Background: A 53-year-old man presented with an acute bilateral posterior uveitis with extensive necrotising retinochoroiditis but without chorioretinal scarring. A thorough workup did not reveal any underlying disease. The possibilities of atypical ocular toxoplasmosis as well as herpetic retinal necrosis were considered and specific therapy instituted, with little improvement. The patient died within 2 months as result of an undifferentiated squamous cell carcinoma.
Methods: Histopathological examination, immunohistochemistry and multilocus polymerase chain reaction confirmed Toxoplasma gondii infection of the retina
Results: Macroscopic examination of enucleated globe showed extensive retinal necrosis and vitreous detachment. Histological examination of retinal tissue identified numerous round-to-elliptical toxoplasmic cysts within the retina, with retinal necrosis and minimal choroidal inflammation. Immunohistochemical analyses confirmed that the cysts were due to T gondii. DNA extracted from formalin-fixed, paraffin-embedded tissue sections was subjected to multilocus polymerase chain reaction (PCR) analysis at the following typing loci: SAG1, SAG2, SAG3, SAG4, B1, NTS2, GRA6 and GRA7. DNA sequencing of positive PCR products at the NTS2, SAG1 and GRA7 loci confirmed the presence of a non-archetypal strain of T gondii infecting the eye of the patient experiencing a severe, atypical ocular toxoplasmosis
Conclusion: A highly divergent, non-archetypal strain of T gondii was responsible for causing a severe, atypical bilateral retinochoroiditis in a patient from Brazil.
Statistics from Altmetric.com
In Brazil, 50–80% of the adult population is infected with Toxoplasma gondii.1 Multiple routes of exposure have been implicated including waterborne transmission and the ingestion of undercooked meat.23 Infection is associated with high levels of ocular disease, especially in the Southern States where incidence can be as high as 18%, and recurrent disease is both serious and common.4 This contrasts with Europe and the USA where the incidence of ocular toxoplasmosis is much lower, ranging from 2 to 5%.5 Epidemiological and population genetic data indicate that many Brazilian patients experiencing ocular disease acquire infection postnatally46 and that they are infected with strains genetically distinct from clonal lines circulating in Europe and USA.178910 Strain type is thought to be an important factor determining the severity of ocular disease, and recent molecular studies in Brazil have identified a rich diversity of strains infecting people and their companion animals. Polymerase chain reaction (PCR)-based typing at a single SAG2 locus has previously identified a Type I lineage allele among outbreak patients experiencing high levels of ocular disease associated with atypical, necrotising lesions.311 However, typing at a single locus fails to identify the true genotype of recombinant and/or atypical strains that bear combinations of archetypal and non-archetypal alleles, as are common in Brazil.1 Typical features of ocular toxoplasmosis include focal acute necrotising granulomatous retinochoroiditis that is often associated with a pre-existing pigmented retinochoroidal scar, called satellite lesions. Atypical presentations include multiple large foci of bilateral acute necrosis, retinal haemorrhages and vasculitis. Such atypical lesions often present a challenge in establishing the diagnosis.1213
Herein, we report a case of a Brazilian patient with bilateral retinal necrosis. Histological examination revealed numerous Toxoplasma cysts, and multilocus PCR-DNA sequencing analyses from formalin-fixed, enucleated eye tissue sections confirmed this patient was infected with a non-archetypal strain of T gondii. This report highlights the necessity of performing multilocus genotyping analyses and describes new primers that amplify from formalin-fixed tissue sections making retrospective analyses possible.
A healthy, 53-year-old man appeared complaining of an acute loss of vision and pain in both eyes. His past ocular and systemic histories were unremarkable. His best-corrected visual acuity was 20/50 in the right eye and 20/200 in the left eye.
Slit-lamp biomicroscopy showed granulomatous keratic precipitates and inflammatory cells in the anterior chamber and vitreous in both eyes. The left eye showed corneal oedema, fibrin in the anterior chamber and posterior synechiaes (fig 1) Funduscopy of the right eye revealed extensive and poorly defined retinochoroiditis involving the inferior retina, associated with 2+ vitreitis, intraretinal haemorrhages, marked periphlebitis and periarteritis, and absence of chorioretinal scars (fig 2). In the left eye, the fundus exam was not possible due to 4+ vitreitis. Fluorescein angiography of the right eye disclosed hypofluorescence of the lesion during the early phases of the study, followed by progressive hyperfluorescence secondary to vascular leakage and occlusive vasculitis (fig 2).
Multiple aetiological agents were considered in the differential diagnosis of bilateral acute uveitis with occlusive vasculitis and retinal necrosis. A workup for sarcoidosis, syphilis, tuberculosis, toxoplasmosis, cytomegalovirus, HIV, herpes, masquerade syndrome by intraocular lymphoma and Behçet disease was negative, except for positive circulating IgG antibodies for toxoplasma and herpes simplex. A thorough systemic workup did not reveal any underlying disease.
