Autophagy-associated immune dysregulation and hyperplasia in a patient with compound heterozygous mutations in ATG9A

ABSTRACT

Autophagy involving core machinery proteins such as ATG9A is part of a cytoprotective process whose dysregulation is associated with carcinogenesis, neurodegeneration and autoimmunity, but few examples exist of monogenic diseases to further insights into human disease. In the present studies, we report a patient with novel compound heterozygous mutations in ATG9A after patient and parental DNA were subjected to whole exome sequencing. The patient developed hyperplastic proliferations of T and B cells in lung and brain and exhibited defects in lymphocyte memory cell populations after developing an infection with Epstein-Barr virus (EBV). Peripheral blood leukocytes from the patient exhibited defects in autophagic activity and EBV-transformed cells redemonstrated this defect which was also associated with defective NFKB (nuclear factor kappa B) signaling and accumulation of the pro-proliferative protein MYC (MYC proto-oncogene, bHLH transcription factor) and anti-apoptotic protein BCL2 (BCL2 apoptosis regulator), leading to increased proliferative capacity. Defects were corrected after transformation with plasmids expressing ATG9A and after treatment of cells as well as the patient with the MTOR (mechanistic target of rapamycin kinase) inhibitor, autophagy-inducing immunosuppressant rapamycin. These results point to a novel role of ATG9A and autophagy in human cellular signaling associated with hyperplasia and lymphocyte biology and provide an example how genetic studies may suggest effective specific therapeutic interventions.

Abbreviations

BCL2: BCL2 apoptosis regulator; BCL10: BCL10 immune signaling adaptor; CARD11: caspase recruitment domain family member 11; CBM: CARD11-BCL10-MALT1; CR2: complement C3d receptor 2; EBNA: Epstein Barr nuclear antigen; EBV: Epstein-Barr virus; FCGR3A; Fc gamma receptor IIIa; GLILD: granulomatous-lymphocytic interstitial lung disease; HV: healthy volunteer; IKBKB/IKB kinase: inhibitor of nuclear factor kappa B kinase subunit beta; IL2RA: interleukin 2 receptor subunit alpha; MALT1: MALT1 paracaspase; MS4A1: membrane spanning 4-domain A1; MTOR: mechanistic target of rapamycin kinase; MYC: MYC proto-oncogene, bHLH: transcription factor; NCAM1: neural cell adhesion molecule 1; NFKB: nuclear factor kappa B; NIAID: National Institute of Allergy and Infectious Diseases; NK: natural killer; PTPRC: protein tyrosine phosphatase receptor type C; SELL: selectin L; PBMCs: peripheral blood mononuclear cells; TR: T cell receptor; Tregs: regulatory T cells; WT: wild-type

Introduction

Autophagy was first mechanistically described in yeast as a pro-survival mechanism during nutrient stress [1]. This highly conserved cytoprotective process targets bulk cytoplasm as well as damaged organelles and intracellular debris or pathogens and recycles degraded building blocks back into functioning materials. In humans, autophagy dysfunction is associated with a large spectrum of diseases. Reports of common genetic variants have included those affecting the central nervous system (PINK1, PRKN, SQSTM1), lung (ATG5), gastrointestinal system (ATG16L1, IRGM) and bone (SQSTM1), as well as systemic diseases involving cancer (BECN1, UVRAG) and metabolic syndromes (BECN1) [2]. Perhaps because of the potential lethality due to the large number of phenotypes involved, few monogenetic syndromes involving autophagy genes have been reported. These include rare alleles in genes such as EPG5 contributing to autophagosome maturation and degradation that have been associated with Vici syndrome, neurodevelopmental and immune defects including developmental delay, and recurrent respiratory infections, as well as chronic mucocutaneous candidiasis [3]. Recently, a homozygous missense mutation in the ATG5 gene was associated with a recessive form of spinocerebellar ataxia [4], and mutations in ATG7 were associated with neurodevelopmental disorders [5]. However, little is known about how human autophagy-related monogenic deficits affect hematopoietic cells.

Immune modulation by autophagy has been described in both innate and adaptive immune processes [6,7]. As examples, macrophages and neutrophils neutralize invading pathogens after phagocytosis by autophagic-like processes termed xenophagy [8], MAP1LC3/LC3-associated phagocytosis (LAP), or sequestosome-like receptor recruitment [9]. In innate immunity, autophagy plays a role in inflammasome-mediated maturation of cytokines. As an example, chronic granulomatous disease is associated with reduced reactive oxygen species and multiple infections as well as an inflammatory colitis which can be successfully treated with the IL1 (interleukin 1) receptor antagonist, anakinra [10,11]. Interestingly, the increased IL1B/IL-1β production in chronic granulomatous disease was linked to a reduction in autophagy that also resulted in defects in killing of internalized bacteria and fungi [12], demonstrating a link between autoimmunity, autophagy, and infectious disease. In adaptive immunity, autophagy facilitates effective major histocompatibility complex (MHC) presentation to the TR (T cell receptor) for T-cell activation [13], both serving to clear pathogens and residual antigens. Engagement of antigen-bound MHC proteins on the surface of antigen-presenting cells by the TR results in the activation of NFKB (nuclear factor kappa B) transcriptional pathway and has been shown to be dependent on autophagy in mice [14]. NFKB is an important transcription factor in immunity and has major roles in T-cell development and functional divergence of a number of helper T-cell subsets [15]. Autophagy has been reported in mice to facilitate NFKB signaling by formation of a SQSTM1-dependent “POLKADOTS” molecular complex after TR-dependent activation of PRKCQ (protein kinase C theta), which phosphorylates a large adapter protein CARD11/CARMA1 (caspase recruitment domain family member 11) with a small adapter protein BCL10 (BCL10 immune signaling adaptor) and a protease MALT1 (MALT1 paracaspase) [16]. However, less is known about how this pathway participates in human disease.

In the present studies, we describe a patient with novel compound heterozygous mutations in ATG9A, encoding a core autophagy protein that directs formation of the primordial autophagosomal structure [17]. The patient was healthy until age 11 when he developed nonmalignant T- and B-cell proliferative foci in the lung and multifocal white matter lesions in the brain accompanied by pulmonary and neurological symptoms following a viremic infection with Epstein-Barr virus (EBV). The studies undertaken represent the first report of a monogenic autophagy defect with immune dysregulation reflected in pathological proliferation and defective differentiation of T cells and abnormal responses to EBV infection as mechanisms underlying the clinical manifestations. These results facilitated successful clinical therapy with the MTOR (mechanistic target of rapamycin kinase) inhibitor, rapamycin (drug name sirolimus).

