Update your browser to view this website correctly. Update my browser now
Since its discovery in the human stomach, the bacterium Helicobacter pylori has been branded as the cause of gastric diseases. This association is linked to the oncogenic toxin CagA produced by certain H. pylori strains, which causes severe damages but needs to be injected into the host cells to exert its toxic effect. Injection is achieved by a special bacterial transport mechanism, the Cag Type IV secretion system (Cag-T4SS). However, in nature, not all H. pylori strains infecting a patient contain the CagA toxin and the Cag-T4SS. In accordance with this, we have performed pre- and co-infection experiments of human cells in vitro with several strains of H. pylori. These experiments revealed that host cells can build up a resistance to the injection of CagA and to cellular damages associated with it. In order to further understand the mechanisms involved in this behavior, we analyzed the scope of it by looking at single aspects of infection. These included an evaluation of time of pre-infection and its effect on CagA translocation and cytokine response of the host cells to the Cag T4SS. Additionally, because of the high genetic variability of H. pylori, it was necessary to study the outreach of this phenomenon during the combination of different wild-type strains. It is remarkable that this phenomenon was not only observed in the epithelial host cell model but as well in primary cells from a hematopoietic origin, suggesting a relevance of the resistance mechanism to the CagA toxin and the Cag T4SS in the way the immune response is triggered during infection.
Self-infection experiments with Helicobacter pylori performed by Barry Marshall labelled H. pylori as the pathogen responsible for gastric pathologies with the cagA gene as one of the first genes correlated with severe gastric pathologies. Its product, the CagA toxin, is injected into human cells by a Type IV secretion apparatus, encoded by H. pylori in the cag pathogenicity island (Cag-PAI). Inside the host cell CagA is phosphorylated by host cell kinases. A functional Cag T4SS is not only able to inject CagA into host cells but it also induces a strong pro-inflammatory chemokine response that includes IL-8 and IL-1ß. Since CagA found inside the host cell, either phosphorylated or not, is able to disturb several signaling pathways, it is necessary to understand under which conditions the bacterium injects the toxin into the host cells. In nature, a human stomach can contain multiple H. pylori strains at the same time. Some strains may contain a functional CagA T4SS, denominated as Type I strains, while others are free of it, and are classified as Type II strains. During competition experiments between different H. pylori strains, the amount of CagA injected into the host cells is strongly reduced. This effect is independent of the strain's binding capacity but highly specific for a given H. pylori strain.
Two Helicobacter pylori strains infecting a host cell reduce the amount of CagA toxin injected by the type I strains. The objective of the study is to discern some aspects of the resistance response of host cells to CagA translocation induced by H. pylori strains. For this, several experiments were performed addressing the following questions: i) How fast can the resistance be formed? How does it behave over time? ii) Is it a specific response of epithelial cells or do immune cells resist as well CagA intoxication? and iii) Can all -Type I strains be blocked by all H. pylori strains in the same level? Since H. pylori CagA translocation is closely related to the production of pro-inflammatory cytokines, all analyses were complemented with measurements of the cytokine IL-8 during the different assays.
The simultaneous infection of epithelial cells in vitro using two bacterial strains with and without a Cag T4SS resulted in a fast reduction of CagA translocation into the host cells. (see Experimental setup, Fig. 1A). This reduction in CagA translocation of the second strain (type I) was even stronger with a 60 min pre-infection of the first (type II) strain. Two possible scenarios for the differences in resistance are i) the co-infection effect is the first effect visible and should increase linearly with longer pre-infection times, or ii) both cellular reactions to pre- and co-infections are independent events. Our experiments evaluated the first 15 min of pre-infection (Lapses of 5 min), up to 60 min response, with lapses of 10 min (Fig. 1B). The effect observed at simultaneous infection was maintained up to 40 min; and even though the variability of the results diminished with time, the differences (mean value/SEM) to the previous time points were not statistically significant. However, the CagA translocation is drastically reduced at the time points 50 min and 60 min, raising the possibility that at 50 min a new mechanism is activated resulting in a stronger blocking of CagA translocation.
For CagA translocation to occur, the Cag T4SS has to be functional, which implies that it also will induce the secretion of IL-8 by epithelial cells. Therefore the question arises whether or not a pre- or a co-infection situation has an effect on the IL-8 production induced by a functional Cag T4SS. These experiments show that both pre- or co-infection situations caused a reduction of IL-8 secretion as a response to the function of the Cag T4SS. The differences in IL-8 production starts to be visible at pre-infections of 15 min and maintained at 60 min pre-infection (Suppl. Fig. 2A). These results are similar to the ones observed with CagA translocation (Fig. 1B). The previous experiments were performed with strains having an identical genetic background using as pre-infecting strain H. pylori P12 lacking all T4SS (P12ΔT4SSs). This kind of experiments may represent the behavior of an H. pylori strain living in the host that turns off their Cag T4SS by genetic switch. However, in human infections, genetically different strains would be expected. For this reason, we compared different wild-type trains, using as a control the P12ΔT4SSs (originally a type I strain) and two type II strains (CagA (-); Tx30a and X47) as pre-infecting strains. Although all pre-infecting strains (P12ΔT4SSs, Tx30a and X47) caused up to 80% reduction in CagA translocation by the secondary strain (P12, P145, G27, P217 and 26695), there was a significant difference between the effect caused by the P12ΔT4SS strain and the natural type II strains (Fig. 1C).
