A RAB27A 5' untranslated region structural variant associated with late-onset hemophagocytic lymphohistiocytosis and normal pigmentation.

To the Editor:

Autosomal-recessive mutations in genes required for secretory lysosome-mediated lymphocyte cytotoxicity cause primary hemophagocytic lymphohistiocytosis (HLH), an early-onset, life-threatening hyperinflammatory syndrome. Similarly, mutations in RAB27A and LYST are associated with HLH, yet manifest hypopigmentation, because Rab27a and LYST also facilitate trafficking of pigment-containing lysosomes in melanocytes. We detail individuals from 5 families from the Baltic area with a novel structural variant at the 5′ untranslated region (UTR) of RAB27A associated with an atypical form of Griscelli syndrome type 2 (GS2) manifesting as late-onset HLH, marked neuroinflammation, skin granulomas, lymphoma, and normal pigmentation (see Table E1 in this article's Online Repository at www.jacionline.org).

Table E1

Clinical and laboratory findings at diagnosis of HLH

Family	1	2	3	4	5
Patient	P1	P2	P3	P4	P5
Ethnical origin	Lithuania	Sweden	Lithuania	Lithuania	Russia
Familial disease	Yes	Yes	No	No	No
Parental consanguinity	No	No	No	No	No
Sex	Female	Male	Male	Female	Male
RAB27A allele 1	Deletion, c.559C>T (p.Arg187Trp)	Dup-Inv	Dup-Inv	Dup-Inv	Dup-Inv
RAB27A allele 2	Dup-Inv	c.239G>C (p.Arg80Thr)	Dup-Inv	Dup-Inv	c.550C>T (p.Arg184*)
Age at diagnosis of HLH (y)	14	14.5	9	No HLH	No HLH
Fever	Yes	Yes	Yes	No	Yes (intermittent)
Splenomegaly	Yes	Yes	No	No	Yes, chronic (splenectomized)
Hepatomegaly	Yes	Yes	Yes	No	Transient, self-limiting
Hemoglobin (g/L)	83	87	90	Within normal	ND
Neutrophils (109/L)	0.45	0.80	0.14	Within normal	ND
Platelets (109/L)	35	54	194	Within normal	ND
Triglycerides (mmol/L)	2.49	5.8	2.71	Within normal	ND
Fibrinogen (g/L)	0.6	0.4	0.6	1.51	ND
Hemophagocytosis	No	No	Yes	ND	ND
Ferritin (μg/L)	2451	49000	12000	Within normal	ND
Soluble CD25 (pg/mL)	20512.4	ND	16437	1760	ND
NK-cell activity	ND	ND	Defective	Defective	Defective
NK-cell degranulation	Defective, 3.1% ΔCD107a	ND	Defective, 6% ΔCD107a	Defective, 3.3% ΔCD107a	Defective, 0.6% ΔCD107a
Neurological manifestations	Yes, before HLH onset	Yes	Yes, before HLH onset	Yes	No
Pathological CSF	Yes	ND	Yes	Yes	ND
Treatment active disease	HLH-2004	Cortico, CSA, ATG	HLH-2004	Cortico, MMF
Remission at 2 mo	No	No	Yes	NA
Age at HSCT	Not done		10	7
Outcome and follow-up	Deceased 48 d after HLH onset		Alive 19 mo after HSCT	Alive at 7 y	Alive at 16 y
Other manifestations	No	Skin granuloma, lung infiltrates	Skin granuloma, lung infiltrates	Skin granuloma, isolated CNS involvement	Hodgkin lymphoma at 13 y, recurrent fever episodes from 10 y

ATG, Antithymocyte globulin; Cortico, corticosteroids; CSA, cyclosporine A; CSF, cerebrospinal fluid; Dup-Inv, Duplication-Inversion; MMF, mycophenolate mofetil; NA, not applicable/available; ND, no data.

In brief, patient 1 (P1) developed HLH at age 13 years after 8 years of recurrent neuroinflammation (Fig 1, A and B; see Fig E1, A, in this article's Online Repository at www.jacionline.org). Her older brother died in infancy because of severe infectious mononucleosis. Patient 2 (P2) developed fatal HLH at age 15 years. Two of his 3 sisters died at age 13 years, one of an HLH-like syndrome and the other because of brain lymphoma. Patient 3 (P3) developed HLH at age 9 years, after 3 years of recurrent neuroinflammation (Fig E1, B). Patient 4 (P4) suffered from isolated neuroinflammation, without fulfilling HLH criteria. P2, P3, and P4 also had skin granulomas (Fig 1, C). Patient 5 (P5) suffered from recurrent fever flares and EBV viremia, and developed lymphoma at age 13 years. Hypopigmentation was not evident in any of the patients (Fig E1, D). Further clinical information and methods are provided in this article's Online Repository at www.jacionline.org.

