We analyzed samples from patients diagnosed with AD, ALS, and Parkinson's disease (PD) in addition to control individuals without neurological disease. The age and gender of the subjects investigated in this study are listed in Supplementary Table 1. All samples were supplied by a brain bank (Banco de Tejidos CIEN) and were analyzed anonymously. The Ethics Committee of the Universidad Autónoma de Madrid approved the study. The transfer of samples was carried out according to national regulations concerning research on human biological samples. In all cases, written informed consent was obtained. For patients with dementia, informed consent for brain donation was given on a postmortem basis by a next-of-kin following the procedure established by the external ethical committee of the brain bank. Accordingly, a next-of-kin of the patient gave credit through an informed consent document to the fact that the patient had never opposed to be a brain donor during his/her life. An ethics committee external to the bank approved all ethico-legal documents, including written informed consent.

Brain samples were processed according to a common postmortem protocol followed by Banco de Tejidos CIEN. Briefly, rapid neuropathological autopsy was performed upon call by the donor's proxies (mean postmortem interval, 4.5 h). Immediately after extraction, the right half of the brain was sliced and frozen, while the left half was fixed by immersion in phosphate-buffered 4% formaldehyde for at least 3 weeks. A full neuropathological study was performed in the left half brain after fixation. Neuropathological diagnosis and staging of all disease entities was performed according to consensus criteria. Various neuropathological variables related to AD, vascular, Lewy and TDP (TAR DNA-binding protein) pathologies in addition to the presence of hippocampal sclerosis were recorded for full classification of cases.

Candida famata, C. albicans, C. glabrata, and Syncephalastrun racemosum were grown in YEP (yeast extract peptone) medium (1% yeast extract, 2% Bacto peptone) as described (Pisa et al., 2008). Fungal cells were centrifuged and washed in phosphate-buffered saline (PBS). Phoma betae was purchased from Allergon AB (Engelholm, Sweden). Fungal cells were autoclaved and lyophilized. Rabbit antisera against C. famata, C. albicans, C. glabrata, P. betae, and S. racemosum were obtained by inoculation of 1 or 2 mg of dried fungi in 0.5 ml PBS, previously mixed with an equal volume of Freund's adjuvant. Rabbits were inoculated up to three times every 3 weeks and the antibody titer and specificity of the sera were tested by immunohistochemistry and immunoblotting using fungal proteins. The protocols employed were approved by the ethics committee of Centro de Biologia Molecular “Severo Ochoa” (identification number: ES280790000180). The optimal dilution for immunofluorescence staining for each antibody was assayed using both isolated Candida spp.

The specificity of the antifungal antibodies obtained was tested by immunofluorescence against different Candida spp. The cross-reactivity of each antibody against the different fungal species can differ, for instance anti-C. glabrata antibody immunoreacted with C. glabrata, C. albicans, C. tropicalis, C. parapsilosis, and C. Krusei, whereas anti-C. albicans does not recognize C. Parapsilosis and C. Krusei. By contrast, anti-S. racemosum only immunoreacts with C. Krusei. Besides, none of the antifungal antibodies obtained immunoreacted with cultured human cells or human brain sections from healthy subjects (Pacheco et al., 2007; Pisa et al., 2015a,b).

CNS tissue was embedded in paraffin following standard techniques and cut into 5-μ m sections using a microtome (Microm HM355s, Walldorf, Germany). For immunohistochemical analysis, paraffin was removed and sections were rehydrated and boiled for 2 min in 10 mM citrate buffer and then incubated for 10 min with 50 mM ammonium chloride. Subsequently, tissue sections were incubated for 10 min with PBS/Triton X-100 (0.1%) followed by 20 min with 2% bovine serum albumin in PBS. Sections were incubated overnight at 4°C with a mouse monoclonal antibody raised against human α-tubulin (Sigma), human phospho-PHF-tau, clone AT100 (Thermo Scientific), or human neurofilament protein, clone 2F11 (Dako), all at 1:50 dilution, or a rabbit polyclonal antibody raised against proteins obtained from C. glabrata at 1:500 or C. famata, C. albicans, P. betae, and S. racemosum at 1:100 dilution. Thereafter, sections were washed with PBS and further incubated for 1 h at 37°C with donkey anti-mouse IgG secondary antibody conjugated to Alexa 555 (Invitrogen) at 1:500 for α-tubulin, tau and neurofilament, and donkey anti-rabbit IgG secondary antibody conjugated to Alexa 488 (Invitrogen) at 1:500 dilution for antifungal antibodies. To visualize nuclei, sections were then stained with DAPI (4,6- diamino-2-fenilindol) (Merck) and treated with autofluorescence eliminator reagent (Merck). The use of this reagent is important to avoid autofluorescence, since there is lipofuschin in the aging brain. All images were collected and analyzed with a LSM710 confocal laser scanning microscope combined with the upright microscope stand AxioImager.M2 (Zeiss), running Zeiss ZEN 2010 software. The spectral system employed was Quasar + 2 PMTs. Images were deconvoluted using Huygens software (4.2.2 p0) and visualized with Fiji/ImageJ (NIH, Bethesda, MD) software.

