Abstract
Mycobacterium marinum (Mm) is a promiscuous pathogen infecting many vertebrates including humans, whose persistent infections are problematic for aquaculture and public health. Among unsettled aspects of host-pathogen interactions, the respective roles of conventional and innate-like (i)T cells in host defenses against Mm remain unclear. Here, we developed an infection model system in the amphibian Xenopus laevis to study host responses to Mm at two distinct life stages, tadpole and adult. Adult frogs possess efficient conventional T cell-mediated immunity, whereas tadpoles predominantly rely on innate-like (i)T cells. We hypothesized that tadpoles are more susceptible and elicit weaker immune responses to Mm than adults. However, our results show that although anti-Mm immune responses between tadpoles and adults are different, tadpoles are as resistant to Mm inoculation as adult frogs. Mm inoculation triggered a robust pro-inflammatory CD8+ T cell response in adults, whereas tadpoles elicited only a non-inflammatory CD8 negative- and iT cell-mediated response. Furthermore, adult anti-Mm responses induced active granuloma formation with abundant T cell infiltration and associated with significantly reduced Mm loads. This is reminiscent of local CD8+ T cell response in lung granulomas of human tuberculosis patients. In contrast, tadpoles rarely exhibited granulomas and tolerated persistent Mm accumulation. Gene expression profiling confirmed poor tadpole CD8+ T cell response contrasting with the marked increase in transcript levels of the anti-Mm iT cell receptor rearrangement (iVα45-Jα1.14) and of CD4. These data provide novel insights into the critical roles of iT cells in vertebrate anti-mycobacterial immune response and tolerance to pathogens.
Introduction
Mycobacteria marinum (Mm) resides in marine and freshwater and is capable of infecting a broad range of aquatic species including fish, reptiles, amphibians (e.g., Xenopus laevis [1]), and mammals including humans (reviewed in [2]). Mm is difficult to eradicate in aquaculture, especially when introduced to a recirculating water system. This pathogen, which causes fish mycobacteriosis outbreaks, can severely impacts aquaculture as reviewed in [3]. In human, Mm causes opportunistic infection in the skin, and poses public health risks [3]. As such, a better understanding of host immune responses to Mm remains crucial to improve diagnostics, treatment options, and vaccine strategies.
Importantly, Mm often serves as a useful Biosafety Level 2 (BSL2) alternative pathogen for Mycobacterium tuberculosis (Mtb), the causative agent for important human tuberculosis (TB; [4]). Like Mtb, Mm has been shown to survive within host macrophages, and is able to induce caseating granulomas in a zebrafish model [5]. Furthermore, studies have shown that Mtb and Mm share selected virulence determinants such as ESX-1 secretion system for phagosomal arrest in the host macrophages [1]. Mm grows optimally at 30 – 33°C, which is lower than the optimal temperature of Mtb (~37°C) and replicates every 4 hours, which is markedly shorter than replication time of Mtb (~20 hr) and more convenient for in vitro studies. Therefore, investigation with Mm will further benefit our understanding of host anti-mycobacterial immune responses providing a valuable and practical model that may ultimately lead to the development of new immunotherapeutic-based strategies for TB.
X. laevis is an attractive comparative immunology animal model due to its fully sequenced genome, and the availability of large genetic and genomic resources, as well as the remarkable similarity of its immune system with that of humans [6, 7]. Unlike mammals, however, X. laevis undergoes metamorphosis and has distinct T cell populations prior to and after this developmental transition. X. laevis tadpoles lack an optimal protein level of classical MHC class I molecule, yet express multiple Xenopus MHC class I-like genes (XNCs). In mammals, MHC I-like molecules have been shown to restrict innate-like T (iT) cells that exhibit features of both innate and adaptive immune cell effectors. Similarly, several iT cell subsets have been identified in X. laevis [8]. These iT cell subsets are predominant in tadpoles representing as much as 80% of CD8low/negative splenic T cells, whereas they represent only a minority compared to conventional T cells in adults X. laevis.
Generally, tadpoles are considered to have a less efficient immune response and, thus, to be more susceptible to natural pathogens (reviewed in [9]). However, a more detailed investigation in X. laevis suggests that distinct specialization of immune response mediated by iT cells allows the immune competence of tadpoles against distinct pathogens [10, 11]. Importantly, one of the iT cell populations exhibiting the invariant TCRα rearrangement iVα45-Jα1.14, has been shown to be critical for tadpole resistance to Mm infection [11].
In mammalian models, two iT cell subsets have been studied in the context of anti-mycobacterial immune response: mucosal associated innate T cells (MAIT) and invariant natural killer T cells (iNKT). A few clinical studies, along with in vitro studies suggest protective functions of both types of iT cell at least at early stage of mycobacterial infections (reviewed in [12, 13]). However, investigation of the specific regulatory and/or effector functions, activation, and recruitment of iNKT and MAIT cells have been challenged by the lack of suitable animal models (reviewed in [14]). As such X. laevis, exhibiting a distinct T cell balance between different life stages, would provide a useful model system to gather new insights into iT cell-mediated immune functions during mycobacterial infection.
Here, we report evidence of a distinct immune response and histopathology between X. laevis adults and tadpoles towards Mm. In support of our hypothesis that the respective response of iT and conventional T cells is distinct between tadpoles and adult frogs, we found markedly different conventional anti-Mm T cell response, histopathology and Mm growth between these two life stages.
Materials and Methods
Animals.
Two-year-old adult frogs and three-week-old pre-metamorphic outbred tadpoles (stages 52) were obtained from our X. laevis research resource for immunology at the University of Rochester, Rochester, New York, USA, http://www.urmc.rochester.edu/smd/mbi/xenopus/index.htm. All animals during experiments were carefully handled under the University of Rochester Committee on Animal Resources regulations (approval number 100577/2003–151).
