Synopsis of Lipid Droplet Formation

Because Tuberculosis (TB), caused by Mycobacterium tuberculosis, can manifest in a variety of ways that include disease of the bone, central nervous system, and other organ systems the analysis of ancient bones looking for TB has been possible. TB has been identified as the causative agent for death in individuals who died more than 4000 years ago [1]. M tuberculosis continues to be a global problem and causes over one million deaths per year [2]. In addition, TB is one of the top 10 causes of death throughout the world and disproportionately causes more deaths in lower income areas of the world.

TB is primarily a pulmonary disease that begins with a period of latency that can last for years as dormant TB infection (not yet disease) in which M. Tuberculosis lives in the lungs without making the person sick. When M. tuberculosis then TB disease occurs this is most often due to a weakened immune system of the host either due to illness, malnutrition, HIV, or other immune compromising conditions. A hallmark of TB infection is the formation of granulomas in the lungs and it is in these that the bacterium is able to live. These granulomas eventually cause enough lung damage that the host dies from suffocation from lack of oxygen. The presence of foamy macrophages in the granulomas, along with the caseous central necrosis provide a lipid-rich environment for the growth of M. tuberculosis [4]. As the genome of M. tuberculosis was researched it was not a surprise to find that a large portion is involved in lipid metabolism as a carbon source for growth [5]. It has been hypothesized that M. tuberculosis that the lipid droplets (LDs) in macrophages are a carbon source but there has been no genetic evidence that shows that host LDs are used by M. tuberculosis. LDs are not just storage facilities. LDs are important in the immune response that include the production of pro-inflammatory eicosanoids [6,7].

This paper reports on research that demonstrated LD formation is a host response and that this response is driven by cytokine IFN-{gamma}and people who are actin CD4+ T cells are very likely to develop TB disease with an example being patients with HIV. Cytokine IFN-{gamma} is needed for the control of M. tuberculosis infection but the manner in which IFN-{gamma} may effect the macrophage metabolism during a TB infection and whether the IFN-{gamma} activation of macrophages has any effect on LD formation or function had not been studied. This interesting paper reports on the researchers demonstration that LD formation and function is a programmed host response that is driven by IFN-{gamma}. The transcription factor for IFN-{gamma} dependent LD formation is identified along with the transcriptional target for the major regulator of LDs formed when macrophages are infected with M. tuberculosis and that this pathway is present in vivo. Finally, this paper also reports on the research that once IFN- {gamma} activation of macrophages occurs then M. Tuberculosis's access to lipids in the macrophages is limited. This research suggests that the paradigm that the lipid droplets in macrophages are a carbon source for M. tuberculosis is unlikely.

The first experiment was designed to learn about the regulation and role of LD formation during M. tuberculosis infection. Primary murine bone marrow derived macrophages (BMDM) that were both resting and had activated INF-{gamma} were infected with fluorescent M tuberculosis. IN Infected esting BMDM showed very few LD accumulation while the BMDM with IFN-{gamma}had nears 100% of the BMDM containing LDs by as early as day 1. In addition, if IFN-{gamma} activation was done with non-infected BMDMs, there was not LD formation. This suggested that it is a synergistic response between the infection with M tuberculosis and the IFN-{gamma} activation that produced the LD formation as an adaptive immune response to the infection. LD formation in human monocyte derived macrophages was also assessed and also showed that the combination of IFN-{gamma} activation and M tuberculosis infection had a large accumulation of LD in the human macrophages.

The next question addressed was whether the acquisition of host lipids by M tuberculosis

correlated with the macrophage production of LD. In order to test this, infected BMDM were pulsed with fluorescent labeled fatty acid that was added to the media. At 2 and 8 hrs, the labeled fatty acid had accumulated into the M tuberculosis and this occurred when there were not macrophage LDs present. Contrary to this, the labeled fatty acid had minimal accumulation inside the M tuberculosis that were in IFN-{gamma} activated BMDM in which host LDs were seen. Together this showed that without the presence of macrophage LDs the M tuberculosis is able to accumulate LDs but when the IFN-{gamma} has induced the macrophage LDs, M tuberculosis has markedly less LD accumulation. Macrophage production of LDs does not correlate with M tuberculosis acquisition of the host lipids.

