Selection of Pneumocystis jirovecii Inosine 5′-Monophosphate Dehydrogenase Mutants in Solid Organ Transplant Recipients: Implication of Mycophenolic Acid

Mycophenolic acid (MPA) targets the inosine 5′-monophosphate dehydrogenase (IMPDH) of human lymphocytes. It is widely used as an immunosuppressant to prevent rejection in solid organ transplant (SOT) recipients who, incidentally, are at risk for Pneumocystis pneumonia (PCP). We hypothesized that MPA exerts selective pressure on P. jirovecii microorganisms considering its in vitro antifungal activity on other fungi. Thus, we analysed impdh gene in P. jirovecii isolates from SOT recipients. P. jirovecii specimens from 26 patients diagnosed with PCP from 2010 to 2020 were retrospectively examined: 10 SOT recipients treated with MPA and 16 non-SOT patients without prior exposure to MPA. The P. jirovecii impdh gene was amplified and sequenced. Nucleotide sequences were aligned with the reference sequences retrieved from available P. jirovecii whole genomes. The deduced IMPDH protein sequences were aligned with available IMPDH proteins from Pneumocystis spp. and other fungal species known to be in vitro sensitive or resistant to MPA. A total of nine SNPs was identified. One SNP (G1020A) that results in an Ala261Thr substitution was identified in all SOT recipients and in none of the non-SOT patients. Considering that IMPDHs of other fungi, resistant to MPA, harbour Thr (or Ser) at the analogous position, the Ala261Thr mutation observed in MPA-treated patients was considered to represent the signature of P. jirovecii exposure to MPA. These results suggest that MPA may be involved in the selection of specific P. jirovecii strains that circulate in the SOT recipient population.


Introduction
Pneumocystis jirovecii (P. jirovecii) is a transmissible, uncultivable fungus responsible for severe pneumonia in immunosuppressed patients. Pneumocystis pneumonia (PCP) remains the most frequent AIDS defining infection in HIV-infected patients in developed countries. In France, PCP represents 30% of the annually reported 1200 AIDS cases [1]. PCP also affects immunosuppressed patients not infected with HIV [2,3]. PCP occurs in 5-15% of solid organ transplant (SOT) recipients in the absence of prophylaxis, with a mortality rate up to 49% [4,5]. The frequency of PCP seems to increase in SOT recipients, specifically in renal transplant recipients (RTR), particularly when the transplantation is not recent, when the patients are no longer subjected to PCP prophylaxis and when the acquisition of P. jirovecii is nosocomial [6].
Mycophenolic acid (MPA) is a widely used anti-rejection treatment included in immunosuppressive regimens for SOT recipients. It targets the inosine 5 -monophosphate dehydrogenase (IMPDH), an enzyme involved in de novo purine nucleotide biosynthesis. MPA specifically inhibits the proliferation of B and T lymphocytes, since these cells synthesize de novo purines while other human cells can use alternate pathways for this synthesis [7]. Therefore, MPA is a powerful immunosuppressive drug. The resulting immunosuppression leads to opportunistic infection occurrence [8]. Nevertheless, data on the occurrence of PCP in patients treated with MPA are contradictory. MPA may have an antifungal activity on Pneumocystis. Oz and Hughes used a rat model which demonstrated that none of the animals treated with MPA alone or in association with dexamethasone developed PCP whereas 90% of rats treated with dexamethasone or tacrolimus alone developed PCP after 4 weeks of treatment [9]. Husain and Singh showed that PCP frequency in RTR treated with MPA was significantly lower than in RTRs without MPA treatment (0/1068 vs. 10/563, p = 0.00006) [8]. On the other hand, we have published three series of PCP case clusters in RTR and heart transplant recipients (HTR) and noted the occurrence of 38 PCP cases in patients effectively treated with MPA [10][11][12]. Considering these aforementioned observations, the putative antifungal activity of MPA on P. jirovecii remains a subject of debate. Conversely, MPA is known to have an in vitro antifungal activity on other fungal species which is useful in fundamental mycology [13].
The emergence of pathogen strains under selection pressure is widespread in microbiology and also observed for P. jirovecii [11,14]. In this context, the hypothesis of the emergence of particular genotypes/phenotypes of P. jirovecii under MPA selection pressure is relevant. The objective of the present study was to investigate this field and we retrospectively sequenced the impdh gene of P. jirovecii isolates from SOT recipients treated with MPA and a control group of non-SOT recipients with no prior exposure to MPA.