A variety of cellular proteins constitute part of CA (Sfanos et al., 2009). To assess the presence of fungal proteins in CA, we carried out immunohistochemistry analysis using a specific rabbit polyclonal antibody raised against C. glabrata, which does not cross-react with human proteins (Pisa et al., 2015a,b). Initially, we tested this antibody on tissue sections from different CNS regions, including lateral frontal cortex (LFC), cerebellar córtex (CEC) and entorhinal cortex/hippocampus (ERH) from one AD patient. Double immunofluorescence staining was performed using a second antibody that detects tau protein. CA were more abundant in ERH regions than in LFC or CEC (Figure 1). CA from all three CNS regions (LFC, CEC, and ERH) stained positive with the anti-fungal antibody (Figure 1). The external laminar structures and the envelope surrounding the central core of CA were clearly immunoreactive for the C. glabrata antibody (green), revealing the presence of fungal proteins in this region (Figure 1). In some instances, the entire envelope was positive, while in other sections only a part of the external envelope was immunoreactive. By contrast, the anti-tau antibody (red) stained only some of the CA, particularly from CEC sections (Figure 1). This result is consistent with previous findings describing tau protein in CA (Singhrao et al., 1993; Day et al., 2015). Of note, not all CA contained tau protein inasmuch as it was undetectable in some CA sections analyzed by confocal microscopy. Thus, although tau can be detected in some CA, it is not an abundant protein.

To further assess the presence of fungal proteins in CA, tissue sections were immunolabeled with additional antibodies raised against C. famata, C. albicans. S. racemosum, and P. betae (green) together with a monoclonal antibody against neurofilaments (red). As shown in Figure 2, CA inclusions from the different CNS regions from AD patient 1 immunoreacted with all four antifungal antibodies, further supporting the existence of fungal proteins in CA. None of the antifungal antibodies recognize cellular proteins from neural cells in CNS sections (Pisa et al., 2015a,b). The localization of the immunopositive structures using the additional antibodies was similar to that observed with the C. glabrata antibody, strengthening the notion that fungal proteins are present in CA from different CNS regions. By contrast, neurofilament staining was more irregular, with strong immunoreactivity in some CA, and weaker or no staining in other CA inclusions. This was evident in CEC sections double-labeled with the anti-C. albicans antibody and in LFC and CEC sections labeled with the anti-S. racemosum antibody; in both cases, the neurofilament signal was more intense than the fungal signal, as manifested by the punctate yellow staining in merged images in some of the sections. The fact that neurofilaments are detected in some CA but not in others underscores the concept that the protein composition of CNS CA is nonhomogeneous. This lack of homogeneity can be revealed only by immunohistochemistry and not by proteomic analyses of purified CA.

We also analyzed ERH sections from other AD patients (AD2-AD11; Supplementary Table 1). Immunohistochemmistry was performed using anti-C. albicans and anti-P. betae antibodies (green) and anti-human α-tubulin antibodies were used to mark microtuble structures (red). Of note, the anti-tubulin antibody immunoreacts not only with human cells but also with several eukaryotic species, including fungal cells. Consequently, in the instances where both green and red signals co-localized, it may be because of the presence of fungal tubulin. Fungal proteins were detected in ERH CA inclusions from all 10 additional AD patients (Figure 3). As the antifungal antibodies are polyclonal, they can cross-react with a number of fungal proteins. The positive immunoreactivity with one of these antibodies does not demonstrate that the fungal species present is the same as that employed to raise the antibody. However, since each antibody immunoreacts with different fungal antigens, differences in the immunostaining provides a clue to indicate that the fungal species differ. Accordingly, this technique cannot establish the precise fungal species present in each sample and DNA sequencing would be required (Alonso et al., 2014a,b; Pisa et al., 2015b). The majority of the CA inclusions from different patients immunoreacted with both anti-C. albicans and anti-P. betae antibodies, although staining for P. betae was more robust in patients AD3, AD8, AD10, and AD11 than in the other patients (Figure 3). This finding suggests that the fungal species present in each patient differ. Additionally these results support the use of a panel of anti-fungal antibodies to comprehensively determine the presence or absence of fungal proteins in CA. An important conclusion from this analysis is that the location of fungal proteins in CA inclusions rules out the possibility that fungal infection was due to postmortem contamination since the formation of CA occurs over long time periods (months or even years).

