Abstract
A 25-year-old woman presented with unilateral red eye and visual blur, and was found to have panuveitis with an inflammatory white mass at the macula, initially presumed to be Toxoplasma retinitis. After failure to respond, she underwent vitrectomy, which produced Candida albicans. Despite intraocular and systemic antifungal treatment, she lost all vision in that eye. Two years later, she developed unilateral hip osteomyelitis leading to total hip replacement and also revealing Candida infection. By clinical exome sequencing, she was then found to have caspase recruitment domain 9 (CARD9) deficiency, an autosomal recessive disorder that causes a specific susceptibility to candidal infections. She remains otherwise well but on lifelong fluconazole prophylaxis.
Background
Endogenous endophthalmitis is a rare but sight-threatening infection. Fungal endophthalmitis is most often seen following line infection in patients undergoing prolonged intravenous treatment, or in ‘mainlining’ drug addicts,1 and sometimes in the immunodeficient. The importance of caspase recruitment domain 9 (CARD9) as a signaller in the Candida immune pathway has only recently been described,2 and susceptibility to Candida infection may be caused by CARD9 gene mutations.3 Candida endophthalmitis associated with CARD9 deficiency has not previously been described; we report a case also associated with fungal osteomyelitis.
Case presentation
A 25-year-old woman presented with a 5-day history of right (R) red eye and blurred vision. She was otherwise completely well with no significant medical history. On examination, the visual acuity was R 6/36, left 6/6. The left eye was completely normal. The right eye showed severe panuveitis (anterior chamber cells +++ with no hypopyon, severe vitritis) and a large white focal inflammatory mass at the R macula (figure 1). She was initially treated on the presumption of primary ocular toxoplasmosis, with cotrimoxazole and prednisolone, with which the uveitis improved significantly. She was, however, found to be seronegative for Toxoplasma (both IgM and IgG), Borrelia and Bartonella, and Mantoux and interferon-γ testing was negative. The ACE level was normal at 35. After tapering the prednisolone dose, the lesion and uveitis worsened (figure 2), and she underwent vitrectomy and internal limiting membrane (ILM) peel, which revealed a choroidal/intraretinal mass that could not be removed; microscopy of vitreous revealed no organisms, and the sample was culture-negative (aerobic and anaerobic) at 10 days. Aqueous humour was PCR negative for herpes simplex virus types 1 and 2, varicella zoster virus and Candida.
Figure 1.

The right fundus at presentation, showing a white macular lesion with associated serous retinal elevation.
Figure 2.

An enlarged macular lesion with posterior vitreous infiltration after tapering of oral steroids.
Over the succeeding month, vision worsened and the patient developed an enlarging mass (figure 3) with vitreous involvement strongly suggestive of fungal infection (figure 4). On further questioning about risks of fungal infection, it transpired that the patient had undergone skin tattooing 4 months before presentation, and, prior to that, a tongue piercing. She had not injected drugs and there was no history of intravenous access. She began oral voriconazole and underwent further surgery including cataract extraction by phacoemulsification with intraocular lens (IOL) implantation, with epiretinal membrane peel, which was complicated by intraocular haemorrhage and retinal detachment requiring silicone oil injection. An intraocular sample was found to be PCR positive for Candida but negative for Aspergillus. Voriconazole 200 mg two times a day was continued for a further 8 weeks. Intraocular inflammation settled but the eye became phthisical with leukokoria and no perception of light.
Figure 3.

A substantially enlarged lesion now covering the posterior pole.
Figure 4.

Vitreous strands and ‘pearls’ suggesting fungal endophthalmitis.
Outcome and follow-up
Two years and 9 months after initial presentation with endophthalmitis, the patient developed progressive pain in the left hip; imaging revealing narrowing of the joint space and a subarticular cyst, progressing to destructive arthritis. Six months after developing symptoms, she underwent total hip replacement. The excised femoral head showed granulomatous osteomyelitis and synovitis with focal osteonecrosis, and fungal hyphae were present, Candida albicans being cultured from granulation tissue (C. albicans is a dimorphic yeast that may show either hyphae or yeast cells in tissue). There was no evidence of mycobacteria. Candida was also present in the vagina, and although there was no fungaemia, 1,3-β-d-glucan was present (>80 pg/mL), indicating systemic fungal infection. She was HIV-seronegative and there was no evidence of cellular immunodeficiency. She was treated with high-dose systemic antifungals, tapering to maintenance fluconazole, and remains well after a further 2½ years using fluconazole 100 mg daily, which is likely to be lifelong.
