Shedding natural light on fungal infections

Abstract

Bioluminescence imaging allows the visualization of the temporal and spatial progression of biological phenomena, in particular infection, by non-invasive methods in vivo. This nature-borrowed technology has been successfully used to monitor bacterial infections but recent studies have also succeeded in tracking fungal infections such as those caused by the two major opportunistic fungal pathogens Candida albicans and Aspergillus fumigatus. The findings of Donat and collaborators published in this issue now show that by combining the sensitivity of the Gaussia princeps luciferase with a surface display expression system it is possible to perform longitudinal infection studies on cutaneous forms of aspergillosis with a small number of animals. Besides providing new and valuable information in the field of aspergillosis, the findings of Donat et al. offer a new perspective on the general applicability of bioluminescence methodologies for eukaryotic pathogens where the bacterial lux operon cannot be exploited.

Keywords: :

• Aspergillus
• bioluminescence imaging
• Candida
• fungal infection
• fungi
• luciferase
• opportunistic infection

Opportunistic fungal infections now represent a substantial threat to the health of populations. Despite remarkable progress in antifungal therapy, late diagnosis and impairment of the immune system maintain a high mortality rate among patients that develop systemic fungal infections such as candidemia and invasive aspergillosis.^1,2 New approaches and interventions to prevent and control these illnesses are clearly needed. Yet, they require in depth knowledge of fungal pathogenicity and host responses to infection. Experimental animal models are a gold standard for the above studies but, unfortunately, they often warrant the use of invasive techniques, both in the “afferent” phase, i.e., the induction of infection, and in the “efferent” phase, i.e., the assessment of infection, together with the sampling of a large number of animals to measure the kinetics of infection and its spread. While in the former phase it is somewhat inevitable, the possibility to otherwise avoid invasive approaches and limit the number of animals tested, particularly when we are dealing with mammals, constitutes a real breakthrough. Bioluminescence imaging is just one case in point: by allowing real time visualization of the temporal and spatial progression of infection,³ it can shed light on infection and its mechanisms. When applied to fungi, bioluminescence imaging provides additional advantages, including the possibility to visualize the infection in its early stages, when traditional means are not adequate, and the possibility to avoid the mistakes incurred by quantification using colony forming unit (CFUs) counts of hyphae growing organisms.^4,5

The use of bioluminescence, i.e., the emission of visible light by living organisms, is a trick that we have learned from nature.⁶ Many organisms, from bacteria and insects to other invertebrates and fish fighting the deep sea darkness, make use of various luminescence colors all based on the chemical oxidation of a luciferin substrate (luciferin) by a luciferase.⁶ Our various methods to exploit this phenomenon and generate bioluminescent organisms as tools to study infection are a genetic “transfer” of the system used in naturally photon emitting organisms.

Making Fungi Bioluminescent

Historically, bioluminescence imaging was first used successfully to monitor bacterial infections, taking advantage of the bacterial lux operons that allow production of a luciferase and its substrate. Bioluminescence imaging of C. albicans and A. fumigatus infections in experimental models is recent and has relied on eukaryotic luciferases and their externally-added substrates. Bioluminescent C. albicans strains have now been generated based on the firefly (Photinus pyralis), sea pansy (Renilla reniformis) and copepod (Gaussia princeps) luciferases. Yet, only firefly luciferase-based and Gaussia luciferase-based bioluminescent C. albicans have been evaluated in animal models of candidiasis. Early attempts to use the firefly luciferase as a reporter in C. albicans were unsuccessful.⁷ The failure was ascribed to the presence of several CUG codons in the firefly luciferase open reading frame, which C. albicans decodes into serine instead of leucine. Hence, a bioluminescent C. albicans strain was generated by replacing CUG codons in the firefly luciferase gene.⁸ This system proved adequate in vitro for both yeast and hyphal cells. Yet, it was critically affected, possibly because of poor permeability of hyphal cells to luciferin, by the yeast-to-hypha switch in C. albicans pathogenicity.^8,9 Nonetheless, bioluminescence was measured in vaginal tissue after vaginal infection in mice and a good correlation between fungal loads determined as CFUs and the bioluminescent signals was detected in animals for over 30 d post infection.⁹ Moreover, the use of bioluminescence to monitor antifungal therapy in models of C. albicans infection was validated. In a recent comprehensive review, Brock suggested that the drawbacks of the current CUG-modified firefly luciferase could be circumvented by synthesis of a fully codon-adapted luciferase gene, thereby enhancing protein production and the related bioluminescence intensity.¹⁰ Brock also suggested that bioluminescence could be additionally improved by preventing peroxisome targeting of the recombinant luciferase. These hypotheses are awaiting evaluation but could prove true given that luciferin, the firefly luciferase substrate, has higher in vivo stability than coelenterazine, the Gaussia luciferase substrate.