Since the accumulation of macrophage LDs was observed to increase during M tuberculosis infect via IFN-{gamma} activation, the next question addressed was what changes are happening in lipid metabolism. Using mass spectrometry it was shown that the levels of most phospholipids and sphingolipids were not changed in BMDM infected with M tuberculosis, or following IFN-{gamma} activation, or in BMDM that were both activated with IFN-{gamma} and infected. It is interesting to note that the total TAG levels were slightly higher in the BMDM activated with IFN-{gamma} and infected with M tuberculosis when compared non activated BMDM that were infected (FIG 3A and 3B). This suggested the elevation of TAG levels alone was not able to induce LD formation in the macrophages.

In order to determine the source of the lipid increase in M tuberculosis infection, RNA-seq data was used to demonstrate that Fasn (the primary fatty acid synthase) was down regulated in BMDM M tuberculosis infection with IFN-{gamma}activation (8). Since fatty acid synthesis had now been shown that it was not what was driving LD accumulation in macrophages in M tuberculosis infection the authors then moved to whether the activity of diacylglycerol acyltransferase (DGAT) activity of catalyzing of the final step in TAG synthesis would affecting TAG synthesis using DGAT inhibitors and what affect that would have on LD formation (9). Treatment of M tuberculosis infected, IFN-{gamma} activated BMDM with the inhibitors prevented LD formation and this continued for 3 days after the infection. These results suggested that TAG synthesis, but not FA synthesis contributes to the formation of LDs in IFN-{gamma} activated and infected macrophages. Since CD36, (also known as fatty acid translocase), is the receptor for LDL and VLDL and long chain fatty acids, the last part of this section of work was used to determine if CD36 was needed to maintain the LDs (10). Using CD36 knock out mice it was shown that the Cd36 knock out BMDM had less LDs 3 days after infection. Taken all together, these data showed that LD requires TAG synthesis, and that maintenance of the LDs needed CD36 to import lipids.

The researchers also used a series of experiments to examine the transcription factor Hypoxia-inducible factor-1 (HIF-1{alpha}) which has been called the "master regulator of the hypoxia response."(11). During hypoxia HIF-1{alpha} mediates LD formation and the the authors had previously shown that is it needed for controlling M tuberculosis infection both in vitro and in vivo and that its activation is IFN-{gamma} dependent (12). Using BMDM with absent HIF-1{alpha}, the authors showed that the macrophages had a lower amount of LDs and that the LDs were smaller. Next lipidomic analysis of the HIF-1{alpha} absent BMDMs showed a small deficit in cholesterol ester and a significant defect in TAG accumulations with these two things suggesting that there was not sufficient accumulation of lipids to make LDs.

Using knowledge from their previous research (12), the next question addressed was which HIF-1{alpha} genes mediate LD formation during M tuberculosis infection. Using infected BMDM with IFN-{gamma} activation, the researchers showed that Hypoxia Iducible Gene-2 (Hig2) was induced in this scenario and that the upregulation was HIF-1{alpha} dependent. Since Hig2 was found to be the HIF-1{alpha} target the authors then hypothesized that this upregulation contributed to LD accumulation. In an elegant design to test this, Hig2 was deleted in BMDM using CRISPR/Cas9 techniques. BMDM with Hig2 knockout were found to have almost a 90% defect in the number of LDs (FIG 5F) and an smaller size of LDs as compared to wild type BMDMs (FIG 5G). Both of these results were identical to BMDMs with a Hif1{alpha} defect. These results demonstrated that the HIF-1{alpha} target gen Hig2 is needed for LD maintenance in M tuberculosis infection.

The final series of questions addressed in this research looked at how and if LDs support host immunity and/or are a source of nutrients for the growth of M tuberculosis in vivo. The authors had previously shown that HIF-1{alpha} is needed for prostaglandin-E-2 (PGE2) production in M tuberculosis infected IFN-{gamma} activated BMDMs (12). In this research they hypothesized that that BMDMs with Hif1a defect would also show problems with eicosanoid production because there is a defect in LD accumulation. Using liquid chromatography - tandem mass spectrometry(LC-MS/MS) eicosanomic profiling was used to look at signaling molecules made by the enzymatic or non-enzymatic oxidation of arachidonic acid or other polyunsaturated fatty acids (PUFAs) that are, similar to arachidonic acid (13). The eicosanomic profiting was done on BMDM with several M tuberculosis situations including inhibition of LD formation via genetic and pharmacologic methods and this was an in vitro system. For many of the eicosanoids, M tuberculosis infection induced production but it was significantly raised with the addition of IFN-{gamma}. In particular, PGE2 correlated with LD formation in M tuberculosis infection. In HIF-1{alpha} deficient BMDMs many eiconsanoids were deficient. Together the data confirmed the hypothesis that LDs are important as a site for prostaglandin and lipoxins productions and that they also enhance the production of these substances. One surprise in this portion of the research was that in Hig2 defective BMDMs there was not seen a defect in eicosanoid production (FIG 6D, 6E). The authors wrote that this may be due to the late timepoints in which the Hig2 defective BMDMs have a large effect in LDs and that this timepoint is when there is limited eiconsanoid production.

