|Home | About | Journals | Submit | Contact Us | Français|
Leprosy is a chronic infectious disease and is a major source of morbidity in developing countries. Leprosy is caused by the obligate intracellular bacterium Mycobacterium leprae, which infects as primary target Schwann cells. Lepromatous leprosy exhibits multiple lesions of the skin, eyes, nerves, and lymph nodes. The sites of infection are characterized by the presence of foamy macrophages, fully packed with lipid droplets (LDs), which are induced by M. leprae. In the last years, it has become evident that M. tuberculosis imports lipids from foamy macrophages and is dependent on fatty acids for growth in infected macrophages. M. leprae seems to have similar mechanisms for scavenging lipids from the host. But due to the inability to culture M. leprae on laboratory media, research progresses only slowly. However, in the last years, substantial progress has been made in the field of lipid metabolism in M. leprae. Herein, we will present and summarize the lipid droplets formation and the metabolism of lipids during M. leprae infection.
Leprosy, a major source of morbidity in developing countries, is a chronic infectious disease caused by the obligate intracellular bacterium Mycobacterium leprae [1, 2]. According to the system of classification of Ridley and Jopling (1966), leprosy patients show two major manifestations of the disease, designated as lepromatous leprosy (LL) and tuberculoid leprosy (TT) . TT is observed in patients with good T-cell mediated (Th1) immunity and is characterized by granuloma formation and death of Schwann cells (SCs) leading to loss of myelin sheath and nerve destruction [3, 4]. TT shows onlyfewlesions, and bacillicanrarelybeidentified. Patients with poor T-cell mediated immunity show the lepromatous-type leprosy (LL). LL leads to massive bacterial load inside host cells specially SCs and macrophages [3, 5–7]. The lesions of TT and LL types are named as T-lep and L-lep lesions, respectively, but damage of the nerves is observed in most of the cases of both types .
Lepromatous leprosy exhibits multiple lesions of the skin, eyes, nerves, and lymph nodes, which are characterized by tumor-like accumulations of foamy macrophages. The foamy macrophages are fully packed with lipid droplets (LDs) and contain high numbers of lepra bacilli. These aggregations of foamy macrophages expand slowly and disfigure the body of the host . Foamy macrophages were described first by the pathologist Rudolph Virchow  and despite these lipid-laden cells are a hallmark of lepromatous leprosy, only recently a few researchers started to elucidate the molecular biology of these cells. In the last ten years, it has become evident that M. tuberculosis induces the formation of foamy macrophages, a process which appears to be a key event in both sustaining persistent bacteria and release of infectious bacilli . Moreover host lipids from LDs are regarded as substantial nutrient source for mycobacteria during infection. Middlebrook already demonstrated in the late 1940s that mycobacterial growth in vitro was enhanced by supplementation with oleic acid . Wheeler et al. reported insufficient fatty acid synthase activity to support growth in M. leprae. This finding supported the hypothesis that M. leprae scavenges lipids from the host cell . Over the last years, it has become evident that survival and persistence of M. tuberculosis is critically dependent on lipid body formation. Furthermore, lipid body formation seems to be the prerequisite for transition of M. tuberculosis to the dormant state. This goes along with the important observation that sputum from tuberculosis patients contains lipid body-laden bacilli [13, 14].
For a great number of mycobacterial species, intracellular lipid bodies were reported [15–25]. In the dormant state, lipids from lipid bodies appear to be the primary carbon source for M. tuberculosis in vivo. For M. tuberculosis, several bacterial genes are upregulated during the dormant state and have been reported to be involved in lipid metabolism such as diacylglycerol acyltransferase (tgs1), lipase (lipY), and isocitrate lyase (icl) [26, 27].
M. leprae has a small genome (3.2 Mb). The obligate intracellular organism shows a moderate genome degradation and several genes are absent when compared with other mycobacterial species. Due to the gene loss, M. leprae is strongly dependent on thehost for basic metabolic functions [3, 28]. Macrophages infected with M. leprae contain oxidized host lipids and it has been observed that M. leprae upregulates 13 host lipidmetabolism genes in T-leplesions and 26 in L-leplesions. The oxidized lipids inhibit innate immune responses and thus seem to be an important virulence factor for the organism .