Mm inoculation.
The Mm strain PM2960 was derived from a clinical isolate (stock number PM2690) generously provided by Dr. Hardy, University of Rochester, Rochester, NY. Fluorescent Mm strain (PM3495) was generated by transformation of a plasmid (pMV261.Kan.DsRed), kindly provided by W. R. Jacobs, Albert Einstein College of Medicine, Bronx, NY, into the parental PM2960 strain of Mm. Mm was cultured in Middlebrook 7H9 broth until saturation, and tittered stocks prepared and frozen at −80°C as previously described [11]. A working concentration for intraperitoneal injection into the animals and for in vitro infection of leukocytes was prepared by diluting the growth media in the amphibian phosphate buffer solution (APBS) with 0.05% of Tween 80. Tissues were harvested from euthanized animals at indicated days post-inoculation or from postmortem animals for further analyses.
Survival analysis of Mm inoculated adults and tadpoles
To examine the host resistance of X. laevis adults and tadpoles to Mm infection, 4-month-old adult frogs and 3 week-old pre-metamorphic tadpoles (stage 52) [15] were intraperitoneally inoculated with three different doses of Mm: low dose (5×105 CFU), medium dose (1×106 CFU), and high dose (2×106 CFU). The survival of the animals was monitored daily for one month and we investigated the effects of host life-stage (tadpole or adult) and dose on survivorship.
Histopathology, immunohistochemistry (IHC) and in situ hybridization (isH).
The liver of post-mortem adult frogs (euthanized at 14 and 50 dpi, n=1 each) and whole tadpoles (euthanized at 30 dpi, n=2) inoculated with Mm were fixed in 10% formalin and processed for routine histologic examination. Based on previous observation, the liver serves as the main organ infected by Mm [11]. Cross sections of the adult livers and full length longitudinal sections of the tadpoles were stained with Hematoxylin & Eosin [16], as well as special stains for bacteria and acid-fast organisms: Gram [17], Ziehl Neelsen [16] and Fite Faraco [18]. Microgranulomas were counted in three random low power fields (10x) from 4 adults (50 dpi) and 4 tadpoles (30 dpi). The average number of microgranulomas per individual was used to represent the number of microgranulomas per 0.5 mm2 of liver tissue. Immunohistochemical (IHC) staining with anti-CD3 receptor antibody (Leica Cat # PA0553; [19]) was performed using an automated platform to identify conventional T cells (Bond-Max IHC/ISH platform, Bond polymer refine DAB kit, Bond polymer refine red kit; Leica Biosystems, Newcastle Upon Tyne, United Kingdom). A negative control (staining without the primary anti-CD3 staining) was run along with the test slides. A tissue was considered positive if strong and distinct staining with anti-CD3 antibody was present in the membrane of individual cells and background staining was either absent or clearly distinct from true specific staining. An in situ hybridization (isH) probe for Mycobacterium sp (ACDBio RNAscope Probe-B-MBovis-23SrRNA, Cat No. 446011, Target region 3–3 – 1286, GenBank NR_076088.1, Brosch 2007) was applied following the manufacturer’s specifications to identify the presence of Mm.
RNA, and genomic DNA isolation from tissues, RT-PCR, and PCR.
Total RNA and genomic DNA were extracted from the animal’s tissues by using TRIzol reagent, following the manufacturer’s protocol (Invitrogen). Total 2 μg of RNA were used to synthesize complementary DNA (cDNA) by a reverse transcriptase, M-MLV (Invitrogen) with a mixture of oligo(dT) primer (Invitrogen). For reverse transcription (RT)-PCR, 125 ng of cDNA were used to determine the expression levels of genes of interest by ΔΔCT value with ABI 7300 Real-Time PCR System and PerfeCTa SYBR Green FastMix ROX. The expression levels are normalized to that of an endogenous housekeeping gene, gapdh, then further normalized against the lowest observed expression. All the primers are validated prior to use. Further, the absolute quantification method was done using the quantitative RT-PCR (qRT-PCR) analysis. Using a Mm 16srRNA PCR fragment was cloned into the pGEM-Easy Vector (Promenga), further transformed into DH10β competent bacteria for amplification. The plasmid stock was serially diluted in the range between 1010 to 101 plasmid copies of 16srRNA Mm in order to generate a standard curve by absolute qRT-PCR. To determine Mm loads, the total 250 ng of genomic DNA was used as a template, then the absolute copy number of Mm 16srRNA genome was extrapolated from the standard curve. All primer sequences are listed in Table 1.
Table 1:
List of qPCR primer sequences
| Primer | Sequence (5’- 3’) |
|---|---|
| CCR2 | F:ATTGGGCAGAACTGTGGTAG |
| R:GGGCGAGTAATCTGAGTCATAA | |
| CD3 | F: TTGGGCTCAGTGTGGAATG |
| R: GGTCCCGGTATCCATCTCTAT | |
| CD4 | F: CATAGTGGTTTCCCTCTGGTTTAG |
| R: CGCAGAGCGTCCATTCATTA | |
| CD8β | F: GGAACACGTTTACCCTGAAGA |
| R: GGGAGGTTCCATTCCCAAAT | |
| CSF-R1 | F: TGTATTCTTTGGACTTGCCGTATCTGG |
| R: TTGTTTAGCTTCAAATTCTGGGTAATA | |
| GAPDH | F: GACATCAAGGCCGCCATTAAGACT |
| R: AGATGGAGGAGTGAGTGTCACCAT | |
| GMCSF-R | F: ACGTGCCAGCTAAACCTCACAGAT |
| R: TGACACAGCCTGGGCGAGAAATAA | |
| IL-1β | F: CATTCCCATGGAGGGCTACA |
| R: TGACTGCCACTGAGCAGCAT | |
| IL-10 | F: TGCTGGATCTTAAGCACACCCTGA |
| R: TGTACAGGCCTTGTTCACGCATCT | |
| iNOS | F: AACCGTAAGCCAAAGAAGGA |
| R: TGGTTCTGGCAGCCACAGT | |
| Mm 16S rRNA | F: AGAGTTTGATCCTGGCTCAG |
| R: CACTCGAGTATCTCCGAAGA | |
| TGFβ | F: CCCACAGGCCAAAGATATAGAC |
| R: CATCAGGTAGGGTTTCGTGTT | |
| TNF-α | F: TGTCAGGCAGGAAAGAAGCA |
| R: CAGCAGAGCAAAGAGGATGGT | |
| R: CCGACACACTGAGCGGAAA | |
| Vα45-Jα1.14 | F: TCCGTTAACGAGAAGGATTCCCAG |
| R: CTCCCAGCCACTACCAGAATAAG |
F: Forward; R: Reverse
Statistical Analysis.