Fig 1

Clinical and laboratory findings. A, Family pedigrees and RAB27A genotype. B, Brain axial magnetic resonance imaging FLAIR images of P1 at diagnosis of HLH showed nonspecific multifocal hyperintense white matter lesions (arrows). C, H&E stain of a skin biopsy from P3 reveals granulomatous infiltrates in the dermis. A nonnecrotizing granuloma (boxed area, left) is presented at a higher magnification (right).D, NK-cell cytotoxic activity of P3 was defective compared with relatives and controls. NK-cell exocytosis (E) and T-cell exocytosis (F) were reduced in P3. NK-cell cytotoxicity (G) and NK (H) and T-cell exocytosis (I) were defective in P4. dup-inv, Duplication/inversion; H&E, hematoxylin and eosin; n.a., not applicable/available; wt, wild type.

Fig E1

A, Brain MRI of P1 at onset of disease, axial series. B, Brain MRI of P3 at onset of disease. C, Lung computed tomography scan of P3. D, Photo of P3 showing normally pigmented hair.

Prompted by a suspicion of HLH, functional investigations of cytotoxic lymphocytes were performed on P1, P3, P4, and P5 (Table E1). Natural killer (NK)-cell cytotoxicity was defective (Fig 1, D and G). Moreover, NK- and T-cell exocytosis was impaired (Fig 1, E and F, H and I), with IL-2 stimulation inducing a partial recovery (see Fig E2, A-F, in this article's Online Repository at www.jacionline.org). Results therefore indicated primary HLH due to defective exocytosis.

Fig E2

A, NK-cell cytotoxicity assay in IL-2–stimulated PBMCs from P3 and relatives. B, NK-cell degranulation assay in IL-2–stimulated PBMCs from P3 and relatives. C, CD8+CD57+ T-cell degranulation assay in IL-2–stimulated PBMCs from P3 and relatives. D, NK-cell cytotoxicity assay in IL-2–stimulated PBMCs from P4 and relatives. E, NK-cell degranulation assay in IL-2–stimulated PBMCs from P4 and relatives. F, CD8+CD57+ T-cell degranulation assay in IL-2–stimulated PBMCs from P4 and relatives. G, Western blot analysis of Rab27a in P4 and relatives. H, Western blot analysis of Rab27a in P5.

Despite the clinical and immunological findings, a molecular diagnosis of primary HLH was not established by exome sequencing of HLH-associated genes. A heterozygous RAB27A missense variant of unknown significance (p.Arg187Trp) was identified in P1. Moreover, previously reported heterozygous RAB27A mutations were detected in P2 (p.Arg80Thr) and P5 (p.Arg184*) (Fig 2, A). To identify possible noncoding mutations, we performed whole-genome sequencing of P1 and P2. Inspection of sequencing data revealed several read pairs aberrant for insert size and orientation at the 5′UTR of RAB27A and over the first 5 exons of the adjacent PIGB, required for glycosylphosphatidylinositol anchor biosynthesis (Fig 2, A). Software-based analysis of structural variants (SVs) confirmed the presence of multiple overlapping SVs in this region, including shared events between P1 and P2 (Fig 2, A).

Fig 2

Genetic findings. A, Genome browser view over the region chr15:55,467,465-55,656,511 (hg19) including RAB27A and PIGB genes, with screenshots for P1 and P2 showing discordant read pairs. B, Model for the complex SV. C, Segregation analysis by MLPA. D, Rab27a expression in PBMCs evaluated by Western blot in P3, relatives, and healthy controls. E, mRNA expression of RAB27A in PBMCs and melanocytes. The expression of the transcript NM_183235.2 (long) was compared with the total expression of RAB27A. dup-inv, Duplication/inversion; GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

To confirm the presence of SVs affecting copy number at the 5′UTR of RAB27A, a custom multiplex ligation-dependent probe amplification (MLPA) assay was designed. In P1, a complex SV with a duplication-normal-duplication pattern inherited from the father was confirmed, in addition to a 65-kb deletion inherited from the mother (Fig 2, A and C). P2 and P5, who carried monoallelic coding mutations, were heterozygous for the complex SV (Fig 2, C). Complex SVs are usually associated with cryptic rearrangements. Further analysis of split reads and discordant read pairs of the complex SV indicated that one of the duplicated regions was inverted (Fig 2, B and C). In this model, only 1 of several transcriptional start sites (TSSs) of RAB27A, encoding the transcript NM_183235.2, is disrupted by the complex SV, while a copy of PIGB remains intact (Fig 2, B). Validation of one predicted breakpoint supported this model (see Fig E3, A, in this article's Online Repository at www.jacionline.org). Together, identification of biallelic mutations in RAB27A established a diagnosis of atypical GS2 in P1, P2, and P5.