Hospital Settings
Brest University hospital is a 2500-bed healthcare facility, characterized by 460,000 outpatient visits and 120,000 inpatients monitored per year. Specialized care is provided for about 550 renal transplant recipients.

Patients and P. jirovecii Specimens
Twenty-six P. jirovecii specimens from 26 patients diagnosed with PCP, between 2010 and 2020, were retrospectively examined. They consisted of one group of 10 SOT recipients (T1 to T10) (sex ratio M/F 7/3; median age 65 years (55-78 years)) and another group of 16 patients who were not transplant recipients (C1 to C16) (sex ratio M/F 12/4; median age 60 years (30-89 years)). The first group had undergone MPA treatment since the day of transplantation, with a median duration of exposure of 51 months (range: 8-216 months) conversely to the second group who did not have any prior exposure to MPA (control group). Patient characteristics are summarized in Tables 1 and 2. Biological diagnoses of PCP were initially based on microscopic detection of P. jirovecii and/or DNA amplification using a qPCR assay targeting the gene encoding the mitochondrial large subunit of ribosomal RNA which was performed after DNA extraction using NucliSens easyMag (bio-Mérieux, France) as we previously described [15]. P. jirovecii DNA specimens were stored at −80 • C for further analysis. The study was non-interventional and did not require inform consents and ethical approval according to French laws and regulations (CSP Art L1121e1.1). Specimen collection was registered with the French Ministry of Research (No. DC-2008-214). The male/female ratio was 7/3, the median age was 65 (range 55-78); Age and sex were not indicated in extenso, since it could be identifiable patient information; BAL: bronchoalveolar lavage; ND: not determined; +/−: positive/negative; HTR, heart transplant recipient; LTR, liver transplant recipient; RTR: renal transplant recipient; SMX-TMP: sulfamethoxazole-trimethoprim. a The post-transplantation period is also the duration of MPA exposure. b Microscopic detection of P. jirovecii was performed using Wright-Giemsa staining and, until December 2017, an immunofluorescence assay (MonofluoKit P. jirovecii, Bio-Rad). c PCR detection of P. jirovecii was performed using a real-time PCR assay, targeting the mitochondrial large subunit ribosomal RNA gene [15]. d First-line treatment. e Second-line treatment.
Six primer pairs were designed using PrimerQuest© software (Integrated DNA Technology, https://www.idtdna.com/Primerquest/Home/Index, accessed on 14 September 2021) in order to sequence the whole IMPDH gene, which is 2070 bp long. Primer specificity was checked using Basic Local Alignment Search Tool (https://blast.ncbi.nlm.nih.gov/ Blast.cgi, accessed on 14 September 2021). Each primer pair amplifies a DNA fragment from 376 to 915 bp in size (Table S1). Each fragment partially spanned at least one other fragment ( Figure S1). The six fragments of the impdh gene were amplified under the same PCR conditions. PCR reactions were performed in a final volume of 50 µL containing 0.5 µM of reverse and forward primers, 3 mM MgCl 2 , 200 µM dNTPs (dNTP set, Eurogentec, Belgium) and 2 U of polymerase (HotGoldStar Diamond Taq ® , Eurogentec, Belgium). Amplification was carried out with activation step at 95 • C for 10 min followed by 45 cycles including denaturation at 94 • C for 30 s, annealing at 60 • C for 45 s and extension at 72 • C for 1 min and a final extension step at 72 • C for 10 min. To avoid contamination, reagent preparations and amplification procedures were performed in separate rooms with different sets of micropipettes. Negative control was included in each PCR. PCR products were electrophoresed on 1.5% agarose gel containing ethidium bromide to visualize the expected band sizes (Table S1). PCR products were purified and sequenced from the two strands with the dideoxy chain termination method on the ABI 3130XL Genetic Analyzer (BigDye ® terminator v1.1 cycle sequencing kit, Applied Biosystem, USA). Electropherograms were analysed using PreGap and Gap software (Staden package version 2.0.0b11-2016, Staden Group). The six fragments were assembled to create the consensus sequence of the impdh gene from each specimen. Nucleotide and protein sequences were aligned with the reference sequences retrieved from complete P. jirovecii genomes as described above using BioEdit software (Version 7.0.0, Thomas Hall, Ibis Biosciences, USA) with the Clustal ® W program. The identification of putative IMPDH alleles was based on the single nucleotide polymorphisms (SNPs).