Clinical whole exome sequencing was undertaken. Genomic DNA was extracted, using standard methods, from a peripheral blood sample. Enrichment was achieved using SureSelect Human All Exon Kit v.5 (Agilent). Paired-end sequencing was run on an Illumina HiSeq2500. The sequence data were mapped to the human reference genome (UCSC Genome Browser hg19) with the Burrows-Wheeler Aligner. Variant calling was performed with a modified Genome Analysis Toolkit (GATK) v.2.4.7 pipeline. A panel of six genes associated with increased susceptibility to fungal infection, including CARD9, was selected for detailed bioinformatic sequence analysis. Confirmation of variants determined by exome sequencing was undertaken by Sanger sequencing on an ABI 3730 DNA sequencer (Applied Biosystems).
The patient was identified to be compound heterozygote for two novel variants in CARD9. A missense variant c.1138G>C, p.(Ala380Pro) and a synonymous variant c951G>A, p.(Arg317Arg). Neither variant was present in control databases (1000 Genomes, Exome Variant Server, Exome Aggregation Consortium) of over 70 000 individuals. The synonymous variant was inherited from her unaffected mother and is predicted to be pathogenic as it disrupts normal messenger RNA splicing. The missense variant arose de novo (absent in the mother or/and father), increasing the likelihood that this is the second pathogenic variant. This means that the risk to the patient's two siblings is negligible. CARD9 deficiency is a recessive condition and therefore the risk to the patient's future offspring would be very low (dependent on the very low risk that her partner was a carrier for a pathogenic variant).
Discussion
The immune response to Candida infections is complex and genetic systemic failures have been well-summarised:4 the initial stage requires recognition of the fungus by either toll-like receptors or dectin-1 (a pattern-recognition receptor for fungi); the latter requires the presence of CARD9 within the antigen-presenting cell. This activates NACHT, LRR and PYD domains containing protein 3, which leads to the release of the cytokines interleukin 1β (IL-1β), IL-6 and IL-18, contributing to T-cell recruitment. CARD9 also interacts with dectin-2, another C-type lectin, via a different signalling pathway that activates Th17 cells.5 CARD9 is therefore a key component of the signalling pathway, and deficiency may contribute to both, mucosal and invasive candidiasis.
The importance of CARD9 in the immune response to Candida was first described by Gross et al,2 and its role in Th induction and IL-17 production, shortly thereafter.6 Deficiency of CARD9 has been described: in a large consanguineous family, 3 of 7 affected members died of invasive candidiasis;3 in 8 families, 17 members developed deep dermatophytosis, of whom 4 died;7 and in 5 offspring of consanguineous marriages, each developed invasive Candida meningoencephalitis or colitis.8
Nosocomial candidaemia is an important and potentially fatal complication, especially on intensive treatment units.9 Candida endophthalmitis is rare, causing only 0.7% of uveitis seen at this tertiary referral centre over a 24-year period,10 but Candida is the most frequent pathogen causing fungal endophthalmitis.11 The great majority of patients are infected parenterally1 and more than half of patients are immunocompromised.11 Reports of endogenous Candida endophthalmitis in patients with no apparent predisposing risk are anecdotal,12 but one case was reported arising soon after skin tattooing in a patient with asplenia.13 Outcomes are often poor with one-third of patients developing severe visual loss (Snellen worse than 6/60).14
Deficiency of CARD9 clearly facilitates invasive candidiasis and, if gaining intraocular access, is likely to cause aggressive endophthalmitis, as in our patient. We suggest that any patient presenting with Candida endophthalmitis in the absence of a route of access and/or clear predisposing cause such as immunocompromised or diabetes, be assessed for potential immune abnormalities including CARD9 deficiency.
Learning points.
Remember the SPUR acronym to initiate investigations for immune deficiency: Severe, Persistent, Unusual or Recurrent infection.