However, when looking at each CagA-translocating strain, it is visible a different response to the pre-infecting strain. While P12, P145 and 26695 strains (Suppl. Fig. A–E) show no significant differences between the amount of translocated CagA blocked by the different pre-infecting strains, the strains G27 and P217 respond differently. G27 can translocate less CagA if the P12∆T4SSs is the pre-infecting strain (Suppl. Fig. 1C), and P217 capability to translocate CagA is much less affected by pre-infections with the Tx30a strain (Suppl. Fig. 1D).
The effect of a co- or pre-infection on IL-8 secretion was similar to the one observed with CagA considering all different strains tested (Fig. 1D), with a significant difference observed between the control infection and the pre-infections, dependent on the type of pre-infecting strain. However, the host cell responded differently to the pre-infection with different strains combinations. In this case, the different levels of IL-8 secreted are stronger with strains P145 and G27 (Suppl. Fig. 2B and 2D), while P12 was affected only by the P12∆T4SSs (Suppl. Fig. 2B), and strains P217 and 26695 could induced similar levels of IL-8 induction in presence of the pre-infecting strains as they do alone (Suppl. Fig. 2E and 2F).
Comparing these data, it is visible that there is no specific correlation between lower CagA translocation and lower IL-8 induction in pre-infecting assays for each strain. Variations might be explained by Cag T4SS independent factors on IL-8 production, such as OipA or others, which have not been analyzed.
The blocking assays performed so far used epithelial-like cells; however, the Cag T4SS affects as well immune cells. And although the specific effect of CagA in these cells is not know, in the host, there is a high probability that H. pylori will have a direct contact with leucocytes in damaged gastric tissue. We, therefore, tested the effects of co- and pre-infection in the standard cell line THP-1, being our model for immune cell and CagA translocation, and verified the observation in primary leucocytes. Co- and pre-infection experiments of primary leucocytes and THP-1 cells (monocytic leukemia cell line) using AGS as control showed that both cell cancer lines (AGS and THP-1) behave similar, while isolated primary leucocytes respond very strong to both treatments without differences between co- and pre-infection conditions (Fig. 1E to 1F). Although cancer epithelial cells are excellent models, the analysis of immune cells response to this phenomena shows us that i) it is not only an in vitro- cancer-cell type effect and; in this specific case, ii) that immune cells will respond strongly to multiple infections. To our surprise, a pretreatment effect on IL-8 could not be evaluated for THP-1 cells and primary leukocytes, since they showed no difference on IL-8 induction between the T4SSs mutant and the wild-type (Suppl. Fig. 3B and 3C), suggesting that IL-8 induction in these cells is independent of the Cag T4SS. This is in contrast to the response from AGS cells (Suppl. Fig. 3A) and previously published association of IL-8 induction and T4SS.
CagA and its associated Cag T4SS are responsible for the injection and phosphorylation of CagA in the host cell, together with the induction of inflammatory chemokines, like IL-8. By measuring both CagA translocation and IL-8 induction as a parameter to evaluate the effectiveness of Cag secretion system in co- and pre-infection assay, our results show that i) different levels of resistance to CagA translocation and IL-8 induction involve fast processes in the host cell, which highlights the importance of host-bacteria interaction in the dynamics of pathology. ii) Although the resistance is a prevalent response to the contact of H. pylori to epithelial or immune cells, each combination of strains has a different effect on CagA translocation resistance and IL-8 induction, making the already complicated life of H. pylori in the stomach fascinating and the outcome unpredictable in cases of multiple infections, and iii) although immune cells respond differently to the multiple infection setup in levels of IL-8 secretion, the effect on CagA translocation is conserved and even stronger than in the epithelia cell line model.
The knowledge we have about the injection of CagA toxin is limited to the use of cell culture. Although mice and gerbils are used as animal models, Helicobacter pylori has adapted to humans. In humans, most of the data about colonization did not contribute to a real picture about the levels of multiple H. pylori strains in one host. Our studies in cells are just an approximation of what could happen in the human mucosa upon having two different strains, but better in vitro models, such as primary polarized gastric cell culture models or human gastric organoids could bring us closer to a better understanding of the real interaction taking place in the host tissue.
The search for the molecular trigger on the bacterial-host interaction is complicated because of the remarkable variability of H. pylori, not only in their genetic level but as well their capacity of changing their proteins through variations, which complicates the work with different strains.