RESULTS

Progressive nodular infiltrates in lung and brain and deficits in T-cell memory subsets are associated with mutations in ATG9A

The patient, an 11-year-old male, presented with progressive pulmonary symptoms unresponsive to antibiotics and corticosteroids. Chest CT demonstrated nodular pulmonary infiltrates (Figure 1(a)). Initial BAL (bronchoalveolar lavage) demonstrated lymphocytosis, with an early predominance of CD4⁺ and CD8⁺ T lymphocytes (Figure S1(a)). The patient’s EBV serology 4 and 6 months after presentation suggested a recent EBV infection with active viral replication (measurable IgG to EBV early antigen, Table S1). At the time of worsening GLILD, the patient demonstrated EBV in the peripheral blood by PCR. The patient underwent VATS (video-assisted thoracoscopic surgery) biopsies of the left upper and lower lobes, which also showed multifocal nodular infiltrates with lymphocytes and histiocytes, and formation of non-necrotizing granulomas with occasional multinucleated giant cells consistent with granulomatous and lymphocytic interstitial lung disease (GLILD; Figure 1(b), i–iii). Lymphogranulomatous infiltrates were located around airways, blood vessels, and pleura (Figure S1(b)B, i–iii). Rare cells stained positive for EBV by in situ hybridization (Figure 1(b), iv). As in the BAL, a predominance of CD3⁺, particularly CD4⁺ T cells was found in pulmonary infiltrates by pathology of biopsy tissue, but CD8⁺ T cells, B cells, and plasma cells were also present (Figure 1(b), v–viii). There was no evidence for the presence of a clonal T-cell population as demonstrated by a polyclonal distribution of TRG (T cell receptor gamma) gene rearrangements visualized following capillary electrophoresis of multiplexed PCR of DNA (Figure 1(c)) [18]. Punctate foci and areas of T2 hyperintensity and enhancement in the middle cerebellar peduncles and posterior body of the corpus callosum were also evident on brain MRI during an exacerbation (Figure S1(c)). Repeated flow cytometric analyses of peripheral blood demonstrated markedly reduced numbers of PTPRC/CD45RA⁻ memory cells of either total memory (SELL/CD62L⁺) or effector memory (SELL/CD62L⁻) CD4⁺ or CD8⁺ cells, and an increase in PTPRC/CD45RA⁺ CD4⁺ or PTPRC/CD45RA⁺ CD8⁺ T cells (Table 1, Figure S2(a)). There was evidence of enhanced thymic T-cell output based on the percentage of recent thymic emigrant T cells (PECAM1/CD31⁺; Table 1, Figure S2(a)) without signs of T-cell exhaustion based on baseline expression of LAG3 and PDCD1/PD1 in CD4⁺ and CD8⁺ T cells or after in vitro TR stimulation with anti-CD3 and anti-CD28, the latter making reduced in vivo survival post activation/ differentiation less likely (Figure S2(b)). The percentage of regulatory T cells (FOXP3⁺) was within the normal range (Table 1). T-cell phenotyping demonstrated overall decreased total numbers of CD3⁺, CD4⁺, and CD8⁺ T cells but normal proliferation to mitogens and antigens (Table 1, Table S2). Changes in the B-cell compartment, notably a reduction in B-cell numbers with a decrease in switched memory B cells, and low natural killer (NK) cells were seen (Table 1). Antibody studies including vaccine responses were within the normal reference range (Table S2). The clinical course is summarized in the Clinical Supplement (Supplemental materials) including description of a sustained improvement in the pulmonary and brain nodules after treatment with rapamycin (drug name sirolimus; Figure S1(d); Clinical Supplement).

Figure 1.

Patient with compound heterozygous mutations of ATG9A presents with interstitial lung disease (ILD) resembling granulomatous-lymphocytic ILD (GLILD). (A) Chest CT at presentation shows multifocal nodular infiltrates. (B) Multifocal nodular infiltrates on lung biopsy specimens (left upper and lower lobe wedge biopsies) at initial presentation (hematoxylin/eosin stain, panels i–iii; immunohistochemistry, panels v–viii), demonstrating lymphocytes, histiocytes, and absence of germinal centers (I), non-necrotizing granulomas (ii) with occasional multinucleate Langerhans-type giant cells (black arrows) (iii), rare cells positive for EBV by in situ hybridization (ISH, black arrows) (iv). Predominance of CD4⁺ T cells (V) and CD3⁺ T cells (vi), some CD8⁺ T cells (vii), and plasma cells (viii). Scale bar: 500 μm (Bi), 200 μm (Bv, vii, viii), 50 μm (Bii, iii, iv). (C) Tissue T cells (TRG) exhibit polyclonality by T-cell receptor gene rearrangement studies. (D) Compound heterozygosity of the patient for 2 novel variants in the ATG9A gene detected by whole exome sequencing and confirmed by Sanger sequencing, the I387V mutation inherited from the mother and the T518I mutation from the father.

Table 1.

Lymphocyte phenotyping during the course of the disease.

Time since presentation	+ 6 mo	+ 1 y 5 mo	+ 2 y 4 mo	+ 3 y 10 mo	+ 4 y 4 mo	+ 6 y 6 mo	Reference range adjusted for age (* based on adult controls)
Therapy
IVIG	X	X	X	X	X	X
Systemic steroids	X	X	X	X		X
Rapamycin				X		X
Lymphocyte phenotyping							Cells/μL	(%)
Total T cells (CD3)	684 (75.5)	876 (86.1)	1437 (93.9)	391 (85.1)	553 (85)	702 (85.6)	1200–2600	60–76
T-cell subsets
CD4+ CD3+	385 (42.5)	529 (52)	1034 (67.6)	264 (57.5)	419 (64.5)	471 (57.4)	650–1500	31–47
CD8+ CD3+	264 (29.1)	305 (30)	366 (23.9)	121 (26.3)	124 (19.1)	211 (25.7)	370–1100	18–35
CD4:CD8 ratio	1.46	1.73	2.83	2.19	3.38	2.23		1.2–6.2
CD3+ CD4− CD8−		35 (2.3)				17 (2.1)	24–223*	1.3–9.2*
CD4+ CD3+ PTPRC/CD45RA+		81.8%				62.9%		9.5–41.9
CD4+ CD3+ PTPRC/CD45RO+		17.6%				37%		16.5–42.2
CD4+ CD3+ SELL/CD62L+ CD45RA+ CD45RA+			936 (61.2)			290 (35.4)	102–1041*	7.6–37.7*
CD4+ CD3+ SELL/CD62L+ PTPRC/CD45RA−			87 (5.7)			112 (13.6)	162–614*	10.4–30.7*
CD4+ CD3+ SELL/CD62L− PTPRC/CD45RA−			11 (0.7)			63 (7.7%)	42–225*	2.3–15.6*
CD4+ CD3+ SELL/CD62L− PTPRC/CD45RA+			0 (0)			6 (0.7)	0–29*	0–1.5%
CD8+ CD3+ SELL/CD62L+ PTPRC/CD45RA+			356 (23.3)			129 (15.7)	85–568*	5.7–19.7*
CD8+ CD3+ SELL/CD62L+ PTPRC/CD45RA−			6 (0.4)			3 (0.4)	25–180*	1.5–10.3*
CD8+ CD3+ SELL/CD62L− PTPRC/CD45RA−			2 (0.1)			29 (3.5)	24–175*	1.1–9.2*
CD8+ CD3+ SELL/CD62L− PTPRC/CD45RA+			2 (0.1)			49 (6)	11–172*	0.7–7.8*
CD4+ CD3+ PTPRC/CD45RA+ PECAM/CD31+			799 (52.2)			249 (30.4)	34–426*	1.6–20.2*
CD4+ CD3+ PTPRC/CD45RA− CXCR5+			11 (0.7)			12 (1.5%)	35–172*	1.8–8.9%
CD4+ CD3+ IL2RA/CD25+ FOXP3+			4.9%			5.1%
Total B cells
CD19+	182 (20.1)	102 (10)	77 (5)	23 (5)	77 (11.8)	39 (4.8)	270–860	13–27
MS4A1/CD20+			77(5)			39 (4.8)	59–329*	3–19*
B-cell subsets
MS4A1/CD20+ CD27+			2 (0.1)			2 (0.3)	12–68*	0.8–3.6*
MS4A1/CD20+ CD38+			75 (4.9)			38 (4.7)	30–282*	1.2–17.6*
MS4A1/CD20+ CD27+ IgM+			2 (0.1)			1 (0.1)	6–50*	0.3–2.5*
MS4A1/CD20+ CD27+ IgM−			0 (0)			1 (0.1)	6–34*	0.3–2.2*
MS4A1/CD20+ CD27− IgD+	85.3%					90%		30–70
MS4A1/CD20+ CD27+ IgD+	9%					4%		8–35
MS4A1/CD20+ CD27+ IgD−	4.5%					3%		8–42
MS4A1/CD20+ or CD19+ CD38low CR2/CD21low	0.8%		1 (0.1)			1 (0.1)		0.5–8
MS4A1/CD20+ CD38high IgMlow	0.9%		0 (0)			0 (0)		0.1–4.7
NK cells
FCGR3A/CD16+ or NCAM1/CD56+ CD3−	29 (3.2)	17 (1.7)	9 (0.6)	45 (9.8)	6 (0.9)	79 (9.6)	100–480	4–17
FCGR3A/CD16+ or NCAM1/CD56+ CD3+			81 (5.3)			30 (3.7)	29–299*	2.2–12.4*