Statistical Analysis
Frequencies of P. jirovecii IMPDH alleles and potential mixed infections, which were defined by the presence of more than one allele in the same specimen, were compared in the two patient groups using Fisher's exact test. A p value of <0.05 was considered significant.

Nucleotide Sequence Accession Number
The nucleotide sequences of the impdh gene obtained in the present study have been deposited in GenBank under accession numbers from MZ272376 to MZ272401.
a Two nucleotides were identified at this position on sequence electropherograms.
Consensus sequences of P. jirovecii impdh gene in each patient were also aligned with the coding DNA sequences (CDS) of the two reference impdh genes (T551_00631 and PNEJI1_002926 loci) [17,18]. SNP at positions 1269 and 1329 were located within introns ( Figure S2). SNP at position 76, 914, 1020, 1737, 1754, 1850 and 1952 were located within exons. Deduced protein sequences were compared to the reference protein sequences (UniProt ID A0A0W4ZU90 and L0P7N3) [17,18]. Four synonymous mutations and 3 missense mutations were identified. Missense mutations were observed at nucleotide positions 76, 1020 and 1737. The first mutation corresponds to a transversion from G to T (G76T) and an amino acid substitution from alanine (Ala) to serine (Ser) at codon 26 (Ala26Ser) ( Figure S2; Figure 1). This change characterising the allele IMPDH-7 was found in one control patient (C10). The second mutation corresponds to a transition from G to A (G1020A) and an amino acid substitution from alanine to threonine (Thr) at codon 261 (Ala261Thr) ( Figure S2; Figure 1). It is noteworthy that this change characterising the allele IMPDH-3 was exclusively found in the SOT recipients and in all of those. The third mutation, found in three control patients (C6, C14 and C16) but not in SOT recipients, corresponds to a transition from G to A (G1737A) and an amino acid substitution from glycine to serine at codon 439 (Gly439Ser) ( Figure S2; Figure 1). This change characterising the allele IMPDH-4 was found in three control patients (C6, C14 and C16) and was not found in SOT recipients.
P. jirovecii IMPDH protein sequences were aligned with available IMPDH proteins from relevant fungi, i.e., the IMPDH of other species of the Pneumocystis genus (P. carinii, P. murina) [17], C. albicans [21], M. guilliermondii [22], S. cerevisiae [23], P. brevicompactum [24] and A. nidulans [24] (Table S2) (Figure 1). The two IMPDHs of P. jirovecii that were retrieved from the two available proteomes (one from Ma and colleagues, the other from Cissé and colleagues) are identical and can be considered as the wild type of IMPDH [17,18]. Among the three aforementioned missense mutations identified in the present study, the mutations Ala26Ser and Gly439Ser are located within non-conserved regions ( Figure  1) [24,25]. Conversely, mutation Ala261Thr is located in a strongly conserved region, Ala261 being itself a strongly conserved residue (Figure 1) [24]. Indeed, an Ala was observed at the analogous positions within the IMPDHs of P. jirovecii (wild type), P. carinii

