Detection of CandidaalbicansADH1 and ADH2 mRNAs in human archival oral biopsy samples

Bakri MM1, Cannon RD2, Holmes AR2, Rich AM2.

1Department of Oral Biology and Biomedical Sciences, Faculty of Dentistry, University of Malaya, Kuala Lumpur, Malaysia (email:); 2Sir John Walsh Research Institute, Faculty of Dentistry, University of Otago, Dunedin, New Zealand (emails:richard.cannon@otago​.ac.nz, .​nz, ​.nz)

Keywords: alcohol dehydrogenase, acetaldehyde, chronic hyperplastic candidosis, Candida albicans, archival biopsy, formalin-fixedparaffin-embedded.

Running title: Candida albicans ADH mRNAs in human oral biopsy samples

Corresponding Author:

Assoc. Prof. Dr. Marina M Bakri

Department of Oral Biology,

Faculty of Dentistry,

University of Malaya, Kuala Lumpur,

50603 Malaysia

email:

tel:+6013-6300716 (Mobile)

+603-79677451/7414 (Office)

fax:+603-79674536

Submission date: 12.9.2013

Abstract

Objectives:The aim of this study was to investigate the relationship between expression of C. albicans Alcohol Dehydrogenases (ADH) genes in archival formalin-fixedparaffin-embedded (FFPE) samples from biopsies of leukoplakia.

Materials and Methods: ArchivalFFPE samples were obtained from four sample groups: normal oral mucosa, non-dysplastic leukoplakia, chronic hyperplastic candidosis (CHC) and non-CHC dysplastic leukoplakia. The presence of C. albicans was determined by periodic acid Schiff staining and by immunocytochemistry.C. albicansADH1 and ADH2 mRNAs were detected using reverse transcription PCR.

Results: C. albicans was detected in FFPE samplesdiagnosed as CHC(the histological diagnoses had been made by specialist oral pathologists, using uniform criteria), but not in any other sample group, including the non-dysplastic leukoplakias. RT-PCR confirmed

a significant correlation between the expression of CaADH1mRNA (P = 0.000) but not for CaADH2mRNA (P= 0.056) in archival FFPE samples (n=31) from biopsies of leukoplakia.

Conclusions: C. albicans was the predominant species in the lesions diagnosed as CHC and the presence of C. albicans in CHC lesions was associated witha high expression of C. albicans ADH1 mRNA.There was no association between the presence of Candida and malignant transformation in the cases examined, however the number of cases was limited and further studies are needed to further elucidate the roleof C. albicans ADH1 in the pathogenesis of oral squamous cell carcinoma.

Introduction

Since the 1960sCandida albicans has been implicated in the progression of some leukoplakias, the most common of the potentially malignant oral disorders (PMOD), to oral squamous cell carcinoma (OSCC).1,2The attributes of Candida that may contribute to cancer progression have been extensively reviewed3,4but the role of Candida in promoting oral cancer has always been controversial. Major risk factors definitely associated with oral cancer are tobacco smoking and alcohol consumption. Alcohol itself is notcarcinogenic but it is the metabolic product, acetaldehyde,that is linked to carcinogenesis.6-10The enzyme isoforms responsible for the reversible production of acetaldehyde from ethanol are alcohol dehydrogenases encoded by the ADH genes. In C. albicans, seven ADH genes have been annotated in the C. albicans genome database ( by homology to S. cerevisiae genes but little is known about the functions of the alcohol dehydrogenase enzymes in the acetaldehyde metabolismof C. albicans. The expression of CaADH1 and CaADH2 were investigated in this study as they have the highest amino acid sequence homology to S. cerevisiae genes involved in acetaldehyde metabolism; ScADH1- the major contributor to ethanol production5and ScADH2- responsible for acetaldehyde production.6,7In this study, the presence of Candida and the expression of particular C. albicansADHgenes in archival biopsies from human oral leukoplakic lesions wasstudied to investigate whether the presence of Candida, or ADH gene activity, could be related to histopathological diagnosis. Histological examination of a lesion with a clinical diagnosis of homogeneous leukoplakiararely shows infection withCandida8. If Candida was detected, the lesion by definition is not leukoplakia and an alternative clinical and histological diagnosis is chronic hyperplastic candidosis (CHC). CHC has the potential to progress to OSCC.9-11

Despite the potential difficulty of obtaining nucleic acids of adequate quality and quantity from FFPE samples, RT-PCR has been successfully used in our laboratory to detect C.albicans in human FFPE specimens.12The aim of this study was to detect the expression of C. albicans ADH genes in archival FFPE samples from biopsies of leukoplakia.In order to optimize RT-PCR conditions for the detection of C. albicansADH genes from human FFPE samples, FFPE samples of rat tissues were initially used to develop the method for detecting C. albicans ADH gene expression. This methodology was then applied to documented human biopsy samples.