Invasive candidiasis in children and young adults without clear underlying causes requires evaluation for caspase recruitment domain 9 (CARD9) deficiency.
Exome sequencing is an efficient and cost-effective means of identifying rare inherited conditions.
Patients found to be CARD9 deficient should use prophylactic fluconazole 100 mg/day to reduce the risk of further invasive candidiasis.
Acknowledgments
The authors acknowledge the assistance of Eleanor Baker in performing gene sequencing. This paper was facilitated by the NIHR Greater Manchester Clinical Research Network.
Footnotes
Contributors: NJ conceived the report. All the authors contributed significant clinical information and opinion. All the authors read and edited the manuscript.
Competing interests: None declared.
Patient consent: Obtained.
Provenance and peer review: Not commissioned; externally peer reviewed.
References
- 1.Chignell AH. Endogenous Candida endophthalmitis. J Royal Soc Med 1992;85:721–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Gross O, Gewies A, Finger K et al. CARD9 controls a non-TLR signalling pathway for innate anti-fungal immunity. Nature 2006;442:651–6. 10.1038/nature04926 [DOI] [PubMed] [Google Scholar]
- 3.Glocker EO, Hennigs A, Nabavi M et al. A homozygous CARD9 mutation in a family with susceptibility to fungal infections. N Engl J Med 2009;361:1727–35. 10.1056/NEJMoa0810719 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Smeekens SP, van de Veerdonk FL, Kullberg BJ et al. Genetic susceptibility to Candida infections. EMBO Mol Med 2013;5:805–13. 10.1002/emmm.201201678 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Robinson MJ, Osorio F, Rosas M et al. Dectin-2 is a Sky-coupled pattern recognition receptor crucial for Th17 responses to fungal infection. J Exp Med 2009;206:2037–51. 10.1084/jem.20082818 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.LeibundGut-Landmann S, Gross O, Robinson MJ et al. Syk- and CARD9-dependent coupling of innate immunity to the induction of T helper cells that produce interleukin 17. Nat Immunol 2007;8:630–8. 10.1038/ni1460 [DOI] [PubMed] [Google Scholar]
- 7.Lanternier F, Pathan S, Vincent QB et al. Deep dermatophytosis and inherited CARD9 deficiency. N Engl J Med 2013;369:1704–14. 10.1056/NEJMoa1208487 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lanternier F, Mahdaviani SA, Barbati E et al. Inherited CARD9 deficiency in otherwise healthy children and adults with Candida species-induced meningoencephalitis, colitis, or both. J Allergy Clin Immunol 2015;135:1558–68.e2. 10.1016/j.jaci.2014.12.1930 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hassan I, Powell G, Sidhu M et al. Excess mortality, length of stay and cost attributable to candidaemia. J Infect 2009;59:360–5. 10.1016/j.jinf.2009.08.020 [DOI] [PubMed] [Google Scholar]
- 10.Jones NP. The Manchester Uveitis Clinic: the first 3000 patients—epidemiology and casemix. J Ocular Immunol Inflammation 2013;23:118–26. 10.3109/09273948.2013.855799 [DOI] [PubMed] [Google Scholar]
- 11.Sridhar J, Flynn HW, Kuriyan AE et al. Endogenous fungal endophthalmitis: risk factors, clinical features, and treatment outcomes in mold and yeast infections. J Ophthal Inflamm Infect 2013;3:60 10.1186/1869-5760-3-60 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Stanbury RM, Chignell AH, Graham EM. Endogenous Candida endophthalmitis with no apparent predisposing factors. Eye (Lond) 1998;12(Pt 2):321–3. 10.1038/eye.1998.75 [DOI] [PubMed] [Google Scholar]
- 13.Alexandridou A, Reginald AY, Stavrou P et al. Candida endophthalmitis after tattooing in an asplenic patient. Arch Ophthalmol 2002;120:518–19. 10.1001/archopht.120.4.518 [DOI] [PubMed] [Google Scholar]
- 14.Sallam A, Taylor SR, Khan A et al. Factors determining visual outcome in endogenous Candida endophthalmitis. Retina (Philadelphia, Pa) 2012;32:1129–34. 10.1097/IAE.0b013e31822d3a34 [DOI] [PubMed] [Google Scholar]