Discussion
In this study, the first data on P. jirovecii impdh gene diversity were obtained. To the best of our knowledge, data on the diversity of the impdh gene in the most studied Pneumocystis species, specifically P. carinii, from rats and P. murina, from mice, are not available. Nonetheless, IMPDH of P. carinii as a potential drug target of MPA was previously evaluated by O'Gara and colleagues [16]. P. carinii's sensitivity to MPA was established using a short-term culture of P. carinii [16]. This sensitivity was confirmed using an IMPDHdeficient strain of Escherichia coli complemented with a plasmid containing the P. carinii impdh gene [16].
Among the nine SNP identified in the overall P. jirovecii specimens from 26 patients, 6 were synonymous mutations, whereas 3 others, Ala26Ser, Ala261Thr and Gly439Ser, resulted in missense mutations. Since the two first mutations Ala26Ser and Gly439Ser are located within non-conserved regions, they might represent natural polymorphisms that might not affect protein function as it has been established in the human IMPDH1 (hIMPDH1) [7,26]. Nonetheless, since the mutation Gly439Ser is located in a hinge region, an effect on enzymatic activity cannot be ruled out.
The mutation Ala261Thr is located in a highly conserved region in fungal IMPDHs, Ala being itself a highly conserved residue, considering our alignments ( Figure 1) combined with data reported by Freedman and colleagues [24] who aligned 20,000 fungal IMPDHs and found that the residues at analogous position 261 (the position identified in P. jirovecii IMPDH in the present study) were 91% Ala, 7.3% Gly, 1.2% Ser, 0.1% Thr and 0.1% Val [24]. Interestingly, fungi displaying natural or acquired resistance to MPA or at least naturally exposed to MPA in their environment were previously described to harbour a Ser or a Thr residue at the corresponding position in the IMPDH protein sequence. For example, a substitution from Ala to Thr (Ala251Thr) has been identified in an MPA-resistant strain of C. albicans selected under a high concentration of MPA (10 µg/mL) (Figure 1) [21]. Similarly, P. brevicompactum IMPDH-A (PbIMPDHA-MPA R ), S. cerevisiae IMPDH2 (ScIMPDH2-MPA R ) and M. guillermondii IMPDH2 (MgIMPDH2-MPA R ) harbour a substitution from Ala to Ser at the analogous position (Ala267Ser, Ala253Ser and Ala251Ser, respectively) [22][23][24] (Figure 1), these three fungi being naturally resistant to MPA. Consequently, the Ala to Ser or Thr substitution in IMPDHs thus, stands out as a critical determinant of MPA-resistance in fungi [22][23][24]. Moreover, the in vitro substitution of Ala267 with Ser made the IMPDH of Aspergillus nidulans, a naturally MPA-sensitive fungus, resistant to MPA [24].
Interestingly, the mutation Ala261Thr in P. jirovecii IMPDH was found only in the 10 SOT recipients enrolled in the study, all of them being still subject to anti-rejection treatment with MPA at the date of PCP diagnosis. Nonetheless, putative resistance of P. jirovecii to MPA is probably not a risk factor for PCP occurrence. Indeed, the patients developed PCP because none of them were undergoing chemoprophylaxis when P. jirovecii infections occurred. In fact, the post-transplantation period ranged from 8 months to 216 months, whereas the duration of recommended prophylaxis is limited to 6-12 months after transplantation [5]. However, the mutation Ala261Thr among P. jirovecii isolates from SOT patients may represent a signature of MPA exposure and the circulation and transmission of specific P. jirovecii strains in susceptible patient populations. This signature of MPA exposure was previously pointed out in environmental fungal species such as Aspergillus glaucus, Aspergillus clavatus, Aspergillus terreus, which encode IMPDHs harbouring Ser or Thr at the analogous position [24]. These environmental moulds have probably been confronted with MPA during their evolution since they are known to develop in natural niches shared with Penicillium species that naturally produce MPA [24].
Consequently, MPA may have favoured the emergence of specific strains of P. jirovecii revealed by this specific mutation in SOT recipients. It is noteworthy that SOT recipients T3 and T9 harboured allele IMPDH-3 (characterised by the G1020A mutation resulting in Ala261Thr) in combination with allele IMPDH-1, whereas the remaining eight SOT recipients harboured IMPDH-3 solely. These mixed infections may be explained by an in-progress process of selection of P. jirovecii organisms under MPA pressure, wild P. jirovecii strains being not entirely replaced by a predominant mutant P. jirovecii strain. While MPA was initially considered to protect against P. jirovecii in the 1990s [8], an increase in PCP cases in SOT recipients has been reported, particularly in RTR, in a context of nosocomial infection over the past 15 years [6,27]. At present, MPA is identified as one of risk factors for PCP [28]. This paradigm shift could be explained in part by MPA pressure on P. jirovecii populations, selection of specific strains characterised by potentially lower sensitivity to MPA and acquisition/transmission of these specific strains among susceptible SOT recipients.