Materials and methods

Thework with human biopsy samples was approved by the University of Otago Human Ethics Committee (reference OTA 02/11/115).The work with rat tongueshad been approved by the University of Otago Animal Ethics Committee (reference 87/06).

Tissue samples

The human tissue samples were archival FFPE biopsies collected between 2004 and 2007 bythe University of Otago Oral Pathology Diagnostic Service (Medlab Dental). There were four groups of samples;onewas a normal oral mucosal control group (Group A) andthree groups comprised lesions clinically diagnosed as leukoplakia with various histological diagnosesas summarised in Table 1.The histological diagnoses had been made by specialist oral pathologists, using uniform criteria (WHO 1997).13 Group A consisted of three patientswith no evidence of oral pathology whohad had tissue removed as part of an unrelated procedure, such as exposure of an unerupted tooth. There was no evidence of Candida in their H & E and PAS stained sections. Group B comprised eight biopsy specimens from eight patients clinically diagnosed with leukoplakia where the biopsy showed epithelial hyperplasia without dysplasia and with no obvious clinical or histological Candida infection. Group C comprisedten biopsy specimens from seven patients clinically diagnosed with leukoplakia where the biopsy showed epithelial hyperplasia with moderate or severe dysplasia and with no obvious clinical or histological Candida infection. Group D consisted oftenbiopsies from ten patients diagnosed histologically with CHC;6 of these were not dysplastic and the remainder showed mild dysplasia.Sections were cut from each block and used for routine PAS staining following the usual laboratory protocol, immunocytochemistry and RT-PCR.

The samples had been fixed in formalin for variable periods of time to allow transportation to the laboratory from distant regions (approx. 5-7 days). To check whether or not viable RNA could be obtained from such samples, rat tongues remaining from a previous project investigating oral colonisation by C. albicans ATCC 10231that had been fixed in 10% neutral buffered formalin for 7 days before being embedded in paraffin wax were retrieved.

Processing of samples for DNA and RNA extraction

RNA was extracted from rat and human FFPE samples. Six 10 m sections were deparaffinized in xylene and washed with 100% ethanol and air-dried before incubation in 1 ml of SK buffer (1 M sorbitol, 50 mM KH2PO4 [pH8.0]) containing 0.5 mg of zymolyase and 0.1% (v/v) 2–mercaptoethanol at 37°C for 1 h to digest C. albicans cell walls. Sample material was then pelleted at 16 000 x g for 5 min at room temperature and as much supernatant as possible was carefully discarded. RNA was then extracted from the pellet using a PureLink TM FFPE RNA isolation kit (Invitrogen, Carlsbad, USA), following the manufacturer’s instructions. The extracted RNA (10 μg) was treated with DNase I (Ambion, Texas, USA) to eliminate residual traces of DNA in the RNA samples.

Immunocytochemistry

Three micron sections obtained from FFPE tissue blocks were processed routinely and washed with 3% hydrogen peroxide in phosphate buffered saline (PBS), pH 7.2,to block non-specific peroxidase binding sites and overlaid with 5% fetal calf serum to block non-specific protein binding sites. The 1°antibody was a C. albicans-specific rabbit antibodyraised against whole heat-killed C. albicans ATCC 10261 cells (Dr. A. R. Holmes, Otago University). Bound antibody was detected with a peroxidase-labelled anti-rabbit2° antibody and visualized with Dako LSAB2System-HRP.

RT-PCR

Thermal cycling for reverse-transcriptase (RT) and PCR reactions were performed using a Mastercycler-Gradient instrument (Eppendorf AG, Hamburg, Germany). RT reactions were undertakenusing Superscript III enzyme (Invitrogen) according to the manufacturer’s instructions. RNase inhibitor was included in the RT reaction and, in addition, the reaction mix was kept at 55°C following the initial template denaturation-primer annealing step, to reduce the potential for mispriming. Specific reverse primers were used to generate cDNA templates in the RT reaction. These primers were also used for subsequent PCR amplification and the primers used are listed in Table 2. The RT negative controls were prepared without the Superscript III enzyme.PCRs were undertaken using Taq polymerase (HotStarTaq; Qiagen, Valencia, Calif.) and the cycling conditions consisted of an initial enzyme activation step of 95°C for 15 min, followed by 30 cycles of denaturation at 94°C for 15 s, annealing for 15 s and extension at 72°C for 6 s. Amplification of the Ca18S rRNA gene from C. albicans genomic DNA was also carried out as a PCR positive control and the reaction mixes without template were used as PCR negative controls.