This paper highlights the importance of the LDs as one of the most unique determinant for M. leprae persistence and virulence. The formation of LDs in M. leprae infected cells has been compared to LDs formation in macrophages infected with M. tuberculosis. Thereafter, we will focus on the interplay between the LDs of the host and the pathogen.
M. leprae infects preferentially macrophages and SCs . A typical feature of lepromatous leprosy is the survival and replication of M. leprae stored within the LDs in the enlarged phagosome of histiocytes. Due to strong immune response against M. leprae and arrested bacilli growth in T-lep lesions, the occurrence of foamy cells of M. leprae-infected macrophages/SCs is a marker only in L-lep lesions [5, 9]. In L-lep lesions, the expression of lipoprotein and fatty acid metabolism, including lipases genes, was 4-fold greater than in T-lep lesions which supports the previous observation . Lipid droplets are thought to be an important nutrient source for the bacillus. A major concern in leprosy is peripheral neuropathy. The damage to nerves of the peripheral nervous system is caused by the infection of SCs by M. leprae. In LL nerve biopsies, highly infected SCs also contain LDs and show a foamy appearance, such as Virchow cells found in dermal lesions .
The biology of these foamy cells has been characterized poorly until now. Neither the origin nor nature of the lipids has been elucidated yet. Only recently, in vitro studies by Mattos could show that M. leprae induces the formation of LDs in human SCs . Moreover, the group found that LDs are promptly recruited to bacterial phagosomes. In SCs, LDs recruiting by bacterial phagosomes depends on cytoskeletal reorganization and PI3K signaling but is independent of TLR2 bacterial sensing .
Important markers for the lipid accumulation in adipocytes or macrophages are lipid-droplet-associated proteins such as adipose differentiation-related protein ADRP and perilipin, which play essential roles in lipid-droplet formation . After phagocytosis of live M. leprae, ADRP expression is constantly upregulated in human monocytes. ADRP and perilipin are localized at the phagosomal membrane (Figure 1) .
Mycobacterium bovis Bacillus Calmette-Guérin (BCG) and M. leprae are recognised by the Toll-like receptors (TLR) TLR6 and TLR2 [32, 37] and induces the formation of foamy macrophages . For BCG it has been shown that surface molecule lipoarabinomannan (LAM) binds to TLR2. .
M. leprae association to macrophages is mediated by binding of the bacteria to TLR2 and TLR6. Heterodimerization of TLR2 and TLR6 leads to downstream signalling and subsequent LDs formation (Figure 1) [32, 33]. Macrophage association is not dependent on binding to TLR2 or TLR6. Neither a TLR2−/− nor TLR6−/− knockout macrophage shows reduced binding to M. leprae. This suggests that both TLR2 and TLR6 can bind M. leprae alone, or/and the presence of other receptors, binding to M. leprae. The TLR2−/− or TLR6−/− knockout macrophages do not also completely abolish LD formation, but show only reduced LDs formation . This suggests the presence of additional signalling pathways for LDs formation. In contrast TLR6, but not TLR2, is essential for M. leprae-induced LDs biogenesis in SCs .
At least one additional factor is required for general phagocytosis, in mycobacterial infection. Members of the CD36 family have to be shown to be required for uptake of mycobacteria . Macrophages infected with M. tuberculosis show an increased surface expression of the type 1 scavenger receptors CD36 and LOX1, which facilitate the uptake of oxidized host macromolecules including OxLDL (Figure 1) .
Macrophages generate and release reactive oxygen species (ROS) during activation of the respiratory burst upon infection with pathogenic bacteria. Oxidative stress results not only in damage to cellular structures but also to oxidation of fatty acids, such as low density lipoproteins (OxLDL) in granulomas. OxLDL binding to LOX-1 also increases reactive oxygen species (ROS) formation [41, 43]. The binding of OxLDL to type 1 scavenger receptors CD36 and LOX1 induces increased surface expression of both receptors, leading to uptake of OxLDL [41, 43, 44]. In addition, CD36 increases the uptake of M. tuberculosis by macrophages . The increased rate of OxLDL uptake results in the accumulation of oxidized lipids, which finally leads to the formation of foamy macrophages (Figure 1) . M. tuberculosis and M. leprae benefit from the accumulated OxLDL in the infected macrophage. OxLDL-laden lung macrophages show enhanced replication of intracellular M. tuberculosis compared to macrophages loaded with nonoxidized LDL . The presence of oxidized phospholipids in M. leprae infected macrophages downregulates the innate immune response and contributes to pathogenesis . Moreover, scavenger receptor-deficient phagocytes are characterized by a reduced intracellular bacterial survival and a lower cytokine response .