For studying the kinetics of Mm loads, a non-parametric Kruskal-Wallis test followed by Dunn test for multiple comparisons was performed. For RNA expression analyses, a mixed linear regression model analysis was performed to compare the kinetics of adult and tadpoles followed by the Kruskal-Wallis test to compare individual time points within adults or tadpoles group. For the survival analysis, a Log-Rank test and the Cox Proportional Hazard Model analysis was performed using R (R version 3.5.2, R studio version 1.1.463). GraphPad Prism 6 software (San Diego, CA, USA) was used for all statistical computation except for Cox Proportional Hazard Model analysis.
Results
Comparison of survival rates following Mm inoculation between X. laevis adults and tadpoles
To examine the respective host resistance of tadpoles and adult frogs to Mm, we monitored survival following different doses of Mm inoculation. Although tadpole death appeared to start earlier than adults for low and medium Mm dose (Fig 1), the mortality rate was not statistically different using a Log-Rank test. To substantiate our analysis, we tested our data using a final statistical model and set different Mm doses and two life-stage (adults and tadpoles) as covariates. The analysis indicates survivorship was not significantly impacted by life-stage (p = 0.1357), instead survivorship appeared to be Mm dose-dependent. The interaction between life-stage and dose was non-significant in our initial model. Further, survivorship decreased with increasing Mm dose: medium and high dose both had significantly lower survivorship than low dose (p = 0.0404; p < 0.0001, respectively) and high dose had significantly lower survivorship than medium dose (p < 0.0001). The median survival time for each dose is indicated in Table 2. Collectively, our survival analysis suggests that tadpoles are not significantly more susceptible against Mm compare to adult frogs.
Figure 1. Survival curves of Mm inoculated adults and tadpoles of X. laevis.
(A) 4 month-old adults (n=5–10/group) and 3 week-old tadpoles (developmental stage of 52, n=10–19/group) were inoculated with different amount of Mm intraperitoneally (5×105 CFU for low dose, 1×106 CFU for medium dose, and 2×106 CFU for high dose). The survival rate was dependent on the doses of Mm based on the Cox Proportional Hazard model analysis (p < 0.05). Comparisons of survival rates between adult frogs and tadpoles for each dose as well as median survival times are listed in table 2. (B) Mm loads in different organs from post-mortem adults. (C) Mm loads of whole individual post-mortem tadpoles. Viral loads were determined by real-time PCR using Mm specific 16srRNA gene.
Table 2.
Median survival days for adult frogs and tadpoles infected with different doses of Mm intraperitoneally (dpi).
| Doses for intraperitoneal infection | ||||
|---|---|---|---|---|
| Life Stage | 3×105 CFU | 5×105 CFU | 1×106 CFU | 2×106 CFU |
| Adult frogs | N/A | 33 | 29 | 12 |
| Tadpoles | 30 | 30 | 17 | 10 |
| P value for Log-rank test | N/A | 0.7164 | 0.1176 | 0.9449 |
N/A means no data
We also quantified the Mm genome copy number of post-mortem animals by qPCR (Figure 1B, C). Relatively modest Mm loads were detected in tissues of post-mortem adult frogs (10 to 100 fold higher in respective tissues of post-mortem animals compared to live animals euthanized at 21 dpi). These Mm loads were not significantly different across organs or at different time of death (Figure 1B). Similarly, there was no correlation (p=0.9738 by Spearman r analysis) between the total Mm loads in tadpoles and time of death. The rapid decay of tadpoles dying from Mm inoculation did not allow to determine the bacterial load for particular organ without a risk of contamination.
Comparison of T cell-mediated immune responses in X. laevis adults and tadpoles
To address our hypothesis that adults and tadpoles have distinct T cell responses during Mm infection, we performed flow cytometry using Xenopus specific anti-CD8 mAb and anti-CD5 mAb to define the response kinetics of two main T cell populations as previously established using adult splenocytes [20]: a CD8+ T cell population defined as cells co-expressing the CD8 and CD5 pan-T cell marker (CD8+/CD5+) and CD5+ T cell population not significantly expressing CD8 (CD8neg/CD5+) that includes iT cells and presumably CD4 T cells and that we will refer here as CD8neg T/iT cell population (Figure 2A–B). In the absence of lymph nodes, the spleen constitutes both a primary and secondary lymphoid organ in Xenopus [21]. Therefore, we monitored the kinetics of the two T cell subsets in spleen (Figure 3). We also examined the liver, which is a major site of mycobacteria infection in Xenopus (Figure 4).
Figure 2. Flow cytometric analysis of T cell in spleen and liver of adults and tadpoles.
4 month-old young adults and three-week-old tadpoles were intraperitoneally inoculated with 1×106 CFU, or 3×105 CFU of Mm, respectively (n=5–6 per each time points). Then, the total lymphocytes at different dpi were stained with Xenopus specific CD5 mAb, and CD8 mAb to analyze two subsets of T cell populations (A-B). After gating on live cells, CD8+CD5+ cells and CD8negCD5+ cells were defined (black boxes).
