged in pretreatedslides (Biocare Medical, Concord CA, USA) using a Cytocentrifuge (Wescor Cytopro7620, USA), and Ziehl-Neelsen staining was performed to identify the bacilliassociated/within cells. The percentage of MDDCs with at least one associatedbacterium was calculated after counting at least 100 MDDCs/sample in high powerfields using a Carl Zeiss inverted Axiovert M-200 microscope (Zeiss, Germany).

Mtb intracellular growth in human MDDCs - The bacterial strainswere cultured at 37ºC in Middlebrook 7H9 medium supplemented with 10% OADC and 0.05%Tween-80 (Sigma-Aldrich, USA). During exponential growth, the bacteria werecentrifuged (2000 × g) for 15 min and resuspended in 1× phosphatebuffered saline (PBS). Clumps were dissociated by passages through a 26-G needle,and then resuspended in RPMI-1640 medium containing 10% FBS. The mycobacterialconcentration was determined by measuring optical density at 600 nm (OD600). To testthe intracellular growth capacity of these strains, MDDCs were collected at day 7 ofculture and seeded in 24-well plates at a density of 5 × 10⁵ cells perwell. These cells were then infected with each Mtb strain individually at amultiplicity of infection (MOI) of 0.2 bacteria per cell in RPMI-1640 medium with10% FBS for 4 h. Cells were then washed twice with 1× PBS before addition ofRPMI-1640/10% FBS. At the indicated time points, the cells were lysed in a 0.1%Triton (Sigma-Aldrich, USA) lysing solution. Serial dilutions of the resultingbacterial suspension were plated on Middlebrook 7H11 solid agar supplemented with10% OADC (BD-Diagnostic Systems, USA) and incubated for 14-21 days at 37ºC for CFUscoring.

Cell-death assessment of MDDCs infected with Mtb strains - Toevaluate apoptosis and necrosis, we analysed MDDCs at day 7 of differentiation usingthe FITC Annexin V Apoptosis Detection Kit II (BD Biosciences, USA), according tothe manufacturer’s instructions. Briefly, MDDCs were infected with each Mtb strainindividually at a MOI of 5 in RPMI-1640 medium with 10% FBS for 4 h. MDDCs were thenwashed twice with 1× PBS before addition of RPMI-1640/10% FBS. After 24 h, MDDCswere washed twice with cold PBS and resuspended in 1× binding buffer, stained with 5µL of Annexin V and 5 µL of Propidium Iodide, and gently vortexed and incubated for15 min at room temperature, protected from light. MDDCs were then and washed usingPBS and resuspended in 4% PFA for 30 min at room temperature for fixation. Finally,the viability status of all MDDC populations was acquired using the FACS Canto IIcytometer (BD Biosciences, USA). For data analysis, Flow Jo VX (Tristar, USA) wasused.

T-cell proliferation assay - The ability ofMtb-infected MDDCs to stimulate CD4⁺ T cells wasassessed using an autologous co-culture system, as previously described.14 Briefly, autologous CD4⁺ T cells (the responders) were freshlyisolated and purified from healthy donors positive for the tuberculin test on theday that the co-culture was initiated. A total of 1.5 × 10⁴ 5-(and6)-Carboxyfluorescein diacetate succinimidyl ester (CFSE) stained CD4⁺ Tcells (responder cells) were added to each well in 200 μL complete media onrounded-bottom, 96-well plates (BD, Pharmingen, USA). Autologous MDDCs (thestimulators) were differentiated until day 6 and pulsed for 24 h with LPS (100ng/mL) (Sigma, Germany). Later, MDDCs were infected with each Mtb strainindividually at a MOI of 0.2 in RPMI-1640 medium with 10% FBS for 4 h, washed twicewith 1× PBS, harvested, and added to the responder cells at a ratio(stimulator:responder) of 1:2.5, 1:5, 1:10, or 1:20. As a control for the responseto Mtb antigens, a group of non-infected MDDCs were grown until day 6 and were thenpulsed for 24 h with a peptide corresponding to the partial sequence of the Mtbearly secretory antigenic target (ESAT-6) protein (20 μg/mL) for 24 h. Of note, thepartial ESAT-6 protein a peptide with the sequence H2N-LNNALQNLARTISEAG-COOH wassynthesised at Fundación Instituto de Inmunología de Colombia, Bogotá Colombia.After six days of culture in 5% CO₂ at 37ºC, the cells were harvested,stained, and gated for CD3 and CD4 positivity, and then analysed for dilution of theCFSE intensity by flow cytometry.