In 1863, Virchow described lipid-laden macrophages. The lipids in these foamy cells form droplets and surround also M. leprae within the phagososomes [46, 47]. This lipid layer or capsule forms a characteristic electron-transparent zone. In contrast to M. tuberculosis, lipid inclusions (lipid bodies) seem to be rather exceptional in M. leprae . The electron-transparent zone contains mycoserosoic acids of phthiocerol dimycocerosates as well as phenolic glycolipids [48, 49]. Brennan reported the full characterization of three phenol-phthiocerol triglycosides by M. leprae . It has been postulated that many of these same molecules together with phosphatidylinositol mannosides and phospholipids are released from the cell wall after synthesis, forming the capsule-like region . The dominant lipid in the cell wall which gives M. leprae immunological specificity is phenolic glycolipid-1 (PGL-1). Phenolic glycolipid 1 has been isolated in relatively high concentrations from purified bacteria and from M. leprae infected tissues . PGL-1 is thought to be a major component of the capsule in M. leprae and constitutes an important interface between bacteria and host . It has been suggested that PGL-1 is involved in the interaction of M. leprae with the laminin of SCs, thus PGL-1 might play a role in peripheral nerve-bacillus interactions . Moreover, phenolic glycolipids seem to be involved in the stimulation of suppressor T-cells in lepromatous leprosy . Daniel et al. demonstrated that M. tuberculosis inside foamy macrophages imports fatty acids derived from host TAG and incorporates them intact into bacterial TAG, which is accumulated in lipid bodies .
Recently, it was reported that also LDs from M. leprae infected SCs and macrophages accumulate mainly host-derived lipids, such as oxidized phospholipids . BODIPY stains infected SCs, indicating that LDs contain neutral lipids, such as triacylglycerols (TAG) .
In M. leprae infected-cells, LDs show high ability to fuse further to form giant LDs. In most mammalian cells, dynein and extracellular signal-regulated kinase-2 (ERK2) facilitate LDs move along the cytoskeletal network . Extracellular signal-regulated kinase-1/2 is one of the most upregulated signals after M. leprae induced-SCs demyelination and nerve injury . Using NIH 3T3 cells, Andersson et al. demonstrated that ERK2 is an essential signal in LDs formation  and also seem to be crucial for the phospholipase D1- (PLD1-) mediated increases of LDs. The ERK2 mechanism is mediated by phosphorylation of dynein, which finally upregulates its association in ADRP-containing LDs .
Lipid droplets induction in all type of cells associates with lipid droplets protein adipophilin (ADFP) . The classic model of LDs postulates that LDs formation take place in the outer leaflet of the endoplasmic reticulum (ER) membrane (the place of TAG synthesis). In this formation, the function of AFDP has been proposed in fatty acid absorption and transport. Cruz et al. showed that ADFP is overexpressed in L-lep lesions . Furthermore, the authors also pinpointed similarity in overall observations to one of the important metabolic disease atherosclerosis. Similar observation in atherosclerosis showed that ADFP is the most expressed LD-associated protein found in macrophages and that the accumulation of LDs in lesional macrophages is associated with ADFP aggregation .