Because CagA translocation is measured by levels of phosphorylation of the toxin by host cell kinases, it could be argued that the cellular resistance to translocation is an effect on kinases and thus the phosphorylation of CagA, rather than translocation. For this, we have verified the translocation reduction using a beta-lactamase-fused CagA (TEM1-CagA), which measures the translocation of the toxin in a kinase-independent manner. The experiments were successful, as we see less beta-lactamase activity during co- and pre-infections (data not shown). Despite the generally high sensitivity of this assay, we were not able to determine small variations in translocation levels as accurately as it can be done by the measurement of phosphorylation with semi-quantitative western blot, probably because this assay is based on an enzymatic activity. Similarly, it might be argued that changes in adhesion of all bacteria (primary and secondary) could have an influence on the translocation efficiency. Our data showed no changes on adhesion of a translocating (type I) strain under pre- or co-infection conditions, as compared to a single infection (data not shown), which rules out diminished translocation caused by a binding competition between strains.
Since the resistance to CagA translocation seems to be caused by two different mechanisms, one early-onset and another delayed, we will evaluate the involvement of different cellular processes separately. As fast responses (shown by co-infection) the best candidates will include membrane processes and membrane lipid composition. For the 60 min response, we will concentrate on “slow” processes, such as de novo protein synthesis, protein recycling processes and protein modification systems. Because of the strong variations between different wild-type strain combinations, it will be necessary to evaluate the cellular processes relevant for each type of CagA translocation event.
With respect to bacteria, we need to verify the validity of the mutations of outer membrane proteins that we have found play a role in the resistance of the host cell. We are, as well, verifying the relevance of the in vitro data to the development of gastric diseases in humans. If confirmed, we would have found a new factor in the development of gastric pathologies in the presence of H. pylori. This will open the possibility that disease only is present when a single Type I H. pylori colonizes its host, and medical treatment can be changed from the actual eradication treatment to a supplement with a type II strain fitted for the strain found in the patient's stomach.
All strains used in this work have been published previously. P12ΔT4SSs is a mutant from P12 lacking all Type 4 Secretion systems, and two flagella components: FlaA and FlaB. Strains P12, P217, P145, G27, Tx30a, X47 and 26695 are all wild-type isolates from human stomach biopsies. List of strains and some genotype and Cag PAI associated phenotypes are shown in Supplement section (Suppl. Fig. 4B).
Cells lines used were AGS cells (ATCC® CRL-1739™) and THP-1 (ATCC® TIB-202™). All cell lines were cultivated in RPMI 1640 media complemented with 10% FBS, at 37°C in an atmosphere containing 5% CO2.
All reagents used are described in the protocols placed in the repository (see below).
Co- and pre-infections
Infections were modified based on the protocol submitted to the Protocol repository Protocols.io under Co-infection: DOI:10.17504/protocols.io.hjpb4mn; Pre-infection DOI:10.17504/protocols.io.gz5bx86.
Semi-quantitative analysis of CagA translocation
The exact protocol for CagA phosphorylation and its semi-quantitative analysis is found in the protocol repository Protocols.io under DOI:10.17504/protocols.io.hjzb4p6. Examples of western blot signals are shown in supplement figure 3F and 3G.
For IL-8 detection, we used a sandwich ELISA assay. The exact protocol and components used are in the protocol repository Protocols.io under DOI:10.17504/protocols.io.gz7bx9n. To calculate the changes in IL-8, the background IL-8 signals were subtracted from all samples. The effect of synchronization of cells on IL-8 induction by AGS cells has been graphed in supplement figure 4A.
Isolation of blood leucocytes
The protocol for isolation of human leucocytes from blood from human donors is in the protocol repository Protocols.io under DOI:10.17504/protocols.io.hjxb4pn.
All values were analyzed using GraphPad Prism Version 5.0. If not otherwise stated, all samples were analyzed with a one-way ANOVA and Newman-Keuls Multiple Comparison Test post-analysis to determine the relevance between treatments. For data analysis of IL-8 experiments and time, a Two-way ANOVA was performed. For figure 1E–G, a Student t-test was performed between co- and pre-infection.
This work was supported by research grants of the DFG (JI 221/1-1) to LFJ-S and the Förderprogramm für Forschung und Lehre (FöFoLe) program of the Medical Faculty of the Ludwig Maximillians University, Reg. Nr. 26/2013 - Munich, LMU to LFJ-S and RH.
The experimental work and analysis was performed by AFZ, LFJ-S. Experimental design by LFJ-S. Manuscript design and writing by KHG, RH and LFJ-S. We thank Dr. Wolfgang Fischer for his support with the leucocyte extraction.
The primary leukocytes were isolated from human blood (Permission 114-16, Ethic commission approval to Dr. Wolfgang Fischer) from human volunteers.