Statistical analyses - The corresponding statistical analysis isannotated in each figure legend. Data was compared using two-way ANOVA analysis withBonferroni’s post-hoc test for internalisation index, Wilcoxon analysis for CFU,while comparison inside the groups was performed with the Kruskal-Wallis test andDunn’s post-test assuming non-normally distributed data for all the other analysedvariables. Statistical analysis was performed with GraphPad Prism software v.5.0(San Diego, CA, USA). P ≤ 0.05 was considered as the level of statisticalsignificance.

Ethical considerations - The present protocol was reviewed andauthorised by the National Research Committee of Instituto Mexicano del SeguroSocial (The Mexican Institute of Social Security, IMSS), that includes asubcommittee for ethical approval (agreement number R-2014-785-042). Informedconsent was obtained from all individual participants included in the study.

RESULTS

Virulence of Mtb strains does not affect recognition/internalisation byhuman MDDCs - In order to assess whether the virulence of the Mtbclinical isolates Ph1 and Ph4 affect their capacity to be recognised/internalised byhuman MDDCs, we infected MDDCs with each bacterial strain, and the bacillary loadwas determined by Ziehl-Neelsen staining. Our results show only a tendency towards ahigher recognition/internalisation of low virulence strains compared to thereference H37Rv strain (Fig. 1A). This wasconfirmed by an analysis of the number of bacilli bound/internalised per MDDC, whichyielded no difference among any clinical isolate compared to H37Rv (Fig. 1B). Therefore, these results suggest thatvariations in virulence do not significantly affect the capacity of Mtb clinicalstrains to be recognised or internalised by MDDCs.

MDDCs control the intracellular growth of the Mtb clinical strainsindependent of virulence - In order to investigate whether thevirulence of Mtb strains modulates their capacity to proliferate in human MDDCs, weperformed CFU assays. After 4 hpi (day 0), we confirmed there was no significantdifference in the bacterial charge in MDDCs infected with the different Mtb strains(Fig. 2). More specifically, we observedthat MDDCs were able to control the intracellular growth of both Ph1 and Ph4 Mtbclinical strains after 120 hpi (day 5). However, at day 5, we did observe a higherinhibition of growth of the Ph1 strain compared to the reference strain (Fig. 2). Altogether, the results suggest thatthese Mtb clinical isolate strains fail to colonise human MDDCs, and that thedifference in virulence among these strains does not play a role in thisprocess.

Mtb clinical strain bearing low virulence leads to high levels of apoptosisin MDDCs - We examined whether the difference in virulence among theMtb clinical isolate strains affects the mode of cell death in infected MDDCs. Flowcytometry revealed that at 24 hpi, the strain with lower virulence (Ph4) inducedmore apoptosis in MDDCs compared to the other Mtb strains and to the uninfectedcells (Fig. 3A-C). As shown in Fig. 3C, the virulence among the Mtb strains doesnot significantly affect the induction of necrosis in MDDCs. Overall, these resultsshow that the hypovirulent Mtb clinical isolate strain (Ph4) differs considerably inits capacity to induce apoptosis, but not necrosis, in infected MDDCs when comparedto the hypervirulent strain (Ph1) and the experimental H37Rv reference strain.

Virulence of Mtb strains does not affect the activation of MDDCs -In order to assess whether the virulence of the Mtb clinical isolate strains affectsthe activation of human DCs, we evaluated the expression of cell surface markersduring infection by flow cytometry. As expected, MDDCs infected with any Mtb straindisplayed significantly higher levels (albeit only a tendency for H37Rv) of theactivation marker CD83, as compared to uninfected cells. Yet, there were nodifferences observed between the Mtb clinical isolate strains or compared to thereference H37Rv strain (Fig. 4A). In addition,we assessed the expression levels of receptors that are important for antigenpresentation, such as CD86 and HLA-DR. Although MDDCs infected with the clinicalisolate bearing low virulence (Ph4) displayed higher levels of CD86 compared touninfected cells, we noticed there were no significant differences detected amongall the infected groups (Fig. 4B). The same wastrue for the expression levels of HLA-DR among all the infected groups, althoughthere was a tendency for higher expression compared to uninfected cells (Fig. 4C). Altogether, these results suggest thatthe difference in virulence among the Mtb strains does not affect the activation ofMDDCs.