Possible acute herpetic retinal necrosis was considered, and the patient wass treated with intravenous and orally acyclovir and oral prednisone (1 mg/kg/day). One month later, his best-corrected visual acuity was improved. However, after 2 months, his visual acuity decreased with worsening of the vitreous charges and increasing areas of retina necrosis in both eyes. An atypical presentation of ocular toxoplasmosis was considered, and systemically trimethoprim/sulmethoxazole was added.
The following month, his systemic condition progressively deteriorated. He presented with loss of weight, a painful left shoulder, hoarseness and cervical lymphadenopathy. A full physical examination, complete laboratory assessment and radiological work-up as well as endoscopic procedures, as indicated, were performed. Lymph-node biopsy established the diagnosis of squamous cell carcinoma from an unknown primary site (CUP). Cervical lymph nodes, bone and liver metastasis were found. The patient died 2 months later, and his eyes were donated for research.
Serology, histopathology and immunohistochemistry
Serum was tested for IgG reactivity for T gondii. Enucleated eye tissue was fixed in 10% neutral buffered formalin, trimmed, paraffin-embedded and 5–10 μm sections cut and stained with haematoxylin and eosin (H&E). Immunohistochemistry was performed to screen for reactivity to antibodies against T gondii on formalin fixed, paraffin-embedded tissues.
Genomic DNA was extracted from 10 μm cut formalin fixed, paraffin-embedded sections according to the manufacturer’s instructions of the DNeasy Tissue kit (Qiagen, Valenciq, California). Paraffin was xylene-extracted from embedded tissue sections prior to DNA isolation. Genomic DNA preparations were screened for the presence of T gondii DNA by nested PCR amplification for the following diagnostic, strain-typing loci: B1, NTS2, SAG1, SAG2, SAG3, SAG4, GRA6 and GRA7. One water (DNA-free) negative control and one Type I strain (RH) positive control were tested with each assay. Both rounds utilised 5 μl of PCR buffer (10× SigmaTaq PCR buffer containing 15 mM MgCl2), 0.1 mM dNTP mix, 10 pmol of each primer and 1.5 U of Taq DNA polymerase in a total reaction volume of 50 μl. Each of the 35 cycles consisted of 94°C for 40 s, 58°C for 40 s and 72°C for 40 s. Samples were amplified in an automated thermocycler (Eppendorf Mastercycler). PCR products were visualised using GelRed staining of 1–2% agarose gels. Primer sequences for B1, SAG1, SAG2, SAG3, SAG4 and GRA6 have been described before.141516 Primer sequences for NTS2 and GRA7 were: NTS2, Fext: TGTGCTCGTGACTTGATGTG, Fint: ACCAGAAAGCCAGTGGAATG, Rext: GGAAAACAAGCGGTGAGATT, Rint: TTTGTTTTTCTCGCGTAGAGG.; and GRA7, Fext: CCAGCATGGATAAGGCATCT, Fint: GACGCTGAAGTGACTGACGA, Rext: GACACTGTCCTCGAGCTCCT, Rint: TTAGCCCCCATATCCTACTGG.
Macroscopic examination of the enucleated globe showed retinal necrosis with ghost retinal vessels and vitreous detachment. Histopathological examination with H&E-stained tissue sections revealed numerous round-to-elliptical toxoplasmic cysts within the retinal necrosis. Immunohistochemistry confirmed the diagnosis (fig 3).
To determine the strain type that caused infection, DNA extracted from the paraffin-embedded, formalin-fixed tissue sections was subjected to nested PCR using eight highly sensitive, genetic typing loci: B1, NTS2, SAG1, SAG2, SAG3, SAG4, GRA6 and GRA7. The B1 gene is widely used for the molecular diagnosis of T gondii infection due to its high sensitivity; there are 35 copies of this gene in the Toxoplasma genome. NTS2 (Non Transcribed Spacer 2) is another multicopy gene locus that lies in the region between the 28S and 18S rRNA genes. The majority of the other markers developed for multilocus genotyping are derived from highly polymorphic surface antigens (SAGs) or parasite proteins secreted from a unique set of Apicomplexan organelles collectively known as dense granules (GRAs). Given the difficulties of PCR amplification using DNA extracted from formalin-fixed specimens, we evaluated a wide panel of genotyping markers for efficacy.
PCR amplification for T gondii was successful using the NTS2, SAG1 and GRA7 loci only (fig 4A). All genetic markers were tested at least twice on independent DNA extractions from two separate tissue sections. All markers were negative against eye tissue extracted from a patient not infected with Toxoplasma (data not shown). Positive PCR products were purified and the allele determined by DNA sequencing. A Type I allele was identified at NTS2, a Type II or III strain allele was identified at SAG1, and a highly divergent, atypical allele was identified at GRA7 (fig 4B). Multiloci genotyping established that the Brazilian patient was infected with a non-archetypal strain that appears to be a recombinant strain possessing a mixture of non-archetypal and archetypal alleles.