Fig E3

A, Sanger trace for breakpoint C-invE in all patients. B, Quantitative PCR of PIGB to quantify RNA from PBMCs isolated in P3 and relatives. C, UCSC Genome browser screenshot showing localization of the RAB27A TSS according to FANTOM5 CAGE data from human cells. p1 is the TSS for the isoform NM_183235.2. D, Tags per million at the different RAB27A TSS from the FANTOM5 CAGE data. E, Tags per million at the different RAB27A TSS during differentiation of human embryonic stem cells into pigmented melanocytes (d, day). CAGE, Cap analysis gene expression.

The clinical resemblance and geographic proximity prompted analysis of RAB27A by MLPA in P3 and P4, patients who displayed defective lymphocyte exocytosis yet lacked a genetic diagnosis. Indeed, MLPA revealed homozygosity for the complex SV in both patients (Fig 2, C), as confirmed through sequencing of a breakpoint (Fig E3, A). Rab27a protein expression was absent in leukocytes from P3, P4, and P5 (Fig 2, D, and Fig E2, G and H). Thus, all 5 patients carried biallelic RAB27A mutations, and were at least heterozygous for the SV affecting the transcript NM_183235.2.

Eight patients with GS2 and normal pigmentation have been reported to date, all with missense mutations that selectively disrupt binding of Rab27a to Munc13-4 but not to melanophilin, explaining defective lymphocyte yet normal melanocyte function.5, 6 Because the complex SV disrupts only 1 of several TSSs of RAB27A, we hypothesized that lymphocytes and melanocytes might selectively use distinct RAB27A TSSs. Quantitative PCR demonstrated diminished expression of transcript NM_183235.2 in peripheral blood leukocytes from P3 (Fig 2, E), suggesting predominant usage of the TSS for this transcript. This observation is also supported by cap analysis gene expression data from the FANTOM5 project (Fig E3, C-E). In contrast, primary as well as embryonic stem cell–derived melanocytes use alternative, downstream TSSs, which were not disrupted by the complex SV (Fig E3, D and E). Quantitative PCR in primary melanocytes confirmed a smaller contribution of transcript NM_183235.2 to total RAB27A expression (Fig 2, E). Transcription of PIGB was maintained by the complex SV (Fig E3, B). Notably, an individual with a homozygous deletion encompassing all RAB27A TSSs displayed classic silvery hair. Our data therefore indicate differential TSS usage between leukocytes and melanocytes, explaining the normal pigmentation observed in patients with at least 1 mutation affecting only the upstream RAB27A TSS.

Nonsense mutations in RAB27A are associated with the development of HLH within the first year of life. Neurologic involvement affects 55% of patients at diagnosis and 67% during the course of disease. Three of the 5 patients reported here displayed severe and recurrent neuroinflammation resembling acute disseminated encephalomyelitis, which preceded onset of full-blown HLH by many years. Functional assays of NK- and T-cell exocytosis should therefore be considered in the diagnostic workup of patients with unexplained neuroinflammatory diseases. Late-onset HLH suggests downstream transcription in activated lymphocytes.

In conclusion, we report 5 patients with atypical GS2 characterized by neuroinflammation, lymphoma, and late-onset HLH. Remarkably, 3 patients manifested skin granulomas. Our results elucidate novel structural aberrations affecting the noncoding region of RAB27A, linking the lack of hypopigmentation to differential RAB27A TSS usage between lymphocytes and melanocytes. The identification of a recurrent complex SV in RAB27A suggests a founder effect in the Baltic population. Assessment of this aberration should be included in the genetic workup of patients with defective exocytosis from this area.