Figure 3. Comparison of the frequency and the number of CD8 and CD8neg T/iT cells in the spleen of adults and tadpoles following Mm inoculation.
Using the flow cytometric strategy shown in the Figure 2, the kinetics of CD8 T cell (A) frequency, and (B) number were determined in adults (white) and tadpoles (black) at different days post inoculation (n=6–7/group from 2 independent experiments). Further, the kinetics of CD8neg T/iT cell (C) frequency, and (D) number were determined. C: uninfected control. Asterisks indicated statistical significance by the Kruskal-Wallis test, non-parametric.
Figure 4. Comparison of the frequency and the number of CD8 and CD8neg T/iT cells in the liver of adults and tadpoles during Mm infection.
Using the flow cytometric strategy shown in the Figure 2, the kinetics of CD8 T cell (A) frequency, and (B) number were determined in adults (white) and tadpoles (black) at different days post inoculation (n=6–7/group from 2 independent experiments). Further, the kinetics of CD8neg T/iT cell (C) frequency, and (D) number were determined. C: uninfected control. Asterisks indicated statistical significance by the Kruskal-Wallis test, non-parametric.
In adult frogs but not in tadpoles, there was an increase of the total number of splenocytes upon Mm inoculation that reached a statistical significance at 21 dpi, suggestive of immune cell expansion (Table 3). Notably, the frequency and the number of CD8 T cells were significantly increased in the adult spleen at 12 dpi (***p=0.001 compared to 6 dpi group, and **p=0.0154 compared to the uninfected group, respectively), whereas no significant change was observed in tadpoles (Figure 3A–B). In contrast, both the frequency and the number of the CD8neg T/iT-like cell were significantly reduced in tadpole’s spleen at 6 dpi, which suggests an egress of this T cell population (Figure 3C–D). Unlike tadpoles, the CD8neg T/iT cell population remained unchanged following Mm inoculation in the adult spleen. (Figure 3C–D).
Table 3.
Total number of lymphocytes (×103) ± standard error.
| Organ | Stages | Control | 6 dpi | 12 dpi | 21 dpi |
|---|---|---|---|---|---|
| Spleen | Adult | 1500 ±158 | 6000 ±1450 | 4860 ±2104 | 9580 ±1632 p*=0.0226 |
| Tadpole | 24 ±11 | 12 ±4 | 34±16 | 22 ±4 | |
| Liver | Adult | 505±63 | 3363 ±309 p*=0.0316 |
3650 ±1656 | 2100 ±710 |
| Tadpole | 30 ±4 | 25 ±6 | 44 ±15 | 32 ±15 |
Compared to control using non-parametric Kruskal-Wallis test followed by Dunn test.
Similarly, the total number of leukocytes recovered from the liver of Mm-inoculated adults but not tadpoles significantly increased at 6 dpi, suggestive of immune cell influx and/or expansion (Table 2). The frequency and number of CD8 T cells peaked at 12 dpi only in adult livers, whereas there were no marked changes in tadpoles’ liver (Figure 4A–B). However, the frequency of CD8neg T/iT cells significantly increased at 12 dpi in tadpole livers suggesting recruitment of these cell types in the liver (Figure 4C–D). In adult liver, the was a significant drop in the frequency of CD8neg T/iT-like cell at 6 dpi compared to uninfected controls that did not affect their cell number. Taken together, the data suggest that conventional CD8+ T cell response is elicited mainly in adults, whereas tadpole anti-Mm T cell response appears to rely more on the recruitment of CD8neg T cells that include iT cells.
Histopathology of Mm inoculated liver of X. laevis adults and tadpoles
Differences were noted in the histopathologic appearance of the livers between adults and tadpoles (Figure 5A–H). In the liver of post-mortem adult frogs, distinct clusters of epithelioid macrophages, indicative of microgranulomas, were scattered randomly throughout the hepatic parenchyma (Figure 5A and D, black arrows). Detailed examination of these microgranulomas in multiple section of different animals revealed that they were composed almost exclusively of histiocytic cells suggestive of epithelioid macrophages as shown in representative sections of Figure 5A and D. Often, the microgranulomas were surrounded by a rim of mononuclear cells with little cytoplasm and darker more condensed nuclear chromatin, consistent with lymphocytes. The largest lesions included basophilic debris, indicative of a poorly demarcated necrotic zone at the core of the microgranuloma (Figures 5A).
Figure 5. Histopathology of the liver of Mm inoculated adults and tadpoles and comparison between microgranuloma numbers.
4 month-old young adults and three-week-old tadpoles were intraperitoneally inoculated with 1×106 CFU, or 3×105 CFU of Mm, respectively (n=5–6 per each time points). Representative of liver sections that were stained with hematoxylin and eosin (A, D and G), an anti-CD3 mAb (B, E and H, brown staining), or an in situ hybridization (isH, red staining) probe for Mycobacterium sp (F and I). Adults inoculated with 106 CFU of Mm and euthanized 14 and 50 days post-infection (dpi): Larger granulomas (A and D, black arrows) surrounded by CD3+ cells in moderate (B) to high numbers (E) and positive isH for Mycobacterium sp in the cytoplasm of cells within the granuloma at 50 dpi (F). Tadpole inoculated with 3×105 CFU of Mm and euthanized 30 dpi: Small microgranuloma including cells with intracytoplasmic melanin granules (G, black and white arrows, respectively), with rare CD3+ cells (H) and positive isH for Mycobacterium sp in the cytoplasm of cells within the microgranuloma (I, arrow) Bars = 50 μm. The average numbers of microgranulomas in adults at 50 dpi (n=4) were significantly higher than those in tadpoles at 30 dpi (n=4) (C, non-parametric 2-way ANOVA, P<0.001).