Statistical analysis

A Pearson product-moment correlation coefficient was computed to assess the relationship between the expression of C. albicans ADHgenes in the biopsies of leukoplakia. The correlation was considered significant (P=0.000) at the 0.01 level (2-tailed).

Results

Detection of C. albicans in FFPE samples

PAS staining and immunocytochemistry

C. albicansyeasts and hyphae in the rat and human FFPE sections stainedmagenta with PAS (Figure 1A and 1C) and brown with immunocytochemistry (Figure 1B and 1D).Candida cells were only detected in the surface keratin layer and did not penetrate deeper into the epithelium. The results for the human biopsy specimens are shown in Table 3. As expected, none of the Group A or B samples tested positive for Candida.

In Group C, the samples that were initially diagnosed as leukoplakia with histological dysplasia but without Candida, produced different results when re-examined for this study. Two out of 10 FFPE biopsy samples (C1 and C2) stained positively for Candida, bothby PAS staining and immunocytochemistry. Out of ten samples in Group D, Candida was detected in nine samples by PAS staining and in six samples by immunocytochemistry. If both the PAS staining and immunocytochemistry results were taken into consideration, then all samples were considered positive for Candida by either PAS staining or immunocytochemistry and none of the samples were negative by both the methods. A summary of the results from all patients is shown in Table 4.

RT-PCR

RT-PCR was initially performed on rat tissue for optimisation of the technique and Ca18S rRNA was chosen as the initial target because the gene is present in each cell in multiple copies and is highly transcribed.Ca18S rRNA was successfully detected and following this, both CaADH1 and CaADH2 mRNAs were detected in the rat FFPE samples but there was less CaADH2 mRNA than CaADH1 mRNA (Figure 2).

Having established that it was possible to detect C.albicans mRNAs in rat FFPE samples, the method was then applied to archival human FFPE samples. As well as the positive and negative PCR controls, additional negative controls were carried out which included RT-PCR of human FFPE samples from Group A which were highly unlikely to have Candida cells present. Detection of C. albicans mRNAs was initially carried out using FFPE samples from Group D (CHC patients) asthe presence of Candida in these samples had been confirmedprior to this study. RT-PCR indicated that Ca18S rRNA and CaADH1 mRNAs were present in nine out of ten samples (Figure3). It was observed that the one sample that tested negative for Ca18s rRNA and CaADH1 by RT-PCR (D8), tested positive for PAS staining but negative with immunocytochemistry. CaADH2 mRNA was detected in three out of ten samples (Figure 3; D2, D5 and D7). D5 and D7 were both positive by PAS staining and immunocytochemistry but D2 tested positive only by PAS staining.

Following this RT-PCR optimization, detection of Ca18S rRNA, CaADH1 and CaADH2 mRNAs was undertaken using the remaining FFPE samples. As expected, C. albicans mRNAs were not detected in Group A(normal oral mucosa) or Group B samples (leukoplakia without dysplasia). Group C (leukoplakia with dysplasia and ostensibly without Candida) produced apparently anomalous results when two of the ten samples (C1 and C2), expressedCa18S rRNA and CaADH1;C2also expressedCaADH2 (Figure4). When the RT-PCR results for Group C were compared to the histology and immunocytochemistry results C1 and C2were retrospectively found to be positive for C. albicans by both histology and immunocytochemistry methods (Table 3).

Statistical analysis showed a positive correlation between the expression of CaADH1 mRNA (r = 0.720) and CaADH2 mRNAs (r = 0.347) in the leukoplakia biopsies. However, the correlation is significant at the 0.01 level (2-tailed) only for the expression of CaADH1 mRNA (P = 0.000) and not for CaADH2 mRNA (P = 0.056).

Clinical outcomes

Clinicians managing the patients in Groups A-D were contacted and asked to provide information obtained from follow-up visits. This information is presented in Table 5.