At times of worsening interstitial lung disease with accumulation of CD4⁺ T cells in the lung, the patient demonstrated an increased percentage of predominantly CD4⁺ T cells in the peripheral blood (2 years-4 months and 4 years-4 months after initial presentation), increased thymic T cell output (CD3⁺ CD4⁺ CD45RA⁺ PECAM1/CD31⁺), an increase in naive T cells (CD4⁺ CD3⁺ PTPRC/CD45RA⁺; CD4⁺ and CD8⁺ CD3⁺ SELL/CD62L⁺ PTPRC/CD45RA⁺), but decreases in memory T cells (CD4⁺ and CD8⁺ CD3⁺ SELL/CD62L⁺ or SELL/CD62L⁻ PTPRC/CD45RA⁻), and associated changes in the B-cell compartment; decrease in immature B cells (MS4A1/CD20⁺ or CD19⁺ CD38^low CR2/CD21^low), decrease in plasmablasts (MS4A1/CD20⁺ CD38^high IgM^low), decrease in class-switched memory B cells (MS4A1/CD20⁺ CD27⁺ IgD⁻), increase in naive B cells (MS4A1/CD20⁺ CD27⁻ IgD⁺), and decrease in NK cells.

Rare compound heterozygous mutations in ATG9A were identified by whole exome sequencing; I387V (c.1159 A > G) and T518I (c.1553 C > T), the first inherited from the mother and the latter from the father. Both mutations were confirmed by Sanger sequencing (Figure 1(d)). The maternal variant, I387V (c.1159 A > G), occurs in a transmembrane domain [19]. The specific residue is not conserved (Consurf score 1 out of 9 [20]). The CADD score is intermediate at 18.2 [21]. In gnomAD, 4 identical alleles were reported in subjects of European descent, giving an allele frequency of 0.00001604 [22]. The 387^th amino acid residue lies within the 14^th alpha helix, which according to the solved EM structure forms the interaction face between monomers in the homotrimer. This intermolecular interaction occurs within the lipid bilayer and is facilitated by hydrophobic bonds. From the structure, it is not clear how important the contribution of isoleucine 387 would be. Moreover, substituting a valine would likely be tolerated if the role is to form hydrophobic bonds. Based on the above, the evidence was not overwhelming that this was a pathogenic variant; however, it could have an unpredictable effect. The paternal variant, T518I (c.1553 C > T), occurs in the c-terminal cytoplasmic domain. The amino acid residue is highly conserved (Consurf score 9 out of 9 [20]). The CADD score is 23.9, signifying that it is likely damaging [21]. In gnomAD, 6 identical alleles are reported, all in European subjects, giving an allele frequency of 0.00002139 [22].

Pathogenic variants in other genes associated with immune dysregulation, lymphoproliferation, and increased susceptibility to infections such as those causing autoimmune lymphoproliferative syndrome, PIK3CD, LRBA, and CTLA4 were not detected, the latter confirmed by Sanger sequencing.

Dominant negative compound heterozygous ATG9A mutations result in defective autophagy

To determine whether the mutant alleles of ATG9A were associated with defective autophagic activity, flow cytometry analyses using peripheral blood mononuclear cells (PBMCs) from the patient were conducted. In comparison to a healthy control, the patient demonstrated accumulation of SQSTM1 by flow cytometry (Figure 2(a)), an autophagic adaptor protein associated with autophagy activity. Since blood sampling in this patient was limited by his small weight, further studies were conducted with EBV-transformed B cells which confirmed the increased accumulation of SQSTM1 by western blot analysis (Figure 2(b)) as well as in additional healthy volunteers (HV; Figure 2(c)) [23]. Cell clones including HV2 (reported as “HV” in Figure 2(d) and in subsequent figures) were selected that demonstrated equivalent expression of EBV as measured by levels of EBNA1 for further study (Figure 2(c)). In the patient, equivalent expression of ATG9A protein as well as ATG9A and SQSTM1 mRNA (Figure 2(d,e)) was observed in the EBV-transfected cells. After treatment with the inhibitor of autophagosome-lysosome fusion and lysosomal degradation bafilomycin A₁ (BafA), lipidated MAP1LC3B/LC3B (microtubule associated protein 1 light chain 3; LC3B-II) required for effective autophagosome formation and SQSTM1 was reduced in EBV-transformed patient cells (Figure 2(f)) [23]. In total, these data suggest that aberrant ATG9A protein in the patient cells reduced overall autophagic activity. Albeit to a lower degree, retention of LC3B-II formation in the father suggested that a predominant contribution to the patient’s overall ATG9A-related phenotype may have been from the T518I mutation of this family member, thus leading to a dominant negative contribution from this allele. Indeed, complementation of ATG9A deletant mutant HeLa cells with plasmids expressing green fluorescent-tagged mutant alleles appeared to rescue the ratio of lipidated to non-lipidated MAP1LC3B (LC3-I:LC3-II) in the KO background as well as reduced LC3-positive puncta size (Figure S3) [19,24]. LC3-positive puncta size is a well-established marker of autophagy activity between mammalian wild-type (WT) and ATG9A deletant mutants [24–27]. In addition, to characterize the interactions of the maternal and paternal alleles, studies were conducted, which demonstrated that while complementation of HV or maternal EBV-transformed cells with plasmids expressing healthy volunteer (WT) ATG9A-GFP resulted in maintenance of autophagic flux, complementation of either HV or maternal cells with the paternal ATG9A-GFP T518 mutant allele resulted in significant reductions in autophagic flux, demonstrated by increased accumulation of SQSTM1 (Figure 2(f)), suggesting a moderate dominant negative contribution of the paternal allele. Furthermore, complementation of paternal cells with maternal ATG9A^I387V-GFP failed to reconstitute autophagic flux in the paternal cells, resulting in increased accumulation of SQSTM1 compared to that after equivalent WT ATG9A-GFP complementation (Figure 2(g)), suggesting a failure of the maternal allele to rescue the paternal allele’s defective function, consistent with a small negative contribution of the maternal allele.