The phenomenon of P. jirovecii selection under drug pressure has already been reported with the description of specific genotypes/phenotypes of dihydropteroate synthase (DHPS) in patients treated with sulfonamides that target the DHPS [14,29,30]. Prior exposure to these drugs was identified as the main predictor of P. jirovecii DHPS mutants [14,29,30]. The patients' city of residence, whether or not they had prior exposure to sulfonamides, has been identified as another predictor of P. jirovecii DHPS mutants [14,29]. This observation was explained by possible P. jirovecii transmission from infected, treated patients to susceptible, untreated patients. Similarly, we recently described a P. jirovecii mutant of the gene of cytochrome b (CYB), which is the target of atovaquone [11]. This mutant was spread in HTR who were effectively subject to PCP prophylaxis by atovaquone, but also in HTR without prior exposure to atovaquone, in a context of PCP outbreak [11]. For these reasons, analysis of DHPS locus and CYB genes serves as a marker of P. jirovecii circulation within human communities.
The presence of allele IMPDH-3 (characterised by the G1020A mutation resulting in Ala261Thr) and the low diversity of the P. jirovecii impdh gene in SOT recipients suggest a common identical source of the fungus in this patient group. Conversely, the high diversity of impdh gene, as well as the absence of the IMPDH-3 allele suggest several different sources of P. jirovecii in the control group. This diversity resulted in the identification of nine SNPs and eight alleles.
Considering the potential of impdh gene analysis in contributing to a method of P. jirovecii genotyping, we evaluated the Hunter index at 0.6952 (<0.95) [31]. This value is insufficient to distinguish P. jirovecii isolates in a genotyping approach based on this sole locus. Nonetheless, impdh gene analysis could be added to the highly discriminative multilocus sequence typing (MLST) method combining mtLSUrRNA, CYB and superoxide dismutase gene analysis, that we have used to investigate Pneumocystis case clusters in hospital [32]. Impdh gene analysis could provide an additional marker of P. jirovecii circulation among SOT recipients who are specifically at risk for PCP occurrence in a context of nosocomial acquisition of P. jirovecii and PCP outbreaks. MLST analysis including impdh gene may strengthen the concept of acquisition and transmission of specific P. jirovecii strains in SOT recipients, which has previously been discussed [6,33,34].
To sum up, we pointed out a missense mutation Ala261Thr in SOT recipients treated with MPA, whereas this change was not found in a control group without prior exposure to MPA. This mutation may represent the signature of MPA exposure and the selection of specific P. jirovecii strains under selection pressure that circulate in the SOT recipient population. A multicentre study, combining the enrolment of patients from diverse geographical regions and the genotyping of P. jirovecii specimens based on the aforementioned MLST approach is now warranted.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/jof7100849/s1, Figure S1: Description of PCR products obtained using the six primer pairs designed for Pneumocystis jirovecii impdh gene amplification, Figure S2: Alignment of impdh CDS and genomic sequences of the two P.jirovecii genome assemblies, Figure S3: Diversity of the impdh gene of Pneumocystis jirovecii based on the analysis of 26 specimens, Table S1: Nucleotide sequences of primers used for amplification of Pneumocystis jirovecii impdh gene, Table S2: List of the IMPDH protein sequences from Pneumocystis species and other fungi that were aligned with the IMPDH sequences obtained in the present study. Institutional Review Board Statement: Ethical review and approval were waived for this study, in accordance with the French public health law (CSP Art L1121e1.1), because this study was noninterventional, focused on P. jirovecii specimens and used samplings and data obtained from the patients' standard routine care.

Informed Consent Statement:
Patients' consent were not required for this study, in accordance with the French public health law, because this protocol only used samplings and data obtained from the patients' standard routine care.

Data Availability Statement:
The nucleotide sequences of the impdh gene obtained in the present study have been deposited in GenBank under accession numbers from MZ272376 to MZ272401.