In tadpoles, only a very small number of tiny epithelioid macrophage clusters were present, usually including distinct melanomacrophages or similar cells with nuclei obscured by melanin (representative shown in Figure 5G). Unlike adults, the microgranulomas present in tadpoles had no evidence of a necrotic zone. Based on the number of microgranulomas per unit of liver tissue, adult frogs exhibited a significantly higher number of granulomas compared to tadpoles (non-parametric 2-way ANOVA, p<0.001) (Figure 5C).
To visualize T cells in relation to granuloma structures, we used a cross-reactive human anti-CD3 Ab [19]. Following Mm inoculation, we detected an accumulation of CD3-positive (CD3+) T cells in both adults’ and tadpoles’ liver (Figure 5 B, E and H). More specifically, CD3+ T cells were conspicuously present throughout the parenchyma, either individually or forming small clusters. In the liver of Mm inoculated adults, the CD3+ T cells were more numerous in the periphery of the microgranulomas (Figure 5B and E, brown staining). In contrast, overall fewer CD3+ T cells were scattered almost uniformly throughout the hepatic parenchyma with little association with microgranulomas in tadpole’s liver (Figure 5H). Though very rare, a few CD3+ T cells were present within the core of microgranulomas in tadpoles’ liver (Figure 5E).
Special stains for bacteria and acid-fast organisms (Ziehl Neelsen and Fite Faraco, respectively) failed to detect the presence of any microorganisms in tissues from both adults and tadpoles [22]. However, an isH probe for Mycobacterium sp, produced amorphous staining in the cytoplasm of random cells in the liver of adults, and tadpoles (Figure 5F and I, respectively). Staining was slightly more structured, occasionally forming round vacuoles, in cells within hepatic microgranulomas compared to those scattered in the parenchyma. Subjectively, positive isH staining seemed higher within microgranulomas than in the rest of the tissue. Histological analyses indicated that Mm were sequestered in microgranulomas, which were larger and rimmed by abundant CD3+ T cells in adult frogs but smaller and less CD3+ rich in tadpoles.
Comparison of Mm dissemination in X. laevis adults and tadpoles
We have previously shown that in tadpoles intraperitoneally inoculated, Mm accumulates in the liver concomitant to a decrease of pathogen loads among PLs [23]. In addition, tadpoles that died at different days post Mm inoculation had no marked increase of Mm loads, suggestive of persistent Mm over time (Figure 1C). To determine whether Mm has a distinct dissemination pattern in adult frogs, we quantified Mm loads in different organs. Following i.p injection of 1×106 CFU of Mm in adult frogs, we assessed Mm loads in PLs that are the first cells to encounter Mm; the liver, the main site of Mm dissemination; and the spleen that is the major lymphoid organ of X. laevis.
Similar to tadpoles, we detected persistent Mm loads among adult’s PLs up until 18 dpi, followed by a significant reduction at 21 dpi (Figure 6A). However, unlike our previously published observation of increased of Mm loads in tadpoles’ liver, we found a significant decrease of Mm loads in adult’s liver at 21 dpi (Figure 6B). In addition, Mm loads significantly dropped at 18 dpi in the adult’s spleen (Figure 6C). Taken together, these data indicate that Mm disseminates systemically in adult frogs as in tadpoles. Unlike tadpoles, however, Mm loads significantly decrease in adult’s PLs, liver, and spleen over the course of Mm infection.
Figure 6. Determination of Mm loads and dissemination using an absolute quantification method and a recovery of live Mm in culture from adult and tadpoles of X. laevis.
Adult frogs were intraperitoneally inoculated with 1×106 CFU of Mm then, Mm loads were determined by using a real-time PCR with Mm specific 16srRNA gene (A-D). (A) peritoneal leukocytes (PLs), (B) liver, and (C) spleen tissues were taken at the indicated days post infection. The dashed line indicates the level of detection by real-time PCR. In order to measure only the live and replicating Mm, we cultured kanamycin resistant Mm from the inoculated adults and tadpoles at 21 dpi using a Middlebrook 7H10 media supplemented with 50 μg/mL kanamycin. (D) Quantitative measurement of CFUs was then normalized to total mg of homogenates.
Although the absolute quantification assay using qRT-PCR is a highly sensitive method to detect even low Mm genome copy number, the assay does not distinguish live infectious pathogens from dead or inactive Mm. As a complementary approach, live Mm from liver tissues were recovered on bacterial culture plates. The CFUs of Mm were measured from infected livers at 21 dpi, which was the time Mm load decrease was detected in adult frogs by qRT-PCR. Due to the size differences of the liver organs between adults and tadpoles, we normalized the total CFUs to mg of total tissue lysates. Notably, we detected significantly higher CFUs of Mm in tadpoles’ liver than that of adults at 21 dpi per 1 mg of total protein (Figure 6E). These data suggest that adult immune responses actively reduce Mm loads in the adults’ liver, whereas live Mm continues to accumulate in the tadpoles’ liver.
Changes in expression profiles of relevant immune genes in the liver
Based on the differences in the histology and the kinetics of Mm loads between adult frogs and tadpoles, we hypothesized that conventional CD8 T cells and innate immune cells were recruited and actively induced inflammation to control Mm in adult but not tadpoles. To address this, we determined the relative expression kinetics of genes encoding T cell co-receptors (CD3ε, CD4, and CD8β) as a proxy of recruitments in the liver. We used gapdh as a mean to normalize the expressions of immune gene transcripts and compared the gene expression profiles between adults and tadpoles. Of note, the Ct value of GAPDH was not markedly different at each time points and between adults and tadpoles (Figure 7E).
Figure 7. Relative expression of T cell-related immune genes in the liver of Mm inoculated adults and tadpoles.