Virulence of Mtb strains affects the capacity of MDDCs to activateCD4⁺T cells - In order to evaluate whether the virulence of the Mtbclinical isolate strains modulates the capacity of MDDCs to activate T lymphocytes,we set up an autologous co-culture system to evaluate T-cell proliferation based onthe dilution of the CFSE dye, as measured by flow cytometry. As expected, T cellsthat were co-cultured with uninfected MDDCs did not proliferate, as measured by theundiluted CFSE dye (Fig. 5A-B). By contrast, weobserved T-cell proliferation with MDDCs pulsed with the ESAT-6 peptide, whichserved as positive control, at any tested ratio (MDDC:T cell), confirming thepresence of an antimycobacterial immune response in healthy donors positive for thetuberculin test (Fig. 5C). Likewise, we noticedthat MDDCs infected with the H37Rv reference strain were also capable of inducingrobust T-cell proliferation (Fig. 5D). However,comparison of MDDCs infected with the different Mtb clinical isolates resulted instriking differences. On the one hand, MDDCs infected with the clinical isolatebearing low virulence (Ph4) were not able to prevent cell death of autologous Tcells, as the co-cultures did not remain viable for the entire six day co-cultureperiod (data not shown). On the other hand, MDDCs infected with the clinical isolatebearing high virulence (Ph1) were capable of inducing T-cell proliferation, but lessefficiently than MDDCs infected with the H37Rv reference strain (Fig. 5A,D). These results were consistent for thefive independent tuberculin-positive donors tested in the antigen presentationassays, even when there were not statistic differences between the rates ofproliferation induced by Ph1 compared to H37Rv. Collectively, these results suggestthat the virulence difference among the Mtb clinical strains is an important factorthat can affect antigen presentation.

DISCUSSION

Most of the information about Mtb strain variation and immunopathology derives fromexperimental animal models.13 Yet, this information cannot be extrapolated directly to human TB infection,and thus highlights the need to study Mtb strain variation using human leukocytes.In the present study, we investigated whether Mtb clinical isolates obtained from aprospective population-based study of pulmonary TB patients in Southern Mexico,whose virulence varied dramatically from high (Ph1) to low (Ph4) in a mousemodel,12 differ in their interaction with human DCs in comparison to the standardexperimental strain H37Rv. We believe this study makes the following contributionsin advancing our understanding of how these Mtb clinical isolates differ in theirinteraction with human DCs.

First, we showed that the Ph1 hypervirulent strain diminishes the capacity of humanDCs to activate autologous CD4⁺ T cells. Compared to DCs infected withH37Rv, Ph1-infected DCs displayed a poor capacity to induce proliferation ofautologous T cells in our in vitro co-cultures. This is not due toa deficiency in DCs to recognise or internalize the Ph1 strain in comparison toH37Rv. Additionally, judging from the induction of antigen presenting molecules(e.g. HLA-DR, CD86) and cell death pathways (e.g. apoptosis, necrosis), we can ruleout a problem with the ability of the Ph1 strain to activate DCs. Rather, we believethat the diminished capacity of Ph1-infected DCs to activate autologous T cells maybe due to their inability to colonise these APCs. Based on our CFU assays, wedetermined that Ph1 failed to achieve a significant level of intracellular growth inhuman DCs. Therefore, we can infer that the Ph1-infected DCs may have less antigenicmaterial available to present and activate naïve T cells. This is in contrast to theresults obtained in the study by Marquina-Castillo et al.12 conducted in a mouse model. In their study, the Ph1 strain grew rapidly inthe lung and peaked at 21 dpi, exhibiting a bacterial burden (via CFU readings)twice as high in comparison to those mice infected with the H37Rv strain.12 We infer this phenotype is likely due to the capacity of the Ph1 strain tocolonise and grow within other leukocytes besides DCs, because, compared to H37Rv,the Ph1 strain displays similar in vitro phenotypes in terms ofcord formation, growth curves, and response to hydrogen peroxide exposure.12 However, the authors did not examine the type of murine leukocytes servingas reservoirs for the bacteria. As previously reported, DCs exhibit anon-permissible phenotype against Mtb intracellular growth,6 and this characteristic is not influenced by the degree of virulence as weare showing in this study. Therefore, our results suggest that the Ph1 strain mayhave a better capacity to colonise macrophages in contrast to DCs.

Nevertheless, our observation that Ph1-infected DCs do not optimally activateautologous T cells is in line with the main finding in the study conducted byMarquina-Castillo et al.,12 which concluded that the Ph1 strain does not induce a protective immuneresponse in a mouse model, such as the one promoted by the H37Rv strain. Indeed,infection of mice with the Ph1 strain was characterised by a delayed IFNγexpression, which is indicative of poor activation of Th1 cells.12 Infection of mice with other hypervirulent Mtb strains, such as thosebelonging to the W-Beijing lineage, also results in a poor protective Th1-drivenimmune response distinguished by the low and temporal expression of IFNγ, TNFα, andiNOS.15^,16 Suboptimal antigen presentation has been shown to contribute to thevirulence of Mtb in an in vivo mouse model.17 Moreover, DCs generated from monocytes after Type I IFN exposure (IFN-DCs)highly resemble naturally occurring DCs induced in vivo, forexample, in a chronic infection context.18 In active TB patients, IFN-DCs showed a diminished capacity to induceAg-specific T-cell responses against Mtb.19 Collectively, our results argue that the interaction of the Ph1hypervirulent strain with human DCs differs from that of H37Rv, resulting in a lowercapacity for these antigen-presenting cells to optimally activate an adaptive immuneresponse.