Several reports have supported the inflammatory role of macrophages and SCs, where LDs induction by these cells associated with modulation in production of inflammatory mediators [59–63]. Moreover, increased LDs formation accompanied by PGE2 and transforming growth factor-β (TGF-β1) synthesis in BCG infected macrophage while strong PGE2 was seen in foamy macrophages during M. tuberculosis infection [64, 65]. Cytokines, chemokines, and growth factors seem to be correlated with LDs formation by M. leprae. For instance, Persson et al. found that IL-1β and TNF-α inflate foamy macrophages formation . M. leprae infection significantly increased NF-κB, TNF-α, and IL-1β expression in NOD1- and NOD2-transfected macrophages cells . Another example, IL-10 and IL-12 secretion correlated with bacilli-LD induction in L-lep lesions [32, 68]. Schwann cells are able to produce several type of cytokines, including IL-1, IL-6, IL-8, IL-10, IL-12, PGE2, TGF-β, and TNF-α [37, 69, 70]. PGE2 and IL-10 overexpression is associated with clog of IL-12 and NO production in M. leprae-infected SCs [32, 68]. In late infection state, M. leprae impede the expression of TNF signaling and induce expression of 9-O-acetyl GD3 as smart strategy to avoid the cellular apoptosis and facilitate its persistence inside the favored cell niche [37, 67, 71, 72]. Obviously, lipid biosynthesis and cytokine production as the whole immune response to M. leprae infection are well connected in SCs . Innate immune response in infected SCs seems to be dependent on LDs accumulation and TLR2/TLR6 signaling. Using a nonsteroidal anti-inflammatory drug and fatty acids synthesis inhibitors, both LDs formation and innate immune responses completely abrogated, supporting the role of bacilli-LDs induction and immune responses [32, 37]. Another modulation observed is downregulation of Th1 responses and bactericidal activity through overproduced PGE2 where TLR2 and TLR6 heterodimerization plays important role . In general, cytokine profile in both lesions well linked with TLRs function. M. leprae modified-TLRs function stimulates monocytes and macrophages in TT and LL disease type, but in L-lep lesions the adaptive response was debilitated more than what was seen in T-lep lesions . Despite the strong evidences that LDs induction is associated with modulation of cytokines, it is not known whether cytokines signaling functions within LDs themselves.
Due to the genome reduction M. leprae is strongly dependent on catalytic activities provided by the host [3, 28]. Macrophages infected with M. leprae contain oxidized host lipids and it has been observed that M. leprae upregulates 13 host lipidmetabolism genes in T-leplesions and 26 in L-leplesions . But still it is not clear what carbon source M. leprae uses under aerobic or anaerobic conditions. The genome contains genes for utilization of sugars, such as glycolysis, tricarboxylic acid cycle (TCA), and monophosphate shunt (HMP) pathways. The M. leprae genome contains all genes for a functional glyoxylate cycle, which would enable the organism to use acetyl-CoA from the β-oxidation of lipids. M. tuberculosis derives much of its energy from the degradation of host-derived lipids [73, 74]. Neutral lipids are hydrolyzed by lipases or esterases, yielding fatty acids for energy generation and anabolism of membrane phospholipids. The genome of M. tuberculosis H37Rv contains twenty genes designated as putative lipases (lipA to W, except K and S) [74, 75]. In the M. leprae genome, only 2 lipase genes (lipG, lipU) were found, suggesting a gene loss in comparison to M. tuberculosis. But recently it has been found that in M. tuberculosis genome only six Lip expressed-enzymes showed reasonable hydrolase activity for long-chain triacylglycerols (LipY, LipC, LipL, LipX, LipK, and LipG). LipY (Rv3097c) has the highest activity, compared to all Lip enzymes. LipG and LipU from M. leprae are homologous with LipG and LipU from M. tuberculosis and show sequence identities of 72% and 79%, respectively. The lipases LipG and LipU from M. tuberculosis show very low and no activity with long-chain triacylglycerols as substrates . M. tuberculosis LipY is suspected to be a major functional lipase, which utilizes stored triacylglycerols (TAG) during dormancy and reactivation of the pathogen [76, 77]. LipY shows only a weak similarity with M. leprae LipU (23% identity). In summary, it appears that M. leprae uses different lipases for the hydrolysis of fatty acids than M. tuberculosis. All M. tuberculosis and their M. leprae homologs enzymes involved in lipid metabolism in lipid bodies are summarized in Table 1.
M. tuberculosis can grow on fatty acids as sole carbon source and it has been demonstrated that fatty acid oxidation is important for survival of the pathogen in the lungs of mice [83, 84]. Fatty acids are oxidized via the β-oxidation cycle and the glyoxylate shunt, to replenish TCA cycle intermediated during growth . The β-oxidation cycle consists of five biochemical reactions, where one molecule acetyl-CoA of the fatty acid is split off per cycle. The genome of M. tuberculosis encodes around 100 genes, designated as fad genes (fatty acid degradation) with putative roles in the β-oxidation of fatty acids. While E. coli has only one enzyme for each step of the β-oxidation cycle, M. tuberculosis seems to have several backup enzymes for each reaction . M. leprae has approximately one-third as many potential fad enzymes. However, there are three times more FadD acyl-CoA synthases than there are FadE acyl-CoA dehydrogenases, whereas these are predicted in equal amounts in M. tuberculosis . The initial step of β-oxidation is the formation of acyl-CoA from free fatty acids and coenzyme A and is catalyzed by acyl-CoA synthase. The Acyl-CoA synthase Rv1683 is suspected to be essential for TAG hydrolysis and growth [80, 90].