Four month-old young adults (white bar) and three-week-old tadpoles (grey bar) were intraperitoneally inoculated with 1×106 CFU, or 3×105 CFU of Mm, respectively (n=5–6 per each time points). Relative gene expression in liver for (A) CD3ε, (B) CD8β, (C) CD4, and (D) iVα45-Jα1.14 was normalized to the housekeeping gene gapdh. (E) Ct values for gapdh of each time point of between adults and tadpoles. C: uninfected control, u.d: undetected value, asterisks indicate a significant difference by the Kruskal-Wallis test.
No increase in expression levels of CD3ε, CD4, or CD8β gene was observed in tadpoles following Mm inoculation (Figure 7A–C). In contrast, CD8β gene expression was drastically increased (more than 3 logs on average) at 6 dpi in adult liver, consistent with infiltration of conventional CD8 T cells in response to Mm inoculation (Figure 7C). To assess the response of iT cells following Mm inoculation, we determined the transcript levels of the invariant (i) Vα45-Jα1.14 rearrangement that is expressed by critical anti-Mm iT cell effectors [11]. In tadpoles, iVα45-Jα1.14 transcripts were undetectable in the uninfected liver, but became rapidly abundant at 6 dpi and remained detectable at 12 dpi (Figure 7D). A similar increase in iVα45-Jα1.14 transcript level occurred in the adult liver during Mm infection. These data support our hypothesis of a dominant anti-Mm iT cell in tadpoles contrasting with a combined iT and conventional CD8 T cell response in adult frogs.
To assess the involvement of innate immune cell effectors in adult and tadpole liver during Mm infection, we monitored the expression of CSF-1 receptor (CSF-1R) as a marker for macrophages; GM-CSF receptor (GM-CSFR) as a marker for neutrophils; and CCR2 as a marker for inflammatory monocytes in the liver (Figure 8A–C). In addition, we determined the expression of pro-inflammatory and anti-inflammatory cytokine genes (Figure 8D–H). Notably, CSF-1R gene expression peaked at 6 dpi in adult frogs and returned the basal level at 12 dpi (Figure 8A). The transcript levels of GM-CSFR and CCR2 did not change at the two time points tested (Figure 8B–C). Interestingly, the increase in CSF-1R expression was correlated with that of TNFα, IL-1β, and iNOS, which are pro-inflammatory cytokines (Figure 8D–F). We did not find a significant increase of anti-inflammatory cytokine expressions such as TGFβ, or IL-10 at 6 and 12 dpi (Figure 8G–H). In tadpoles, we observed a reduced expression of CSF-1R at 12 dpi compared to uninfected tadpoles. (Figure 8A). Although the kinetics of TNFα, IL-1β, and iNOS did not drastically change in tadpoles, we found that the expression of TGFβ significantly dropped at 12 dpi in parallel to CSF-1R (Figure 8G). Taken together, these data suggest that the acute influx or activation of macrophages (indicated by CSF-1R) and conventional CD8 T cells (indicated by CD8β cells) may contribute to early inflammation in adult frogs, whereas mainly iT cells (indicated by Vα45-Jα1.14 rearrangement) are recruited and persistent in tadpoles without marked inflammation.
Figure 8. Relative expression of immune receptor genes and pro- and anti-inflammatory cytokine genes in the liver of Mm inoculated adults and tadpoles.
Four month-old young adults (white bar) and three-week-old tadpoles (grey bar) were intraperitoneally inoculated with 1×106 CFU, or 3×105 CFU of Mm, respectively (n=5–6 per each time points). The relative gene expression in the liver was determined for the immune gene receptors: (A) CSF-1R (macrophage recruitment marker), (B) GMCSF-R (neutrophil recruitment marker), (C) CCR2 (inflammatory monocyte marker); for the pro-inflammatory cytokine genes: (D) TNFα, (E) IL-1β, (F) iNOS; and for the anti-inflammatory cytokine genes: (G) TGFβ and (H) IL-10. All the data were normalized to the housekeeping gene gapdh. C: uninfected control, u.d: undetected value, asterisks indicate significant difference by the Kruskal-Wallis test.
Discussion
The wide range of host species infected by Mm poses a concern not only for the aquatic ecosystem and fish industry, but also for its potential risk to public health. Aquatic species notoriously infected by Mm includes X. laevis, which we utilized in this study as a model to understand host anti-Mm immune response. Although the immune system of tadpoles is generally considered more immature and/or less efficient in controlling pathogens than that of adult frogs, our study rather suggests a distinct adaptation and specialization of immune responses for each these two life stages in X. laevis. Notably, unlike adult frogs that exhibit potent conventional CD8 T cell effectors, tadpoles do not express an optimal level of MHC I protein until the onset of metamorphosis and predominantly rely on immunity driven by MHC class I-like and iT cell subsets expressing [8, 23]. Here, we report evidence that suggests a distinctive host-pathogen interaction between adults and tadpoles during Mm infection.