Does M tuberculosis have the ability to accumulate lipids in IFN-{gamma} activated BMDM when the LD formation is inhibited? To investigate this the researchers used Hif1a defective BMDM that were infected with M tuberculosis because these macrophages had been shown previously to have equivalent TAG levels as wild type BMDM (FIG 4G). Both types of BMDM (wildtype and Hif1a defective) were infected with M tuberculosis. Using fluorescent labeling, lipid accumulation was assessed and it was shown that at 3 days after infection there were many LDs in the cytosol of the infected and IFN-{gamma} activated wild type BMDM. No staining was noted on the bacteria. In BMDM defective in Hif1a, the bacteria was shown to have the ability to accumulate lipids again. These experiments showed that M tuberculosis was unable to accumulate the host lids in IFN-{gamma}expressing macrophages. It is important to recall that this IFN-{gamma}is activated through HIF-1{alpha}.

Lastly, the researchers examined the role of LDs as a source of nutrients for M tuberculosis in vivo. To understand why this is important the paradigm that the lipid droplets (LDs) in macrophages are a carbon source for M tuberculosis. First the researchers tested whether in vivo M tuberculosis LDs in lung lesions require IFN-{gamma} driven, HIF-1{alpha} dependent expression of Hig2. Using Oil Red O neutral lipid staining on lung sections of 5 different genetic types of mice including wild type and His 2 defective and it was found that those who had a defective pathway for IFN-{gamma} dependent LD had fewer LDs as compared to wildtype mice. This suggests that the pathway the researchers had identified earlier for LD formation in vitro was also working in vivo. Next the it was shown that the lack of LDs in Hig2 defective had an effect on in vivo growth of M tuberculosis by looking at the bacterial burden in the lungs at days 18 and 28 post infection. There was no difference seen between wildtype and Hig2 defective, (which had less LDs), in the numbers of colony forming units. These results counter the paradigm that LDs in macrophages area carbon source for M tuberculosis.

In summary this paper reported on several experiments that showed the following things:

1. IFN-{gamma} is required for macrophage formation during M tuberculosis infection.

2. LD formation in infected macrophages is not correlated to M tuberculosis attainment of host lipids.

3. LD formation is associated with increases in intracellular triglycerides and cholesterol esters. This LD formation requires the fatty acid translocate CD36.

4. TAG synthesis contributes to LD formation during M tuberculosis infection of IFN-{gamma} activated macrophages, but FA synthesis does not.

5. LD formation during M tuberculosis infection requires HIF-1{alpha}.

6. LD maintenance during M tuberculosis infection requires the HIF-1{alpha} target gene Hig2

7. Host immunity is supported by LDs in M tuberculosis infected macrophages.

8. M tuberculosis in unable to acquire the host lipids in IFN-{gamma} activated macrophages that express HIF-1{alpha}.

9. LDs are not a source of nutrient for M tuberculosis growth both in vitro and in vivo.

Due to the elegant and stepwise nature of this research, the conclusions as they are enumerated above are justified. Contrary to the working model and paradigm that in M tuberculosis LDs in macrophages use these LDs as a carbon source for nutrition, I agree that this work demonstrates M tuberculosis does not induce formation of LDs. The work clearly demonstrates that IFN-{gamma} is part of the immune process of the host both by examine the immune processes that occur via IFN-{gamma} activation of HIF-{alpha}. The designs and methods of this research are compelling and I see not reason to think that the conclusions are not justified. I found this work exciting in its refusal to accept a precious paradigm and the use of newer research techniques, (ex: CRISPR/CAS9), to challenge the paradigm.

The further work done to examine whether LDs are or are not a source of nutrients in the in vitro environment is well done and the results justify the conclusions. However, this in one line of research that needs further exploration. Since there have been studies that show the tandem positioning of phagosomes with M tuberculosis in LDs in vitro which contrast with the work in this paper, it will be important to tease out in future work whether this is from an HIF-1{alpha} dependent mechanism to counteract the infection or if it is changes in bacterial metabolism in the IFN-{gamma} mediated immune response that was discussed in this paper. Finally, this sort of work into which system affect the host's ability to fight or contain M tuberculosis could eventually lead to pharmacologic research into new drugs to treat TB especially since TB the cause of death in over one million people a year through the world.

Cite this page