Figure 2.

Compound heterozygous ATG9A mutations lead to defects in autophagy. (A) PBMC-derived lymphocytes of the patient (blue) or a healthy volunteer (HV1, red) were subjected to flow cytometry assessing the expression of SQSTM1. (B) Cell lysate SDS-PAGE of EBV-transformed B-cells from a healthy volunteer (HV2), patient (PT) or parents (mother = Mo or father = Fa) were subjected to western blotting utilizing the indicated antibodies. (C) Cell lysate SDS-PAGE of EBV-transformed B-cells from a HV2 or 3 additional healthy volunteers (HV3, HV4, HV5) or patient (PT) were subjected to western blotting utilizing the indicated antibodies. All subsequent designations of HV2 will be referred to as “HV” for brevity. (D) RNA from cells identical to Figure 2B was isolated and assessed by qRT-PCR to measure ATG9A or (E) SQSTM1 gene expression. (F) Western blots of the indicated cell lysate after treatment of cells for 2 hours with 1 μM bafilomycin A₁ (+) on no treatment (-) using the indicated antibodies. (G) Indicated EBV-transformed cells were transfected with plasmids expressing ATG9A containing the indicated allele (WT = wild-type; Mo = ATG9A-GFP I387V; Fa = ATG9A-GFP T518) or empty vector alone (EV), recovered for 2 days and analyzed by western blot using the indicated antibodies. Error bars indicate s.d., n = 3 (Figure 2(bf)) independent experiments. *P < 0.05, **P < 0.01, NS = non-significant.

Reduced function of ATG9A is associated with deficits in NFKB signaling

T cells obtained from the patient did not appear to exhibit evidence of immunological exhaustion (Figure S2(b)). We therefore posited that deficits in T-cell receptor signaling mediated through the NFKB pathway, previously associated with autophagic function in mice and human cells [16,28], may have contributed to the patient’s preponderance of naive T-cells (Table 1). As shown in Figure S4(a), siRNA knockdown of ATG9A in Jurkat T cells resulted in significantly decreased expression of ATG9A mRNA and protein, and accumulation of SQSTM1 (Figure S4(a,b)). ATG9A suppression also resulted in statistically reduced SQSTM1-dependent recruitment of BCL10 after stimulation with anti-CD3 and anti-CD28 by co-immunoprecipitation (Figure S4(c)). BCL10 is a key member of CBM (CARD11-BCL10-MALT1) complex involved in the activation of the inhibitor of IKK (inhibitor of NFKB kinase) complex required for NFKB signaling. Interaction of BCL10 with SQSTM1 within protein “polkadot” puncta was recently found to be required for effective IKK phosphorylation and NFKB activation [29]. Commensurate with this mechanism, ATG9A siRNA transfection of Jurkat cells resulted in reduced activation of IKK, as evidenced by reduced phosphorylation of CHUK/IKK1/IKKA (component of inhibitor of NFKB kinase complex):IKBKB/IKK2 (inhibitor of NFKB kinase subunit beta) or “CHUK-IKBKB” in response to TR stimulation with anti-CD3 and anti-CD28 compared to non-targeted, control treated Jurkat cells (Figure S4(d)). Furthermore, knock-down of ATG9A reduced expression of NFKB downstream target transcripts including IRF1, MYD88, RIPK1, TBK1, and TRAF6 after TR after TR stimulation (Figure S4(e)) [30,31]. Some reductions in NFKB-dependent genes were evident after ATG9A suppression, which could have been due to unopposed suppressive effects in the absence of NFKB-antigen dependent activation as described previously for the CTLA4 inhibitor pathway [32].

These findings were confirmed in cells from EBV-transformed B-cells from a healthy volunteer (HV), which demonstrated induction of p-CHUK:IKBKB in response to stimulation by phorbol myristate acetate (PMA)/ionomycin (rather than anti-CD3 and anti-CD28 because of the absence of TR signaling in B-cells) whereas this pathway could not be induced in cells from the patient after stimulation unless recombinant ATG9A was supplemented (Figure 3(a)). Dynamic range of EBV-transfected cells was less, perhaps due to EBV induction of gene products such as PDLIM7/LMP1, an inducer of NFKB [33]. Following reconstitution of recombinant ATG9A in patient cells using a vector plasmid expressing ATG9A, autophagy activity was induced, as demonstrated by reduction of SQSTM1 (Figure 3(b)) and induction of p-CHUK:IKBKB (Figure 3(a)). Reduced signaling in the patient was associated with decreased expression of the same NFKB downstream targets as observed in Jurkat cells after ATG9A siRNA suppression (Figure 3(c)). Moreover, expression of NFKB downstream targets was restored after transfection using a vector plasmid expressing ATG9A, compared to empty vector controls (Figure 3(d)).

Figure 3.

Defects in NFKB signaling are restored after complementation with plasmids expressing recombinant ATG9A in patient cells with compound heterozygous ATG9A mutations. (A) Phosphorylated CHUK/IKK1/IKKA/ IKKα/β (p-CHUK) western blots of cell lysates from EBV-transformed patient cells after transfection of cells from a healthy volunteer (HV) or cells from the patient (PT) with a vector expressing either ATG9A (+ATG9A) or an empty vector control (EV) before (-) or after (+) stimulation with PMA-ionomycin (P/I). (B) Western blots of cell lysates from EBV-transformed PT cells after transfection with a vector expressing either EV or +ATG9A, probed with antibodies against the indicated proteins. (C) RNA was isolated from EBV transformed cells from HV or PT harvested either before (-) or after (+) stimulation with P/I and was subjected to qRT-PCR of indicated NFKB signal targets, normalized to RNA18S/18S rRNA. (D) qRT-PCR of indicated NFKB target genes using RNA of indicated individuals expressing either empty vector control (EV) or an ATG9A expression vector (+ATG9A). Cells were harvested after stimulation with P/I. Error bars indicate s.d., n = 3 independent experiments. Student’s t-test, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, NS = non-significant.