First, in apparent contradiction of the general view of weaker tadpole immunity, we found a comparable survivorship between adults and tadpoles challenged with different doses of Mm. Considering the smaller size of the tadpoles compared to the adults, the inoculated tadpoles were remarkably efficient in tolerating high amount of Mm. This implies that tadpole and adult immune defenses, albeit distinctive, achieve similar survival against mycobacterial infections. To get further insights into the distinct adaptation of host responses against Mm by these two life stages, we first examined the T cell response by flow cytometry. Although specific antibodies to detect iT and CD4 T cells are currently missing in Xenopus, we were able to take advantage of CD5, a pan T cell marker in X. laevis in combination anti-CD8 mAb to monitor conventional CD8 T cells (CD8+/CD5+) versus a population of CD8neg T/CD5+ T cells, which have been shown to contain T cell expressing CD4 ([24]). CD8neg T/CD5+ T cells also include iT cells (e.g., iVa45 T cells) since they express very low level to no CD8 at all [8, 23]. Based on this approach we found that adults frogs exhibit a strong conventional CD8+ T cell response against Mm, which is not the case in tadpoles where the increase of CD8neg T cells and iT cells in the liver concomitant with their drop in the spleen suggests their rapid recruitment upon Mm inoculation. It is noteworthy that splenic IgM+ B cells are not CD5+ in X. laevis, except following strong stimulation with PMA [25]. Further investigation by gene expression profiling confirms that Mm inoculation induces a significant increase of CD8β transcript levels indicative of CD8 T cell response together with iVα45-Jα1.14 mRNAs in the adult liver. In contrast, only increased of iVα45-Jα1.14 but not CD8β transcript levels were detected in tadpoles in response to Mm inoculation (Fig 7). It is also interesting to note that CD4 expression significantly decreased, while iVα45-Jα1.14 transcript levels remained persistent in Mm inoculated tadpole’s liver (Figure 7B, and 7D). This suggests that iVα45 T cells rather than CD4+ T cells are major effector cells during Mm infection. This is corroborated by previously published reverse genetic loss-of-function by transgenesis evidence showing that iVα45 cells have a critical host protective function in tadpoles [11]. The mechanisms of activation and functions of these iT cells remain to be elucidated.
A striking difference between adult frogs and tadpoles revealed by this study was the granuloma formation resulting from Mm inoculation, which consistently occurred in adults but rarely in tadpoles. These granulomas, which consisted of epithelioid macrophages surrounded by adaptive T cells, were common in adult liver and to a lesser extent in other organs such as lung and spleen. Notably, IHC staining revealed an accumulation of CD3+ T cells at the rim of these granulomas, which is suggestive of a vigorous T cell response against an acute Mm infection. Histologic examination of the livers of adults and tadpoles occurred at slightly different stages post-inoculation and a definitive conclusion would require further confirmation. However, sampling times included early and late infection for adults and mid-term infection in tadpoles. Thus, we believe our findings support an age-dependent difference in immune response. It is uncertain why the special stains that normally pick out mycobacterial organisms failed to do so in the livers of adults or tadpoles. Confirmation of infection was achieved, however, though PCR and culture results, and with the positive staining of several cells with an isH probe for Mycobacterium sp.
The formation of granulomas is generally associated with inflammation in mammals, especially in cases of mycobacterial infections [26]. Likewise, in adult X. laevis, the expression of several inflammatory genes (TNFα, IL-1β, and iNOS) was significantly increased following Mm inoculation in the liver. Collectively, the immune histological analyses further strengthen the critical function of T cells in sequestering and controlling Mm infection in adult frogs. Importantly, the collective evidence of CD8 T cell-mediated response, inflammation-mediated granuloma formation, and the clearance of Mm in adult frogs are highly similar to the critical role of CD8 T cells in granuloma structures from Mtb infected human [27]. Therefore, our findings suggest that Mm infection in adult X. laevis may complement host mycobacterial infection in mammalian models.
In sharp contrast, microgranulomas were rarely observed in tadpoles (Figure 3C). In addition, Mm inoculation did not induce significant expression of inflammatory genes. Indeed, transcript levels of IL-1β remained low following Mm inoculation, whereas TNFα or iNOS expression was not induced by Mm. Importantly, the high CFU counts of live Mm recovered from the infected liver in this study substantiates the previous report of Mm accumulation in tadpoles [23]. Therefore, it is possible that the dominant iT cell response in tadpole allow the larval host to tolerate high and persistent Mm load and consequently to maintain the survivorship at a similar level as that of adult frogs. Alternatively, it also possible that the larval CD8neg T cell population includes regulatory T cells that contribute to maintain tolerance to Mm infection. Although, Tregs function has not been characterized in amphibians, gene orthologs encoding cytokines and transcription factors critical for differentiation of Tregs (e.g., FOXP3; CTLA4) are present in the genome of X. tropicalis and X. laevis (Robert, unpublished observations).
For an evolutionary standpoint is tempting to speculate that the predominant anti-Mm iT cell response in tadpole avoid the activation of inflammation, whereas a conventional T cell response in adult is associated with inflammation. As such the adaptive host response against Mm in two different life stages occupying distinct environmental niches is fundamentally different: the adult conventional T cell-based system is adapted to eradicate Mm pathogens by killing infected cells and generating an inflammatory response, whereas the tadpole iT cell-based system upon detection of the infection is designed to minimize inflammation and tolerate pathogen burden. Consistent with these contrasted immune responses, live Mm recovered from tadpole liver did not markedly decrease during infection, whereas it decreased in adults. The high Mm loads found in the adults’ post-mortem tissues by qPCR, further suggest an insufficient clearance of Mm resulting in the death of the adult frogs.
Active tolerance to pathogen has been reported in various cases including mycobacteria infection [28–30] Although not as well defined as immune mechanisms behind host resistance, specific tolerance mechanisms has recently raised attention owing to their relevance for better understanding host-pathogen interactions and potentials for developing more effective treatments for infectious diseases. Our data in tadpoles are consistent with an adapted tolerance to Mm infection that may involve iT cells as specific effectors controlling this tolerance. It will be interesting to examine in more detail how iT cells can establish and control host tolerance to Mm in tadpoles by determining the specificity (e.g., ligands presented by MHC-like molecules and recognized by iT cells) and the mechanisms involved (cytokine produced, cell types targeted).
Key points.
Xenopus tadpoles and adults have adapted distinct immune responses to Mycobacteria
A disease tolerance involving non-inflammatory iT cell response in tadpoles
A disease resistance with granuloma, inflammation and CD8 T cell response in adults
Acknowledgements
We would like to thank Tina Martin for the expert animal husbandry.
This research was supported by a predoctoral fellowship to RKH (T32AI118689), as well as R21AI139718 and R24AI059830 from the National Institute of Allergy and Infectious Diseases (NIH/NIAID) and IOS-1456213 from the National Science Foundation.