The second key contribution of our study is the demonstration that the Ph4hypovirulent strain induces apoptosis in human MDDCs, preventing them fromsupporting an autologous co-culture system with T lymphocytes. Similar to the cellsinfected with H37Rv, MDDCs were able to recognise, internalise, and control theintracellular growth of the Ph4 strain. However, MDDCs infected with thishypovirulent strain underwent rapid and massive levels of apoptosis, as compared tothe other Mtb strains. These results suggest that less virulent Mtb strains caninduce apoptosis in DCs, as well as macrophages, and in this manner, become a sourceof antigen for other antigen-presenting cells to pick up and promote an efficientadaptive immune response, as postulated previously.7^,20 Interestingly, while the Ph4 strain did not differ in its capacity toactivate MDDCs in terms of the up-regulation of cell surface receptors involved inantigen presentation (albeit always displaying the highest tendency among the testedstrains), MDDCs infected with this strain were not able to support the establishedautologous co-culture system with CD4⁺ T cells. In the context of TB,infected migratory DCs are poor antigen-presenting cells,21 which suggests that infected cells have a mechanism to transfer antigen touninfected cells so that uninfected cells could prime CD4⁺ T cells.In vivo studies demonstrated that migratory DCs mustcollaborate with one or more resident DCs to successfully prime CD4⁺ Tcells.22^,23 Cultured DCs and macrophages release multiple Mtb protein antigens into theextracellular medium that can be taken up, processed, and presented toCD4⁺ T cells by uninfected DCs.23 This form of antigen transfer occurs without transfer of the pathogenitself, providing a mechanism for host cells to bypass the inhibitory effects of Mtbon antigen presentation and allow for effective priming of antigen-specificCD4⁺ T cells. We inferred that MDDCs infected with the Ph4 strainbecome the source of mycobacterial antigens for uninfected bystanderantigen-presenting cells, which may consequently activate an efficient T-cellresponse. Moreover, we believe that these results are in line with thecharacterisation performed in the mouse model by Marquina-Castillo andcolleagues.12 Indeed, the authors determined that infection of mice with the Ph4 strainculminated in an increased survival rate of the mice in comparison to infection withthe H37Rv strain. This was also correlated with a better protective acquired immuneresponse characterised by high and sustained (albeit delayed) IFNγ expression, andearly high levels of TNFα expression.12

It remains to be determined what signalling pathways promote the rapid and highlevels of apoptosis triggered by the Ph4 strain in MDDCs. In the context of murine:bone marrow-derived-dendritic cells (BMDC) for example it was demonstrated that Mtbinduces caspase-1/11-independent apoptosis but not necrosis.7 In the context of macrophages, avirulent Mtb strains induce strongexpression of prostaglandin-endoperoxide synthase 2 (PTGS2), which consequentlyleads to high rates of apoptosis-promoting prostaglandin E2 (PGE₂)production.20 Lipoxin A4 (LXA4) is known to inhibit PTGS2 and PGE₂, leading toan increase in necrosis instead of apoptosis, which is common in macrophagesinfected by virulent Mtb strains.5 Altogether, our results argue that the interaction of the Ph4 hypovirulentstrain with human DCs greatly differs from that of the H37Rv strain, resulting inrapid and massive rates of apoptosis that may facilitate the acquisition ofantigenic material, thus leading hypothetically towards an efficient activation ofthe adaptive immune response.

In conclusion, this study represents the first prospective assessment of the humanleukocyte response to the Mtb strains endemic in Southern Mexico that vary invirulence and transmission as previously assessed in mouse models. In general, ourfindings in human DCs suggest that the variability in virulence among these Mtbstrains affects the capacity of leukocytes to respond to pathogenic challenge andmount an immune response against it, highlighting important parallels from studiesperformed in mouse models. The association of the observed effects in this study andspecific virulence factors previously described for Mtb warrant furtherinvestigation. We support the notion that our Mtb clinical isolate strains make goodcandidates for further investigation using genome sequencing, transcriptomehybridisation, and comparative proteomics, probably leading to the eventualidentification of Mtb genes associated with virulence and the interference of humanDC biological functions, as previously proposed.13

ACKNOWLEDGEMENTS

To Dr Hudrisier for his help and sharing of reagents and equipment for theantigen-presentation assays.

Footnotes

REFERENCES