Even though M. leprae genome contains less necessary β-oxidation cycle genes than M. tuberculosis, transcripts analysis revealed active expression of acyl-CoA metabolic enzymes including echA1 (ML0120, putative enoyl-CoA hydratase), echA12 (ML1241, possible enoyl-CoA hydratase), fadA2 (ML2564, acetyl-CoA-acetyltransferase), fadB2 (ML2461, 3-hydroxyacyl-CoA dehydrogenase), fadD19 (ML0352, acyl-CoA synthase), fadD26 (ML2358, fatty acid-CoA-ligase), fadD29 (ML0132, probable fatty-acid-CoA synthetase), fadD28 (ML0138, possible fatty-acid-CoA synthase), fadE25 (ML0737, probable acyl-CoA dehydrogenase), and fadE5 (ML2563, acyl-CoA dehydrogenase) [91, 92]. Combing observations from leprosy lesions, this gives strong evidence that host lipids provide the main carbon and energy sources for M. leprae during infection.
Together with malate synthase, isocitrate lyase (ICL) is the key enzyme of the glyoxylate cycle that catalyzes the cleavage of isocitrate to glyoxylate and succinate [88, 93]. The M. leprae genome contains a gene, coding for an isocitrate lyase, aceA. AceA is upregulated in both M. leprae-infected nude mouse and human lesions. . The amino acid sequence of aceA (ML1985c) shows 80% identity with its homologue from M. tuberculosis ICL2 (Rv1915/1916). A second icl gene is not present in M. leprae, as observed in M. tuberculosis. This finding is of particular interest, because two M. tuberculosis isocitrate lyases, icl and icl2, are jointly required for in vivo growth and virulence [83, 84]. Deletion of icl1 or icl2 has little effect on bacterial growth in macrophages . So far the evidence indicates that M. leprae aceA might play a slightly different role in both isocitrate lyases in M. tuberculosis.
For persistent M. tuberculosis, TAGs seem to be the major carbon and energy source. Biosynthesis of TAG consists of the sequential esterification of the glycerol moiety with fatty acyl-residues by various acyltransferases. The fatty acids are supplied by de novo biosynthesis or β-oxidation. Esterification occurs via sequential acylation of the sn-1,2 and 3 positions of glycerol-3-phosphate, and removal of the phosphate group before the last acylation step. The terminal reaction is the esterification of diacylglycerol (DAG) with acyl-CoA by an diacylglycerol acyltransferase (Figure 2) . Bacteria do not contain DGATs but only bifunctional wax ester synthase/acyl-CoA: diacylglycerol acyltransferases (WS/DGAT). WS/DGATs mediate next to TAG formation the synthesis of waxes by esterification of acyl-CoA with alcohol . The genome of M. tuberculosis codes for 15 genes which contain the highly conserved putative active site motif of WS/DGATs (HHxxxDG). These genes were designated as “tgs,” triacylglycerol synthases, but have only a weak sequence similarity to other WS/DGAT sequences. All 15 expressed mycobacterial tgs proteins show diacylglycerol acyltransferase activity and tgs1 has the highest activity of all enzymes . Tgs1 appears to be a major contributor to TAG synthesis in M. tuberculosis so far [26, 78]. And moreover, two homologue proteins to tgs1 and tgs2 (BCG3153c and BCG3794c) and another poorly characterized acyltransferase (BCG1489c) were found to be exclusively associated to lipid bodies. Disruption of the tgs genes BCG3153c, BCG3794c, and BCG1489c reduces TAG accumulation during hypoxia-induced nonreplicating state, revealing that the enzymes are involved in TAG synthesis during latency and pathogenicity .