Table E2

MLPA probes

Gene	Size (bp)	Used for Figure 2, C
GABRA4_fam-pilot	84	Control probe
RAB27A_probe5	87	Probe C in Figure 2, C
RAB27A_probe6	90
RAB27A_probe7	93
RAB27A_probe8	96	Probe D in Figure 2, C
RAB27A_probe4	99	Probe B in Figure 2, C
RAB27A_probe3	102
Stern PCLN13q	105	Control probe
RAB27A_probe10	108	Probe E in Figure 2, C
RAB27A_probe9	111
RAB27A_probe1	114	Probe A in Figure 2, C
PIGB_probe11	117	Probe F in Figure 2, C
RB1ex23	129	Control probe
MRPL41_e1_132	132	Control probe
SRY	135	Control probe
GABRA4_fam-pilot_A	gggttccctaagggttggaCAGCCTGTTGTCATAACCATCG
GABRA4_fam-pilot_B	/5Phos/AGCAAACTGTCCAGGATGCGtctagattggatcttgctggcac
Stern PCLN13q_A	gggttccctaagggttggaGACACAAGGGTGTAAAATGCACG
Stern PCLN13q_B	/5Phos/TTTCAGGGTGTGTTTGCATATGATTTAATCAATCAGTATGtctagattggatcttgctggcac
RB1ex23_A	gggttccctaagggttggaGTCACCAATACCTCACATTCCTCGAAGCCCTTACAAGTTTCCT
RB1ex23_B	/5Phos/AGTTCACCCTTACGGATTCCTGGAGGGAACATCTATATTTCACCtctagattggatcttgctggcac
MRPL41_e1_132_A	gggttccctaagggttggaGACCCTGACAACCTGGAAAAGTACGGCTTCGAGCCCACACAGGAG
MRPL41_e1_132_B	/5Phos/GGAAAGCTCTTCCAGCTCTACCCCAGGAACTTCCTGCGCTAGCTGtctagattggatcttgctggcac
SRY_A	gggttccctaagggttggaCAGTGCAAAGGAAGGAAGAGCTTCTCCGGAGAGCGGGAATATTCT
SRY_B	/5Phos/CTTGCACAGCTGGACTGTAATCATCGCTGTTGAATACGCTTAACATAGtctagattggatcttgctggcac
RAB27A_p5_A	gggttccctaagggttggaGGACTTCAGGAAGCTGCAATGTTT
RAB27A_p5_B	/5Phos/GCTTTTGTGAATTTCCTTCCCtctagattggatcttgctggcac
RAB27A_p6_A	gggttccctaagggttggaCACAGAACCTCAGAGAAGCCTTGA
RAB27A_p6_B	/5Phos/GGGGCAACTGGTCAACCAATTAGGtctagattggatcttgctggcac
RAB27A_p7_A	gggttccctaagggttggaCAGACCTCGTGGTTCCCACTCAGA
RAB27A_p7_B	/5Phos/GGCAGCCAAGTTCATCCCTCTCAGTGGtctagattggatcttgctggcac
RAB27A_p8_A	gggttccctaagggttggaGCCTTCTAAGCCGTCTCGCTGACTTGT
RAB27A_p8_B	/5Phos/GTCTACCTCCACCGCAAATTCCAGCTGtctagattggatcttgctggcac
RAB27A_p4_A	gggttccctaagggttggaGATCCAACTGCTCCCTTCAAGAAGTT
RAB27A_p4_B	/5Phos/GGTAATTAGGGTGAGGTGGAATGATGTACTCtctagattggatcttgctggcac
RAB27A_p3_A	gggttccctaagggttggaCAACTGGCCAGCTGTCACTCAAATGCTAATT
RAB27A_p3_B	/5Phos/GTGTCTATCATCTGCTTTCTCTAATAGCCtctagattggatcttgctggcac
RAB27A_p10_A	gggttccctaagggttggaGTGGCAAGATGGGTGGTAAGTCCTAAATACTTTA
RAB27A_p10_B	/5Phos/GAAGCTGTATGCCAGTTATTTCGTTCCTATGGtctagattggatcttgctggcac
RAB27A_p9_A	gggttccctaagggttggaGCTTTAAGAATGGTGTGGAGGGACCAGAGGTCACTA
RAB27A_p9_B	/5Phos/CTGTGCCTTACAAGGAGCCAACCAGAGCAGCAGtctagattggatcttgctggcac
RAB27A_p1_A	gggttccctaagggttggaGAGGCATGACCATTTGATCGCACCACTCCTTCAGGAAT
RAB27A_p1_B	/5Phos/CCAGGACTTGTCCACACACCGTTCCATTCGCTTCtctagattggatcttgctggcac
PIGB_p_11_A	gggttccctaagggttggaGAGGACCATTTATGTTTCCGGAACAGAATACCAATGCTA
PIGB_p_11_B	/5Phos/CAGAATGTTGAGTCCCCCTACTGACCTACTTCCCTCtctagattggatcttgctggcac