Patient cells demonstrate increased proliferation in cell culture related to increased expression of BCL2 and MYC

In order to characterize the proliferative capacity of patient cells that may have contributed to the propensity to form nonmalignant lymphocyte proliferations in tissue, we followed growth rates of cultured EBV-transformed cells. As shown in Figure 4(a), patient cells demonstrated increased proliferative capacity over six days compared to EBV-transformed healthy volunteer or parental cells. Of note, closer similarity in proliferation between cells from the patient and the father suggested a predominant contribution of their shared mutation and a trend toward increased proliferation in the mother. Similar to effects exhibited by NFKB signaling, complementation of patient cells with ATG9A reduced proliferative capacity, compared to patient cells transformed with a control vector (Figure 4(b)). Patient (and father’s) cells also overexpressed the antiapoptotic protein BCL2 (BCL2 apoptosis regulator) and MYC (MYC proto-oncogene, bHLH transcription factor) [34], previously shown to be degraded by an autophagy-dependent mechanism [35] (Figure 4(c)). Suppression of MYC utilizing siRNA reduced proliferative capacity compared to equivalent cells treated with non-targeted control siRNA (Figure 4(d)). Furthermore, treatment of cells with the MTOR inhibitor and autophagic stimulant, rapamycin led to increased turnover of SQSTM1 (Figure 4(e)) and reduced proliferative capacity of patient cells (Figure 4(f)). However, the EBV-inducible growth factor TNFAIP3 (TNF alpha induced protein 3), which is associated with cellular proliferation [36,37], was not increased in patient cells suggesting no significant contribution of this factor to proliferation in the patient cells (Figure S4F). In addition, as shown earlier, reductions of NFKB signaling were noted in patient cells, suggesting that the TNFAIP3 pathway was unlikely to be related to the patient’s hyperproliferative capacity. Treatment of the patient with rapamycin (sirolimus) resulted in successful control of pulmonary hyperplasia and symptoms both initially and on re-challenge after a 5-month washout period, Figure S1D) as well as stabilization of findings on brain MRI (Figure S1(c)) corresponding with the ex-vivo findings.

Figure 4.

Patient cells with compound heterozygous mutations of ATG9A demonstrate accelerated growth characteristics, which are reduced after complementation with plasmids expressing recombinant ATG9A. (A) Growth curve of EBV-transformed cells of a healthy volunteer (HV), patient (PT), mother (Mo) or father (Fa). Cells were harvested and replated at 2 × 10⁴ cells/well in 24-well plates in triplicate and quantified daily by hemocytometer. (B) Equivalent growth curves of PT EBV-transformed cells expressing either empty vector control (EV) or vector expressing ATG9A (ATG9A). (C) Western blots of indicated proteins from indicated EBV-transformed cells. (D) Growth curves of EBV-transformed patient cells transfected with MYC (MYC siRNA) or a non-targeted control siRNA (NT). Western blots of indicated cells showing indicated protein expression (E) or growth curves (F) of indicated cells either treated (+) or untreated (-) with 50 nM rapamycin. Error bars indicate ± s.d., n = 5 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, NS = non-significant.

Discussion

A previously healthy 11-year-old male presented with progressive pulmonary symptoms after a respiratory infection and was diagnosed with GLILD. Based on serology testing 4 and 6 months after symptom onset, EBV infection as a participating trigger was suspected. The presence of multi-nucleated giant cells and EBV staining pointed toward modest ongoing tissue infection. The development of GLILD has been described as a noninfectious complication in up to 20% of patients with common variable immune deficiency (CVID) [38], a heterogeneous disorder of impaired humoral immunity with often undetermined culprit gene mutation. LRBA and CTLA4 deficiencies are examples of immune dysregulation as seen in CVID, but with known genetic causes, and presenting with defective antibody production, autoimmunity, lymphoproliferation, and often GLILD [39,40]. Inadequate expression and function of LRBA results in increased and dysregulated recycling of CTLA4 on T cells, which negatively affects the suppressive function of regulatory T cells (Tregs). Mutations in LRBA and CTLA4 were ruled out in the present patient. Humoral immunity appeared intact as serum immunoglobulins levels were normal with mildly elevated IgE. Furthermore, protective antibody levels to H. influenzae, S. pneumoniae, and Varicella were demonstrated. Evidence for immune dysregulation was most obvious in lymphocyte subset numbers and distribution with reductions in numbers of memory T cells, but normal numbers of CD4⁺ FOXP3⁺ Tregs.

Following identification of two missense mutations in ATG9A in the patient, studies were initiated to determine if the clinical findings were the result of defects in autophagy-dependent events. ATG9A in mammalian cells is hypothesized to be a key protein involved in nucleation of membranes essential for autophagosome formation due to its action in yeast as a “scramblase” that translocates phospholipids between outer and inner leaflets of the phagosome assembly complex membrane [41,42]. This might be expected to result in a decrease in autophagosome maturation evidenced by reduced lipidated LC3B and SQSTM1 after autophagosome degradation was normalized using BafA. However, while the yeast ATG9 is incorporated into phagophores at nucleation and remains inserted in the external membrane of the developing autophagosomes until closure [43], the mammalian ATG9A may only associate transiently with phagophores [44]. Further complicating its characterization in mammals, Atg9a mutations in mice are embryonically lethal or lethal within the first week of life [45], thus requiring a partial suppression phenotype as in the present patient. We showed for the first time that heterozygous ATG9A alleles in a patient leads to defective ATG9A function measured by altered levels of the SQSTM1 receptor protein and faulty lipidation of LC3B, both useful as a measure of autophagic activity [23]. Interestingly, while each mutant allele was able to complement ATG9A KO HeLa cells to some extent, interactions with partner proteins within the tripartite structure of the ATG9A complex may have become disrupted between the WT and paternal allele [19], resulting in a dominant negative-induced reduction in autophagic activity. Indeed, insertion of the paternal mutant allele within an altered complex may also have interfered with effective autophagosome degradation as evidenced by accumulation of SQSTM1 in the absence of lysosomal degradation inhibition, although further detailed kinetic studies to further elucidate these effects was felt to be beyond the scope of these studies. Furthermore, the maternal allele, while not as disruptive as the paternal contribution, was not able to reconstitute the function of the ATG9A complex to the same extent as expression of an additional WT copy in the paternal cells, thus exerting a small additional contribution to the overall patient phenotype. These data are thus consistent with the established role for ATG9A in autophagy.

Moreover, the ATG9A protein appears to be intolerant of loss of function mutations [22,31]. The gnomAD probability of being loss-of-function intolerant score (pLI) is 0.97, indicating a high likelihood that haploinsufficiency occurs in the absence of a single allele. In addition, there appears to be intolerance of missense variations, as the missense Z-score was 2.6, indicating less observed rare SNVs than would be expected. No individuals homozygous for loss-of-function variations were identified in the database. A group in Belgium interrogated the region of the paternal variant T518I (c.1553 C > T) in vitro by deleting amino acids 515 through 519 and found that the mutated ATG9A failed to travel through the Golgi apparatus [46]. It was shown that a similar mutant, 516A/519A, maintained an ability to trimerize with either a WT or mutant allele. T518 was not specifically evaluated beyond the 4 amino acid deletion; however, given its high conservation and proximity to these important residues, it is reasonable to predict a dominant negative effect of the paternal mutation.