References
- 1.Tobin DM and Ramakrishnan L, Comparative pathogenesis of Mycobacterium marinum and Mycobacterium tuberculosis. Cell Microbiol, 2008. 10(5): p. 1027–39. [DOI] [PubMed] [Google Scholar]
- 2.Franco-Paredes C, et al. , Cutaneous Mycobacterial Infections. Clin Microbiol Rev, 2018. 32(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hashish E, et al. , Mycobacterium marinum infection in fish and man: epidemiology, pathophysiology and management; a review. Vet Q, 2018. 38(1): p. 35–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Cosma CL, Sherman DR, and Ramakrishnan L, The secret lives of the pathogenic mycobacteria. Annu Rev Microbiol, 2003. 57: p. 641–76. [DOI] [PubMed] [Google Scholar]
- 5.Cronan MR and Tobin DM, Fit for consumption: zebrafish as a model for tuberculosis. Dis Model Mech, 2014. 7(7): p. 777–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Session AM, et al. , Genome evolution in the allotetraploid frog Xenopus laevis. Nature, 2016. 538(7625): p. 336–343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Robert J and Ohta Y, Comparative and Developmental Study of the Immune System in Xenopus. Developmental dynamics : an official publication of the American Association of Anatomists, 2009. 238(6): p. 1249–1270. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Edholm ES, et al. , Nonclassical MHC class I-dependent invariant T cells are evolutionarily conserved and prominent from early development in amphibians. Proc Natl Acad Sci U S A, 2013. 110(35): p. 14342–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Robert J and Ohta Y, Comparative and developmental study of the immune system in Xenopus. Dev Dyn, 2009. 238(6): p. 1249–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Edholm ES, et al. , Nonclassical MHC-Restricted Invariant Valpha6 T Cells Are Critical for Efficient Early Innate Antiviral Immunity in the Amphibian Xenopus laevis. J Immunol, 2015. 195(2): p. 576–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Edholm ES, et al. , Distinct MHC class I-like interacting invariant T cell lineage at the forefront of mycobacterial immunity uncovered in Xenopus. Proc Natl Acad Sci U S A, 2018. 115(17): p. E4023–E4031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Le Bourhis L, et al. , Antimicrobial activity of mucosal-associated invariant T cells. Nat Immunol, 2010. 11(8): p. 701–8. [DOI] [PubMed] [Google Scholar]
- 13.Huang S, Targeting Innate-Like T Cells in Tuberculosis. Front Immunol, 2016. 7: p. 594. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hyoe RK and Robert J, A Xenopus tadpole alternative model to study innate-like T cell-mediated anti-mycobacterial immunity. Dev Comp Immunol, 2019. 92: p. 253–259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zahn N, Levin M, and Adams DS, The Zahn drawings: new illustrations of Xenopus embryo and tadpole stages for studies of craniofacial development. Development, 2017. 144(15): p. 2708–2713. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.LG L, Manual of histologic staining methods of the Armed Forces Institute of Pathology 1968, McGraw-Hill: New York: p. 40–41, 220. [Google Scholar]
- 17.K B, A phenol crystal violet stain for differential staining of Gram positive and negative bacteria Miles Scientific, 1984. VIII, No. 4: p. 216. [Google Scholar]
- 18.RD L, Histologic technic and practical histochemistry 1954, New York: McGraw-Hill. [Google Scholar]
- 19.Gobel TW, Meier EL, and Du Pasquier L, Biochemical analysis of the Xenopus laevis TCR/CD3 complex supports the “stepwise evolution” model. Eur J Immunol, 2000. 30(10): p. 2775–81. [DOI] [PubMed] [Google Scholar]
- 20.Morales HD and Robert J, Characterization of primary and memory CD8 T-cell responses against ranavirus (FV3) in Xenopus laevis. J Virol, 2007. 81(5): p. 2240–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Neely HR and Flajnik MF, Emergence and Evolution of Secondary Lymphoid Organs. Annu Rev Cell Dev Biol, 2016. 32: p. 693–711. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.do you mean Gram staining and ZN staining? Both ZN and FF staining are for acid-fast bacteria but FF is modified for less acid-fast mycobacteria such as M. leprae.
- 23.Edholm ES, et al. , Distinct MHC class I-like interacting invariant T cell lineage at the forefront of mycobacterial immunity uncovered in Xenopus. Proc Natl Acad Sci U S A, 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Chida AS, Goyos A, and Robert J, Phylogenetic and developmental study of CD4, CD8 alpha and beta T cell co-receptor homologs in two amphibian species, Xenopus tropicalis and Xenopus laevis. Dev Comp Immunol, 2011. 35(3): p. 366–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Jurgens JB, et al. , Identification of a candidate CD5 homologue in the amphibian Xenopus laevis. J Immunol, 1995. 155(9): p. 4218–23. [PubMed] [Google Scholar]
- 26.Sasindran SJ and Torrelles JB, Mycobacterium Tuberculosis Infection and Inflammation: what is Beneficial for the Host and for the Bacterium? Front Microbiol, 2011. 2: p. 2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Lin PL and Flynn JL, CD8 T cells and Mycobacterium tuberculosis infection. Semin Immunopathol, 2015. 37(3): p. 239–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.McCarville JL and Ayres JS, Disease tolerance: concept and mechanisms. Curr Opin Immunol, 2018. 50: p. 88–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Schneider DS and Ayres JS, Two ways to survive infection: what resistance and tolerance can teach us about treating infectious diseases. Nat Rev Immunol, 2008. 8(11): p. 889–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Meunier I, et al. , Unravelling the networks dictating host resistance versus tolerance during pulmonary infections. Cell Tissue Res, 2017. 367(3): p. 525–536. [DOI] [PMC free article] [PubMed] [Google Scholar]