Our groups recently developed an alternative system by targeting the Gaussia luciferase to the cell surface of C. albicans. The luciferase reporter was constructed by fusing a synthetic, codon-optimized version of the G. princeps luciferase gene to C. albicans PGA59, which encodes a glycosylphosphatidylinositol-linked cell wall protein. Luciferase expressed from this PGA59-gLUC fusion (referred to as gLUC59) was localized at the C. albicans cell surface.⁴ We were thus able to perform real-time monitoring of cutaneous, subcutaneous and vaginal infection.⁴ Importantly, the bioluminescence photon measurement correlated with yeast and hyphae counts evaluated as CFUs, suggesting that the gLUC59 system provides an excellent tool for monitoring the course of these infections regardless of the form of fungal growth.^4,5 Of note, we also observed that in vivo imaging was more reliable than vaginal CFU counts for assessing the extent and duration of vaginal infection.^5,11 The ability to perform real-time monitoring of vaginal candidiasis in animals is particularly important since adequate eradicating therapies for chronic recurrent vaginal candidiasis, a very widespread disease in women, are lacking, and approaches using anti-Candida vaccines and immunotherapy are being developed and evaluated.¹² In line with this, using a gLUC59-expressing C. albicans strain we could monitor the efficacy of a β-glucan-conjugate vaccine formulated with the human-compatible MF59 adjuvant in a murine model of vaginal candidiasis.⁵ In addition this model proved useful to test new therapeutic compounds¹³ and the beneficial role of cytokines such as Il-17.¹¹ More recently, the gLUC59 system has been used to investigate by real-time monitoring the potential of photodynamic therapy for prophylaxis and treatment of cutaneous C. albicans infections in mice.^14,15

Bioluminescent A. fumigatus

Bioluminescence imaging of the filamentous fungus A. fumigatus was first performed by Brock et al.¹⁵ They expressed the firefly luciferase gene from the strong and constitutive promoter of the A. fumigatus gpdA gene encoding glyceraldehyde-3-phosphate-dehydrogenase. Effective light emission from filamentous cells indicated that luciferin reached the intracellular compartment of hyphae. Interestingly, this system also provided a useful tool for screening the activity of antifungal compounds. When the in vitro system was translated into an invasive bronchopulmonary aspergillosis, strong bioluminescence from lungs 24 h after infection was observed in immunosuppressed live animals, providing the first demonstration that fungal infection in deep tissue can be monitored. A major limitation of this system was a strong and rapid decrease of light emission 24 h post infection, despite the high fungal load. Given that all luciferases require dissolved oxygen for oxidation of their substrate, it was speculated that the rapid and strong inflammatory and necrotic process seen in the lungs might provide an anaerobic microenvironment that compromises light production from luciferin. A subsequent study by Ibrahim-Granet et al. has brought credit to this hypothesis.¹⁶ Indeed, these authors have shown that the limitation in fungal quantification was overcome in leukopenic mice displaying numeric and functional reduction of phagocytes. Hypoxia was shown to be a major factor causing the decline of the bioluminescent signal which can be used as an indicator of the imminent death of the animals. By performing bioluminescence monitoring of invasive aspergillosis in a host undergoing myeloablation with cyclophosphamide it should be possible to assess the efficacy of some therapeutic compounds.¹⁶

The work by Donat et al. provides an alternative approach for monitoring A. fumigatus infections using a genetically-engineered A. fumigatus strain that expresses a cell-surface exposed Gaussia luciferase.¹⁷ This was achieved in a manner similar to that used in C. albicans with gLUC59, i.e., through addition of secretory and GPI-anchoring signals to the Gaussia luciferase. In this study, the authors additionally developed a relevant model of cutaneous aspergillosis and showed that the surface display of the Gaussia luciferase allowed spatial and temporal monitoring of the infection using bioluminescence imaging, a significant further improvement on the recent experimental model described by Ben-Ami et al.¹⁸ The advantages include dynamic imaging which reduces the number of animals required, as well as the detection of a minimal number of 1,000 infectious conidia. Moreover, the effect of antifungal therapy could be monitored by real-time imaging in this model. Without detracting from the importance of cutaneous aspergillosis, which is the second most common type, invasive pulmonary aspergillosis remains the most important and serious type of infection. Difficulties in monitoring the latter were highlighted because the low sensitivity of the system did not allow for detection of infection in the lung even though the substrate was administered intravenously. The authors suggest that the rapid decline in photon emission by the Gaussia luciferase coupled with the instability of the substrate does not allow for a reliable model of pulmonary aspergillosis. Indeed, it is important to note that in order to achieve a clear and well defined bioluminescent signal in the cutaneous aspergillosis model, the coelenterazine substrate had to be injected at the site of infection, reflecting problems due to the difficulty of distribution and rapid degradation of this substrate. Similar limitations were noticed when using engineered C. albicans expressing gLUC59.¹⁹

Caveats and Drawbacks

It should be noted that all attempts to monitor systemic or deep-seated C. albicans infection in live animals using bioluminescence imaging have been relatively unsuccessful until now. While this could be due to the limited distribution of the luciferase substrate, we also tested multiple doses of coelenterazine, and various routes including intravenous injection, without success.¹⁹ As discussed by Brock,¹⁰ a variety of alternative explanations can be offered for these failures, including absorption by hemoglobin of light emitted from internally located deep-seated animal organs, as well as still non optimized codon usage in the case of the firefly luciferase. In the case of experimental aspergillosis, the intensity and stability of photons emitted by internal organs, including lungs, remain issues of concern. Even accounting for the significant differences between the results obtained with C. albicans and A. fumigatus, it appears overall that the major limitations in bioluminescence imaging of fungal infections have been experienced in animal models used to study deep-seated infections. Since these infections are of major medical concern, increased efforts should be made to address and possibly resolve the underlying issues. Bearing this in mind, we can be satisfied momentarily with the successes in bioluminescent monitoring of superficial and subcutaneous infections, and recurrent vaginal infections, which are devastating the quality of life of millions of women worldwide.