Of note, there was a lack of neurodevelopmental delay in the patient, which has been associated with a homozygous missense mutation of another autophagy gene, ATG5 [4]. This suggests that clinical manifestations of genetically inherited pathogenic mutations in autophagic components may have unique and distinguishing features. In addition, the father has not developed overt pulmonary symptoms although high resolution CT scanning could not be conducted to assess for asymptomatic lesions due to a lack of approval under the institutional review board’s approved protocol.

Deficits in NFKB signaling were demonstrated in patient cells and, importantly, were restored after complementation of cells with plasmid vectors expressing WT ATG9A. These deficits may be responsible for the reduction in CD4⁺ and CD8⁺ memory cell populations as intact NFKB pathways have been implicated in persistence, function, and phenotypes of the T-cell memory pool [47]. In contrast, we did not see evidence of lymphocyte exhaustion, though our study of blood lymphocytes may have missed exhausted sub-populations which may exhibit functional heterogeneity [48]. Rather, the deficit was traced to aberrant formation of the TR and B-cell receptor (BCL) dependent CBM-complex, recently described in mice as being dependent on autophagy [14]. Additional effects of ATG9A on NFKB signaling could be present including alterations of mobilization and oxidation of fatty acids from mitochondria that has been implicated in NFKB signaling as well as direct effects on cellular dynamics of cell protrusion as well as permeability of the plasma membrane and may be the subject of future study [49–51]. However, the effects of ATG9 on NFKB signaling may provide an important clinical comparator as recently, several patients having homozygous or compound heterozygous mutation in each of the CBM components (the CBM-opathies) have been described with shared phenotypes reflective of inborn errors of immunity [52]. For example, deficits in CARD11, expressed in T and B-cells have associated with normal CD4⁺ and CD8⁺ numbers, but with a predominance of naive T cells and defective NFKB signaling, similar to the present patient [53]. Correspondingly, patients with MALT1 and BCL10 deficits also present with normal numbers of CD4⁺ and CD8⁺ T cells also associated with a predominance of naive T cells and defective NFKB signaling [54,55]. However, the clinical phenotype of the ATG9A deficit was less severe than in the CBM-opathies as the present patient did not have hypogammaglobulinemia and uncontrolled opportunistic viral, bacterial and fungal infections typical of the CBM-opathies, although the patient did have prolonged low copy EBV viremia. This lessened severity compared to the classic CBM-opathies may be related to the modulatory role of autophagy in CBM-related signaling or comparably modest effects on autophagy of the parental mutations.

The present patient also presented with nonmalignant cellular proliferations of lymphocytes which is not typical of the CBM-opathies and could be a direct result of dysfunctional autophagy unrelated to NFKB signaling. Indeed, EBV-transformed B cells from the patient exhibited increased proliferative capacity which was reduced after ATG9A reconstitution. As autophagy has been associated with turnover of the antiapoptotic protein BCL2 and the pro-proliferative transcription factor MYC [56,57], we identified increased protein levels in the patient. In addition, proliferative capacity was reduced after MYC suppression utilizing MYC-specific siRNA and by treatment with the MTOR inhibitor rapamycin. However, it is uncertain whether these studies alone may explain the patient’s lung and brain lymphocytic nodules where tissue proliferation may be influenced by a number of tissue-specific factors [58]. Development of tissue lymphocyte nodules can be enhanced by inflammatory signals such as CXCR3 in response to viral infections [59], which may explain the development of lung and brain nodules after EBV viral infection. EBV is also a well-known agent of tumorigenesis, primarily associated with B-cell transformations [60], but T-cell proliferations more similar to the present case are also known. However, EBV-positive T-cell lymphomas per se tend to exhibit a rapid clinical progression with multiple organ failure, sepsis, and death, often associated with a hemophagocytic syndrome [61]. Less aggressive are the extranodal NK/T cell lymphomas, but these have so far not been associated with autophagic defects, and histologies are frequently ulcerated and necrotic with associated vascular damage, which the present patient did not demonstrate [62]. Thus, we posit that a considerable contribution to disease was the patient’s mutation in ATG9A, given the cellular reversal after ATG9A complementation and disease control with rapamycin. Rapamycin, an MTOR inhibitor, is a potent activator of autophagy [63]. A contribution of EBV to the patient’s condition is suggested by the temporal association of EBV infection with development of the patient’s symptoms. One potential mechanism could be through changes of the microRNA milieu as EBV is known to express over 20 microRNAs that can be transmitted to neighboring cells via exosomes [64], including the MIR196 (microRNA MIR196) family [64,65]; MIR196 can act as an inhibitor of autophagic processes which has previously been described in inflammatory bowel disease and esophageal cancer [66]. Further studies will be required to delineate the exact mechanism related to the role of EBV to the potentiation of the patient’s clinical syndrome including a possible role of microRNA species. Most importantly, the association between reduced autophagy and mutated ATG9A was informative for the clinical treatment strategy as the introduction of rapamycin was effective in the sustained control of the proliferative lesions. Treatment was also associated with a modest improvement in memory cell populations and clearance of low level EBV viremia, though it is difficult to establish causality. These findings represent the first description of significant lymphocyte-related pathology from inherited mutant alleles in a core autophagy gene and provide important insight into the roles of autophagy and autophagy-targeting treatment strategies in human disease.

Study Limitations

There are a few clinical and experimental caveats to keep in mind. As the first report of human mutant alleles in the autophagy-related gene ATG9A, it is important to note that this study is limited by the inclusion of only one patient. Additional information may require the recruitment of additional patients and we hope that this initial report will encourage others to consider mutant alleles in ATG9A in the workup of nonmalignant hyperplasia and syndromes similar to the CARD11-BCL10-MALT1 (CBM)-opathies having deficits in NFKB signaling. Secondly, the focus of deficits was on autophagy-related deficits that could result in the patient’s signs and symptoms; however, additional non-autophagy related phenotypes may be present and may require multiple patients to better identify consistent mechanisms. In addition, the role of EBV as a contributor was difficult to assess and may require several infected and uninfected patient samples to elucidate further. Furthermore, the use of EBV-transformed cells, while necessary because of the patient’s small statue, may benefit from further study with additional adult patients that can provide sufficient volumes of untransformed PBMCs.

Materials and Methods

The patient and parents were followed at National Jewish Health, Denver, CO, and at the National Institutes of Health Clinical Center, Bethesda, MD. Informed consents were obtained under an institutional review board-approved protocol at National Jewish Health, protocol HS-2756, “Characterization of Patients with Immunodeficiency Followed at National Jewish Health – Research Protocol”, and the National Institute of Allergy and Infectious Diseases (NIAID) protocol NCT00404560, “Screening Protocol for Detection and Characterization of Infections and Infection Susceptibility”.

Sanger Sequencing

Sanger sequencing of the ATG9A gene was performed as previously described on an Applied Biosystems 96-capillary 3730XL DNA analyzer (ThermoFisher Scientific, 4,331,246) using the BigDye Terminator v3.1 cycle sequencing kit (ThermoFisher Scientific, 4,337,454) [67]. The following primers (F, forward; R, reverse; Table S3) were used: ATG9AF1 and ATG9AR1 to determine the maternal variant (reference sequence GCTCCATCCTG; Figure 1(d)), and ATG9AF2 and ATG9AR2 to determine the paternal variant (reference sequence AGATACCTGCT; Figure 1(d)).