Atypical toxoplasmic retinochoroiditis has been described in immunocompromised patients including those with AIDS as well as in immunocompetent patients receiving immunosuppressive drug therapy or transplant recipients and elderly individuals.512 Variation in the clinical presentation and severity of disease has been attributed to several factors, including the genetic heterogeneity of the host and the genotype of the parasite responsible for infection.1217 This study identified a novel recombinant strain of T gondii that caused an atypical ocular toxoplasmosis. A Type I allele was identified at NTS2, a Type II or III strain allele was identified at SAG1, and a highly divergent, atypical allele was identified at GRA7. Thus, the strain infecting this patient appeared to be a novel recombinant genotype possessing a mixture of non-archetypal and archetypal alleles.
This study highlights the development of several robust, new multilocus genotyping markers that were both highly sensitive and capable of annealing to parasite DNA extracted from formalin-fixed clinical tissue specimens. Primers against eight genetic typing loci (B1, NTS2, SAG1, SAG2, SAG3, SAG4, GRA6 and GRA7), were tested and three markers (NTS2, SAG1, and GRA7) consistently amplified parasite DNA extracted from different cut sections from the paraffin-embedded, formalin-fixed block of the patient’s eye tissue. These new polymorphic markers now extend the arsenal of molecular typing tools to formalin-fixed tissues, making it possible to retrospectively perform multilocus genotyping on archival clinical specimens. It is important to note that the B1 primers, which are routinely used for the molecular diagnosis of T gondii infection, did not work against DNA preparations from the formalin-fixed specimen; nor did the SAG2 primers, which have previously been applied successfully against acetone-fixed tissue specimens.9 These facts underscore the necessity of applying a battery of molecular markers to maximise the potential for assigning a multilocus genotype against DNA preparations extracted from formalin-fixed specimens. Application of these molecular markers should enable researchers and clinicians to test whether specific genotypes of circulating T gondii are associated with the development of ocular toxoplasmosis. Multilocus typing is preferred because the exclusive use of only a single marker is not sufficiently resolved to accurately determine strain genotype in regions of the world (eg, Brazil) where strain diversity is greater and non-archetypal and recombinant strains are detected more frequently.12319 And for those cases of ocular toxoplasmosis where T gondii DNA is not available or cannot be amplified due to prolonged fixation in formalin, it is now possible to ascribe a serotype by reacting patient serum in an ELISA format against a panel of synthetic polymorphic peptides derived against several T gondii antigens that pull out strain-specific antibodies circulating in infected patients.18 These new tools should make it possible to ask whether strain “type” is a predictor for the development of ocular disease and may ultimately become an important diagnostic and prognostic tool for management of ocular disease.
Whether the strain infecting this patient is related to, or identical to, strains previously identified infecting ocular toxoplasmosis patients from Brazil is not possible to address because the majority of previous investigations simply investigated a single typing locus, SAG2. In a waterborne outbreak in Parana State, Brazil, a high incidence of associated ocular toxoplasmosis was diagnosed and genotyping performed on two epidemiologically linked isolates from the outbreak identified only Type I lineage alleles at SAG2.3 A study of nine immunocompetent patients from Brazil bearing a severe, atypical ocular toxoplasmosis also identified a Type I lineage allele at SAG2.311 The authors of these two studies concluded that Type I strains were associated with ocular toxoplasmosis in Brazil, but in the absence of multilocus typing using a wider set of sequence-based markers, there is no way of telling whether the strain(s) associated with ocular toxoplasmosis in these patients were truly Type I or non-archetypal, recombinant strains like that found infecting the patient described herein. In this study, a Type I lineage allele was identified at NTS2, so typing using only this locus would have led to an erroneous conclusion. It is, however, possible to distinguish this strain from the one(s) that caused a severe or atypical toxoplasmic retinochoroiditis in 12 otherwise healthy adult patients in the US because it possessed a Type II or III lineage allele at SAG1, whereas the US strains all possessed a Type I lineage allele. These results make it clear that future work should incorporate multilocus typing to identify the diversity and epidemiological origins of the strain(s) associated with the development of ocular disease in people both regionally and worldwide. Future studies should be applied retrospectively against banked clinical material and prospectively against patient cohorts with ocular disease to address what strains pose the greatest risk for developing a severe form of toxoplasmic retinochoroiditis.
The case presented herein also showed a patient with bilateral atypical ocular toxoplasmosis as the first sign of CUP origin, which is perceived to be a very aggressive and ominous disease carrying a poor prognosis. Patients with immunodeficiency, including those with cancer or treated with immunosuppressive drugs, have a higher rate of opportunistic infections. The compromised immune system of the host may have been an additional factor responsible for the atypical and severe presentation in this case, considering that the strain involved was also an atypical one.
The ability to identify parasite types, although not routinely performed, may now provide new insights into the pathogenesis of this disease and, ultimately, become an important diagnostic and prognostic tool.
The authors would like to thank E James and S Magargal for technical assistance and for the grants from FADA-UNIFESP and CNPq (Rubens Belfort Jr).
Funding This work was supported by the Intramural Research Program of the NIH and NIAID. MEG is a scholar of the Canadian Institute for Advanced Research (CIFAR) Program for Integrated Microbial Biodiversity.
Competing interests None.
Provenance and Peer review Not commissioned; externally peer reviewed.
Ethics approval Ethics approval was provided by Federal University of São Paulo.
Patient consent Obtained.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.