The lesions in all patients with leukoplakia with no histological evidence of dysplasia (Group B), about whom follow-up information was available, resolved and did not recur. There were 10 lesions from seven patients included in the leukoplakia with dysplasia without Candida category (Group C). Follow-up information was obtained for six of the seven patients. Three of the six had no recurrence with a minimum follow-up period of three years. In this group was a patient (C2) who had an incisional biopsy diagnosed as epithelial hyperplasia with moderate dysplasia. PAS stained slides at that time, and in a retrospective review, did not show Candida cells present. The patient was referred by the treating clinician for laser removal of the dysplastic lesion. The surgical specimen was diagnosed at another laboratory as CHC. In light of this, and the detection of Candida in the investigation of this case for the current study, the likely correct diagnosis was CHC. One patient had two areas biopsied (C7 and C8) due to recurrence with increased dysplasia six months after the initial biopsies. There was a further recurrence three years later which showed mild dysplasia. The patient is under regular review. One patient (C5) had lesions which progressed to squamous cell carcinoma.

For the ten patients with CHC (group D), follow-up information was obtained for eight. One had died of unrelated causes. In four patients the lesion resolved with no recurrence. Three had recurrent lesions, which were biopsied. All subsequent biopsies showed CHC with no dysplasia. No patients in this group progressed to further dysplasia or OSCC.

Discussion

Prolonged immersion in formalin decreases the ability to retrieve usable RNA 14 and so rat FFPE tongues, known to be colonized by C albicans and that had been fixed in formalin for seven days were used to confirm that Ca18S rRNA, CaADH1 and CaADH2 mRNA could be detected by RT-PCR. This was successful and then RT-PCR was applied to the human archival FFPE samples, beginning withGroup D samples (CHC)since these were known to contain C. albicans cells. All ten Group D samples tested positive for Candida with either the PAS or the immunocytochemistry technique. While the PAS staining could not differentiate C. albicans from other Candida species, the immunocytochemistry technique provided information on the Candida species present in the FFPE samples as the method utilizes an antibody which was raised specifically against C. albicans and the technique would not detect other yeast species. The results suggest that C. albicans is the major species present in CHC, in line with other published studies,2,4,15 although only culture of un-fixed specimens, or PCR with species-specific primers would completely confirm this.

When molecular technique involving RT-PCR was applied to the FFPE samples from Group D, there was a high expression of the Candida mRNAs as bothCa18S rRNA and CaADH1 mRNAs were detectable in nine out of the ten samples.It is interesting to note that the sample (D8) that did not test positive for Ca18S rRNA or CaADH1 by RT-PCR also tested negative for C. albicans by immunocytochemistry, but positive with PAS staining. The explanation for this could be that C. albicans was not present in that sample but some other Candida species associated with CHC was present instead, such as C. tropicalis, C. pintolopesiior C. glabrata.16As for CaADH2 mRNA detection, a method used by Schofield et al., (2003) was employed. In order to increase the sensitivity of mRNA detection, the number of PCR cycles was also increased from 30 to 45 cycles and the amount of template (CaADH2 cDNA)used for PCR amplification was doubled (to twice that used for detection of Ca18S rRNA and CaADH1 mRNA). Under this condition, only 3 out of the 10 samples tested positive, suggesting differential expression for CaADH2 and similar results had also been reported by Beggs et al., (2003).

Having validated that RT-PCR could be used to detect C. albicans gene expression in Group D, the presence of C. albicans in Group C was investigated using RT-PCR. The Group C FFPE samples had been diagnosed as being dysplastic, but without obvious Candida infection on the basis of histological examination of H&E and PAS stained slides by the Oral Pathology Diagnostic Serviceprior to their inclusion in this study. However, when the Group C samples were re-examined in this study, the results obtained were different to those obtained for diagnostic purposes. Two of the ten samples tested positive with both PAS staining and immunocytochemistry. This discrepancy may have been because the later sections, which would have been deeper in the block, revealed Candida not previously present in the plane of section, or it may have been due to more carefully focused observation on the part of the examiner, although sections were examined without knowledge of the diagnosis group of the specimen. RT-PCR examination of Group C samples revealed that of the ten samples, the two that tested positive for PAS and immunocytochemistry were also positive for Ca18S rRNA and CaADH1 mRNA. This would suggest that the Candida species present in the samples wasC. albicans. As for CaADH2, it was only detected in one of the two samples that tested positive for Ca18S rRNA and CaADH1 mRNA, again suggesting possible differential expression of CaADH2.