Flow Cytometry

Flow cytometry of cells from the patient and healthy volunteers was performed on purified peripheral blood mononuclear cells (PBMCs) using indicated antibodies. PBMCs (3 x 10⁶ cells) were obtained from the patient or controls and resuspended at a concentration of 10^6/ml in RPMI-1640 containing 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin in a 12-well plate. The plate was incubated at 37°C and 5% CO₂. After an overnight incubation, cells were harvested and washed twice using ice cold 2% bovine serum albumin (BSA; Alfa Aesar, J64100) in PBS (KD Medical, RGF-3190), and then fixed in fixation buffer for 1 h at room temperature or overnight at 4°C. Once the cells were fixed, cells were washed twice with permeabilization buffer and permeabilized for 30 min at room temperature. Goat serum (10 μl; Abcam, ab7481) was added and incubated for another 30 min at room temperature. Anti-SQSTM/p62 monoclonal antibody (0.5 μg; Abcam, ab56416) was added and a monoclonal IgG2a (Abcam, ab91361) was used as an isotype control. After an hour incubation at room temperature, cells were washed twice with 2% BSA in PBS. As secondary antibody, goat anti-mouse IgG H&L secondary antibody, Alexa Fluor^TM 594 (ThermoFisher Scientific, A11032) was utilized to stain the cells for another hour at room temperature. Cells were washed twice using 2% BSA in PBS and then analyzed by flow cytometry.

Exhaustion T-cell Assays

Exhaustion T-cell assays were performed on patient and healthy volunteer PBMCs by the method of Dong et al. [68]. Briefly, T cells were purified from patient/donor PBMC by EasySep Human T Cell Isolation Kit (STEMCELL Technologies, 17,951). Purified cells were cultured on 96-well plates at 10⁵ cells per well with rh-IL-2 (20 IU/mL) in RPMI-1640 medium (ATCC, 30–2001) containing 10% FBS. Cells were stimulated with Dynabeads Human T-Activator CD3/CD28 (ThermoFisher Scientific, 11131D) then collected at time points: Day 0, Day 1, and Day 3. Cells were fixed with BD Cytofix/Cytoperm (BD Biosciences, 554,714) and stained with anti-PTPRC/CD45, -CD3, -CD4, -CD8, -LAG3 (lymphocyte activating 3), and -PDCD1/PD1 (BD Biosciences, 557,833, 563,800, 557,852, 563,677, 565,616, and 562,516, respectively). Stained cells were analyzed by a BD LSR Fortessa flow cytometer.

Cell Culture of EBV-transformed B Cells, Jurkat, and HeLa Cells

Cells from the healthy volunteer, patient, and family were transformed with EBV [69], and transformed cells expressing equivalent levels of EBV nuclear antigen (EBNA) were selected by western blot using an anti-EBNA 3A antibody (Abcam, ab16126). Cells were maintained in RPMI 1640 with 20% FCS, 2 mmol/L L-glutamine, 100 U/ml penicillin and 100 mg/ml streptomycin at 37°C in a humidified 5% CO₂ incubator.

Jurkat Cells (E6-1) were freshly obtained from the American Type Culture Collection (ATCC, TIB-152) and cultured in RPMI-1640 medium supplemented with 10% FBS and 1% penicillin-streptomycin solution. HeLa cells (ATCC, CCL-2.1) were grown in Dulbecco’s Modified Eagle’s Medium (DMEM [Quality Biological, 112–319-101]) supplemented with 10% fetal bovine serum (Corning, 35–011-CV) and 100 U/mL penicillin – 100 μg/mL streptomycin solution (Corning, 30–002-CI) equaling complete DMEM, at 37^οC and 95% air: 5% CO₂.

Generation of ATG9A Mammalian Plasmids and Transfection

Experimental methods to measure the expression of ATG9A were adapted from previous methods [69]. In brief, Jurkat cells were transfected with plasmid pCMV-GFP (Addgene, 11,153; deposited by Connie Cepko) containing PCR-amplified fragments of ATG9A as described [19]. To generate ATG9A-pCMV-GFP plasmids, the coding region of ATG9A was amplified with the primers ATG9A-S-KPN1 and ATG9A-A-NOT1 using human wild-type cDNA followed by digestion with KpnI and NotI restriction enzymes and purification with a PCR purification kit (Qiagen, 28,104). After ligation with the DNA T4 ligase (Invitrogen/ThermoFisher, IVGN2104), the plasmids were transformed into electro-competent E. coli (Bioline, C3020K). Individual colonies were picked and cultured in LB broth with ampicillin. The plasmids containing ATG9A were isolated with the Qiagen Miniprep kit (Qiagen, 27,106 × 4) and verified by DNA sequencing. For transfection, 3 × 10⁶ Jurkat cells were transfected with 10 μg of plasmid using Cell Line Nucleofector Kit V (Lonza Bioscience, VVCA-1003); EBV-infected patient B cells were transfected using Human Monocyte Nucleofector Kit (Lonza Bioscience, VPA-1007) following the manufacture’s protocol. The cells were cultured for 48 h before further analyses. Following sample collection, transfection efficiency and protein levels were determined using RT-qPCR and western blotting.

Transient plasmid transfection of HeLa cells was performed using Lipofectamine 2000 (ThermoFisher Scientific, 11,668,019) according to the manufacturer’s instructions. To generate the stable HeLa cell lines expressing the different ATG9A-pCMV GFP plasmids, transiently transfected cells were incubated with 500 μg/mL Geneticin (ThermoFisher Scientific, 10,131,035) for 1 week until cell death was no longer evident. GFP-positive cells were selected on a FACS Aria II Flow Cytometer, collected in bulk, and analyzed by immunofluorescence and immunoblotting to confirm ATG9A-GFP expression and further characterize the stably transfected cells phenotypes. ATG9A-KO HeLa cells were generated by CRISPR-Cas9 system and were previously described [70].

RNA Interference

RNA interference was conducted in Jurkat cells as previously described [71]. Jurkat cells (3 x 10⁶) were transfected with 1 μl of 30 nM Accell Human ATG9A (Horizon, 9065) siRNA – SMARTpool, 5 nmol, (Horizon, E-014294-00-0005) specifically targeting ATG9A or non-target control siRNA (Accell non-targeting pool, Horizon, D-001910-10-05) using the Cell Line Nucleofector^TM Kit V (Lonza Bioscience, VVCA-1003). Cells were cultured for 48 hours before all experiments to achieve optimal downregulation of ATG9A. Following sample collection, knock-down efficiency was determined using RT-qPCR to quantify ATG9A transcript levels and to perform all other downstream applications.