Ten of the 15 tgs genes in M. tuberculosis are located adjacent or proximal to 11 lip genes that are annotated as probable phospholipases or lipases-esterases carboxylesterases. Some tgs genes may be cotranscribed with neighboring lip genes and may synthesize triacylglycerols from the released fatty acids from the host . Lip gene products may be important for utilization of TAGs during dormancy and upon reactivation after dormancy. The tgs gene Rv0221 is located near LipC (Rv0220), lipW (Rv0217c), acyl-CoA synthetase (Rv0214), acyl-CoA dehydrogenase (Rv0215c), and an integral membrane acyltransferase (Rv0228). This clustering of genes of the fatty acid metabolism suggests that these genes may be cotranscribed and may release fatty acid from host TAG, carry out the transport of fatty acids and finally catalyze the resynthesis of TAGs in the pathogen. Rv0221 and LipC have to be shown to be catalytical active [26, 76].
The M. leprae genome shows only one predicted gene product which has a significant degree of identity to any tgs enzymes from M. tuberculosis . The tgs gene product ML1244 shows 72% identity to Rv2484c from M. tuberculosis. Rv2484c is located next to a carboxylesterase lipQ, (Rv2485c), a probable glycerol-3-phosphate acyltransferase, (Rv2482c), a lysophosphatidic acid acyltransferase-like protein (Rv2483c), and a probable enoyl-CoA hydratase (Rv2486). The gene cluster of lipid metabolism genes suggests a possible involvement of the gene products in the synthesis of TAG . A few tgs genes (Rv3234c, Rv3233c, Rv2285, and Rv1425) are located proximally to lipoproteins, which may serve as donors or acceptors of fatty acids .
Ag85A, a mycoly transferase that is known to catalyze the formation of the cord factor was recently found to have additional DGAT activity . The homologue gene of M. tuberculosis Ag85A is expressed in M. leprae. Although, mycoly transferase 85 complex genes (A, B, and C) transcripts were one of the major secreted group during mycobacterium infection of the mouse, Ag85A is one of the most genes shown to be relatively highly expressed either in infected nude mouse or human skin lesions .
Macrophages, infected with M. leprae, alter the lipid metabolism. Especially an increased accumulation of cholesterol and cholesterol esters has been found . The increased cholesterol level was also associated with a lower level of esterase activity in infected macrophages .
Mattos et al. recently showed by high-performance thin-layer chromatography that the amount of cholesterol increases in SCs upon infection with M. leprae . Cholesterol utilization was also identified to be required for mycobacterial persistence . In 2008, Pandey and Sassetti found that M. tuberculosis can grow using cholesterol as a primary carbon source and that the mce4 transporter is required for cholesterol uptake.
While the M. tuberculosis genome contains four homologous mce operons, mce1-mce4, which are thought to encode lipid transporters [96, 97], M. leprae genome encodes only one operon (mce1). All M. leprae five mce genes were overexpressed during intracellular growth in mouse and human biopsies [91, 92]. This observation suggests that the intracellular bacilli population induces cholesterol uptake of the infected cell and subsequently uses the stored cholesterol as carbon and energy source.
Cholesterol is also essential for survival of mycobacteria in macrophages. Cholesterol accumulates at the site of mycobacterial entry in macrophages and promotes mycobacterial uptake. For M. tuberculosis and M. leprae it has been shown that cholesterol mediates the recruitment of TACO from the plasma membrane to the phagosome . TACO, also termed as coronin-1A (CORO1A), is a coat protein that prevents phagosome-lysosome fusion and thus degradation of mycobacteria in lysosomes [34, 35]. The entering of mycobacteria at cholesterol-rich domains of the plasma membrane and their subsequent uptake in TACO-coated phagosomes promotes intracellular survival [35, 36].
Studying M. leprae metabolisms during infection is as complex as the choice of invasion mechanisms traffic and as their multifarious factors for entry are. In this scenario, most of publications emphasize that leprosy maintains metabolic remodeling between both host and pathogen. A hallmark of intracellular infection is the formation of foamy host cells. Several results support the view that M. leprae actively catabolizes fatty acids for energy and produces a wide array of secretory proteins. Huge gaps remain in our knowledge of the structure and function of M. leprae lipid bodies, as the host metabolism required alterations to switch the organism to the utilization of lipids as alternate carbon sources. Investigating the metabolic aspects of M. leprae especially LDs formation is an important topic to learn more about metabolic pathways that could potentially be exploited for developing novel drugs against leprosy and tuberculosis.
This work was supported by NOPERSIST (EC Grant Agreement no. 232188) grant from European Commission, Seventh Framework Programme (FP7) to M. Singh.