Our recent finding of fungal infection in CNS from ALS patients (Alonso et al., 2015), prompted us to test for fungi in CA inclusions in ALS samples. We examined tissue sections from motor cortex (MC), medulla (MD) and different levels of the spinal cord (SC1, SC2, and SC3) of an ALS patient (ALS1) using anti-C. albicans and anti-P. betae antibodies. Double immunolabeling of the CNS from patient ALS1 with α-tubulin (red) revealed fungal proteins (green) at the periphery of CA inclusions in different regions (Figure 4). Of note, CA inclusions were also detected in different regions of the CNS of this patient, including the spinal cord. The numbers of CA inclusions observed in these samples were, however, lower than those found in AD patients. We also tested tissue sections from different CNS regions of five additional ALS patients (ALS2-ALS6) using the same antibodies. CA inclusions were detected in all ALS patients examined and all were positive for fungal protein. For clarity, only one field with each antibody and only one CNS region for each ALS patient is shown (Figure 4). In general, fungal proteins (green) were detected at the periphery of CA inclusions, but in some instances fungal proteins were observed throughout CA bodies, including the central portion. Additionally, anti-α-tubulin (red), reactivity was in the main detected in association with material present in CA, but in a few instances α-tubulin was detected in the surrounding areas. The observations in ALS tissue indicate that there is a similarity between ALS CA and AD CA (Figures 1, 2), and also further support the concept of CA protein heterogeneity.

We also tested for fungal proteins in CA from brain samples of one PD patient. As before, several CNS regions were analyzed using antifungal antibodies (green) and anti-α-tubulin antibodies (red) (Figure 5). The CNS regions examined in this PD patient included pons (PN), mesencephalon (MSP), hypothalamus (HT), callosal body (CB), and caudate and lenticular nuclei (CLN). CA were also detected in this PD patient and were found in all CNS regions analyzed. Furthermore, immunoreactive fungal proteins were prominent in CA bodies and the distribution of the immunolabels was similar to those observed in ALS and AD patients. Similarly to ALS patients, the number of CA inclusions in the different regions examined in the PD patient was much less than in AD patients. Analysis of CNS samples from five additional PD patients using anti-C. albicans and anti-P. betae antibodies is shown in Figure 5. Once again, CA inclusions were detected in all PD patients analyzed and immunoreactivity to fungal proteins in these CA was revealed with the two antifungal antibodies employed. The location of fungal proteins (green) and the distribution of human α-tubulin were again similar to the pattern observed with CA from ALS patients.

It is thought that the formation of CA inclusions are related to the aging process, even in healthy subjects (Song et al., 2014). We examined ERH sections from the CNS of five control subjects using the two antifungal antibodies indicated above. In general, CA inclusions were clearly much less abundant than those observed in CNS tissue from patients diagnosed with neurodegenerative diseases. Nevertheless, the CA inclusions in control subjects exhibited a modest immunoreactivity against the two antifungal antibodies employed. For example, CA from control subjects C2 and C3 immunoreacted with the anti-C. albicans antibody, while C4 and, to a lesser extent, C3 immunoreacted with the anti-P. betae antibody (Figure 6). Similar to CA from neurodegenerative patients, the external perimeter of CA exhibited punctate immunoreactivity when detected; however, the labeling intensity was lower than that observed in AD patients. These findings indicate that in general the amount of fungal proteins in CA from control individuals is much lower than from equivalent neurodegenerative disease patients and in some cases no immunoreactivity is detected. The absence of fungal proteins in some CA might be determined by the particular section examined. Nevertheless, the amount of fungal proteins in control CA inclusions is very low. Alternatively, it is theoretically possible that the fungal proteins, if present, cannot be detected with the antibodies employed in this study. An estimation of the number and positiveness of CA in different patients and control subjects is shown in Figure 7. Certainly, the vast majority of CA from the different patients analyzed is stained with antifungal antibodies, whereas only a very minor portion of CA can be considered as positive in control subjects. On the other hand, in general this quantitation reveals higher numbers of CA in AD patients, particularly from ERC areas where CA are more abundant both in patients and controls. However, one limitation of this quantitation is that these numbers can vary depending on the specific tissue section analyzed. As previously noted by other researchers, the amount of CA is higher close to blood vessels (Nishio et al., 2001).

Collectively, these findings reveal the presence of CA inclusions in several regions of the CNS from patients with neurodegenerative diseases. CA are more abundant in the ERH of AD patients than in other regions but there are higher quantities of CA from ALS and PD samples than in controls. Importantly, fungal proteins are detected in CA from all patients with neurodegenerative diseases tested by means of specific antibodies.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.