Western blotting and immunoprecipitation

Western blotting and immunoprecipitation using indicated anti-human target antibodies was performed. For western blotting, cells as indicated were lysed using cOmplete Lysis-M (Roche, 04719956001) with the addition of PhosSTOP Phosphate Inhibitor Cocktail Tablets (Roche, 04906837001), and protein concentrations were determined using a colorimetric assay (Bio-Rad, 5,000,006). Fifty micrograms of each sample were subjected to a 4–20% or NativePAGE™ 3–12% Bis-Tris Protein Gel (ThermoFisher, BN1001BOX), 1.0 mm, 10-well, and proteins were transferred onto 0.45-μm PVDF membrane (Millipore, IPVH15150). Membranes were immunoblotted with rabbit anti-ATG9A antibody (Abcam, ab108338), rabbit anti-LC3B antibody (Abcam, ab221794), mouse anti-BCL10 antibody (Santa Cruz Biotechology, SC-5273), rabbit anti-SQSTM1/p62 antibody (Abcam, A302-857A), mouse anti-LC3 antibody (Novus Biologicals, NB600-1384), HRP-conjugated anti-GFP (Miltenyi Biotec, 130,091,833), HRP-conjugated anti-ACTB/actin (Sigma, A3854), or mouse anti-actin (Millipore, MAB1501). Membranes were then probed with rabbit anti-horseradish peroxidase-conjugated secondary antibodies (Abcam, ab205718) or anti-mouse secondary antibodies (Abcam, ab97040) and proteins were visualized using Clarity Western ECL Substrate (Bio-Rad, 170–5061).

For immunoblotting, 1 mg of total protein from 5 × 10⁶ Jurkat cells was precipitated using 5 μg of mouse mAb antiSQSTM1/p62 antibody (Abcam, ab56416) attached to 50 μl protein G magnetic beads (Cell Signaling Technology, 0024S) for 2 h at room temperature. Samples were then washed 6 times with phosphate-buffered saline, pH 7.0 and prepared for gel electrophoresis. Lysate input was used as a control. Fifty μg of each sample was subjected to electrophoresis using 4–20% Mini-Protean® TGX™precast gels (Bio-Rad, 4,561,094) and proteins were transferred onto 0.45 μm PVDF membranes (Millipore, IPVH15150). Membranes were immunoblotted with Rabbit mAb anti-BCL10 antibody 1:1000 (Abcam, ab33905). Membranes were then probed with horseradish peroxidase-conjugated goat pAb anti-rabbit IgG secondary antibodies 1:5000 (Abcam, ab205718) and proteins were visualized using Clarity™ Western ECL Substrate (Bio-Rad, 170–5061).

RNA and RT-qPCR

RNA was isolated, and expression of indicated genes was measured by RT-qPCR as previously reported [71]. Primer sequences (Table S3) are provided in Supplemental Materials and Methods. Total RNA was extracted using the RNeasy mini kit (Qiagen Sciences, 74,104). To eliminate genomic DNA contamination, an additional DNase treatment was performed according to the RNeasy kit instructions with the RNase-free DNase set (Qiagen Sciences, 79,254).

RT-qPCR was employed to demonstrate gene expressions using SensiFAST SYBR No-ROX One-Step Kit (Bioline, BIO-72005). The volumes of reaction mix composition given below are based on a standard 20 µl final reaction mix. 2x SensiFAST™ SYBR No-ROX One-Step Mix 10 µl, 10 µM Forward Primer 0.8 µl, 10 µM Reverse Primer 0.8 µl, Reverse transcriptase 0.2 µl, RiboSafe RNase Inhibitor 0.4 µl, H2O up to 16 µl and template 50 ng in 4 µl. The 2-step cycling RT-qPCR was utilized as descripted below. Reverse transcription: 45°C 10 min, polymerase activation: 95°C 2 min, denaturation: 95°C 5 s. Annealing/extension: 60°C 20s with a total of 40 cycles. After completion of these cycles, melting-curve data were collected to verify PCR specificity.

TRG (T cell receptor gamma) gene rearrangement studies

DNA was extracted from FFPE tissue and polymerase chain reaction (PCR) amplified for detection of TRG locus gene rearrangements. A multiplexed PCR was performed following a published protocol [72], using primers that interrogate TRG rearrangements involving all of the known Vγ family members and the Jγ1/2, JP1/2, and JP joining segments. To allow for fluorescence detection, each joining region primer was covalently linked to a unique fluorescent dye. The products were analyzed by capillary electrophoresis on an ABI 3130xl Genetic Analyzer, and electropherograms were analyzed using GeneMapper^TM software version 4.0 (ThermoFisher Scientific, 4,440,915). Duplicate reactions were run for each sample, and the reaction was repeated on independent sections from the lesion, producing 4 separate assay plots.

Confocal Fluorescence Microscopy

HeLa cells (5 x 10⁴) were cultured on coverslips in a 24-well plate, transfected with the indicated ATG9A mammalian plasmid as in the Section, “Generation of ATG9A mammalian plasmids and transfection”, cultured for an additional 24 h, washed with PBS, and fixed for 12 min with 4% paraformaldehyde in PBS containing 0.1 mM CaCl₂, 1.0 mM MgCl₂ (PBSCM buffer) and 4% sucrose (Sigma-Aldrich, 57,903–1 KG) at room temperature. Cells were then washed twice in PBS and permeabilized in PBSCM containing 0.2% Triton X-100 (Sigma, T8787). Cells were blocked for 30 min in 0.2% BSA (GoldBio, A-412) in PBS (blocking solution) and sequentially stained for 30 min with primary and Alexa-conjugated secondary antibodies in blocking solution at 37°C. Coverslips were then washed with PBS and mounted on glass slides with DAPI Fluoromount-G (Electron Microscopy Sciences, 17,984–24). Confocal images were obtained using a Zeiss LSM 780 microscope with a Plan Apochromat 63x objective. Images were processed in ImageJ as described [73].

Image Analysis

LC3 puncta size analysis was carried out using the “Analyze Particle” plugin for ImageJ. Channels of confocal images were split and only untransfected or transfected cells were analyzed, depending on the experiment. The channel for LC3 staining (red) was converted to grayscale, manually thresholded to highlight the puncta, and processed with the “watershed” plugin to automatically cut apart big clumps of large LC3 aggregates. Then, the “Analyze Particle” plugin was used with default parameters except for “Size” that was set at a range “0.02-Infinity” to avoid counting very small dots from the background (noise puncta). Average size (μm²) from 4 random areas per coverslip where consider in each replicate which includes more than 15 cells per condition. Statistical significance was analyzed by one-way ANOVA followed by Dunnett’s test using GraphPad Prism (GraphPad Software).

Statistics

Indicated statistical tests are shown within each Figure legend; Student t-tests were conducted using an unpaired design, using GraphPad Prism Software version 7 (GraphPad Software).

Funding Statement

This work was supported by NIH grant AI-77609 (to EWG), and the Eugene F. and Easton M. Crawford Charitable Lead Unitrust, Pohs, Siegel, and Dreizzesun Funds (to EWG), and by the Intramural Research Programs of the National Institute of Allergy and Infectious Diseases, (NIAID) (AI001123 and AO001124), and the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) (ZIA HD001607), National Institutes of Health, Bethesda MD.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2022.2093028