Novel Variants of PPP2R1A in Catalytic Subunit Binding Domain and Genotype–Phenotype Analysis in Neurodevelopmentally Delayed Patients

Round 1

Reviewer 1 Report

Through whole exome sequencing studies, Qian et al. find new variants in PPP2R1A, a regulatory subunit of protein phophatase 2A. Variants in PPP2R1A have been observed previously in individuals with neurodevelopment disorders, and the variants described in this paper expand this list. Four of the variants identified were used to carry out functional studies using FLAG-tagged, over expressed, protein in HEK293T cells. Specifically, the authors examined total protein levels and the ability of mutant proteins to interact with PPP2R5D and PPP2CA by co-IP. The data presented is generally clear, although I have several concerns.

1. M180T, R498L and D282fs clearly show deficits in binding to PPP2R5D and/or PPP2CA. This is consistent with these variants being pathogenic.  This, however, is not clear for V470A, which seems to show normal interactions. It was not clear why this variant was classified as pathogenic? 

2. It was not clear how many variants (total) were detected in these patients. More transparency on the genotypic information would be nice. 

Editing for language needs to done. 

Author Response

Through whole exome sequencing studies, Qian et al. find new variants in PPP2R1A, a regulatory subunit of protein phosphatase 2A. Variants in PPP2R1A have been observed previously in individuals with neurodevelopment disorders, and the variants described in this paper expand this list. Four of the variants identified were used to carry out functional studies using FLAG-tagged, over expressed, protein in HEK293T cells. Specifically, the authors examined total protein levels and the ability of mutant proteins to interact with PPP2R5D and PPP2CA by co-IP. The data presented is generally clear, although I have several concerns.

1. M180T, R498L and D282fs clearly show deficits in binding to PPP2R5D and/or PPP2CA. This is consistent with these variants being pathogenic. This, however, is not clear for V470A, which seems to show normal interactions. It was not clear why this variant was classified as pathogenic? 

Response: Thanks for the reviewer’s comment. We agree that the V470A showed normal interactions, and we also classified the V470A as an uncertain significance variant as in Table 1.

2. It was not clear how many variants (total) were detected in these patients. More transparency on the genotypic information would be nice. 

Response: Patients 2 and 3 were tested trio whole exome sequencing (Trio-WES), Patients 1, 4 and 5 were tested proband-only clinical exome sequencing. Except for the variants of PPP2R1A, other genetic-causing variants were excluded in these five cases.

The de novo and compound heterozygous variants detected from Trio-WES, and the candidate variants from proband clinical exome sequencing are listed in the following table.

The variants were annotated using Variant Effect Predictor (VEP-Ensembl 73). Functional exonic and splicing variants were considered in the analysis. Allelic frequencies were annotated from public datasets, including 1000 Genomes (1KG), Exome aggregation Consortium (ExAC), GNOMAD, and our internal databases, which include over 100k samples. Further analysis of single nucleotide variants (SNV) and small indels follows an in-house pipeline, including off-target filtering, public population frequency filtering, internal population frequency filtering, variant damage prediction and inheritance model scoring. The analysis pipelines filtered out an average of ~200 variants in each sample for manual analysis. The total variants detected in these five patients were 237, 233, 164, 115 and 188, respectively. 

The de novo and compound heterozygous variants detected from Trio-WES, and the candidate variants from proband clinical exome sequencing are listed in the following table.

ID	Gene	location	Variant	Zygo	OMIM	Inherit	Source	Comments
Patient1	ANKRD11	chr16: 89357415	NM_013275:exon5: c.397+6C>T	Het	KBG syndrome, [MIM:148050]	AD	NA	1 normal individual in in-house dataset
	PPP2R1A	chr19: 52715973	NM_014225:exon5: c.538A>G(p.M180V)	Het	Houge-Janssen syndrome 2 [MIM:616362]	AD	De novo	Pathogenic variant matches the patient's phenotype
	PRSS12	chr4: 119234489	NM_003619:exon7:c.1356C>T(p.D452D)	Het	Intellectual developmental disorder, autosomal recessive 1, [MIM:249500]	AR	Maternal	Synonymous variant, homozygous in-house dataset without intellectual disability feature
	PRSS12	chr4: 119273715	NM_003619:exon1:c.161C>G(p.P54R)	Het	Intellectual developmental disorder, autosomal recessive 1, [MIM:249500]	AR	Paternal	homozygous carrier in in-house dataset without intellectual disability feature
Patient 2	DHX58	chr17: 40259777	NM_024119:exon8: c.841_842delinsAA(p.A281K)	Het	.		De novo	Non OMIM disease gene
	NUP210	chr3: 13383363	NM_024923:exon23: c.3113A>G(p.D1038G)	Het	.		Paternal	Non OMIM disease gene
	NUP210	chr3:13420488	NM_024923:exon8: c.977-8A>G	Het	.		Maternal	Non OMIM disease gene
	PPP2R1A	chr19:52715974	NM_014225:exon5: c.539T>C(p.M180T)	Het	Houge-Janssen syndrome 2 [MIM:616362]	AD	De novo	Pathogenic variant matches the patient's phenotype
	SLC27A3	chr1:153749134	NM_001317929:exon2: c.808C>T(p.P270S)	Het	.		Maternal	Non OMIM disease gene
	SLC27A3	chr1:153747984	NM_001317929:exon1: c.11_23delinsT (p.P8_L11del)	Het	.		Paternal	Non OMIM disease gene
Patient 3	AMER1	chrX:63412601	NM_152424:exon2: c.566A>G(p.Q189R)	Hemi	Osteopathia striata with cranial sclerosis, [MIM:300373]	XLD	Maternal	From normal mother
	MYO5B	chr18:47431204	NM_001080467:exon20:c.2415-6C>G	Het	Microvillus inclusion disease, [MIM:251850]	AR	Maternal	Phenotype not match
	MYO5B	chr18:47429121	NM_001080467:exon21:c.2654A>G(p.D885G)	Het	Microvillus inclusion disease, [MIM:251850]	AR	Paternal	Phenotype not match
	PPP2R1A	chr19:52719066	NM_014225:exon7: c.842delinsCA (p.D282Rfs*14)	Het	Houge-Janssen syndrome 2 [MIM:616362]	AD	De novo	Likely pathogenic variant matches the patient's phenotype
Patient 4	ARHGAP31	chr3:119134635	NM_020754:exon12: c.3859A>G(p.T1287A)	Het	Adams-Oliver syndrome 1, [MIM:100300]	AD	NA	6 individuals without intellectual disability were in the in-house dataset
	CHRNA4	chr20:61981756	NM_000744:exon5: c.1007G>A(p.R336H)	Het	Epilepsy, nocturnal frontal lobe, 1, [MIM:600513]	AD	NA	Phenotype not match
	COL2A1	chr12:48369765	NM_001844:exon50: c.3578C>T(p.S1193L)	Het	Achondrogenesis, type II or hypochondrogenesis, [MIM:200610]; [MIM:156550]; Stickler syndrome, type I, [MIM:108300]; et al	AD	NA	Phenotype not match
	GJB2	chr13:20763612	NM_004004:exon2: c.109G>A(p.V37I)	Het	Deafness, autosomal dominant 3A, [MIM:601544]; Deafness, autosomal recessive 1A, [MIM:220290]; et al	AR	NA	Likely pathogenic variant, carrier
	HARS1	chr5:140056390	NM_002109:exon10: c.1043G>A	Het	Charcot-Marie-Tooth disease, axonal, type 2W, [MIM:616625]; Usher syndrome type 3B, [MIM:614504]	AD/AR	NA	Phenotype not match
	HTT	chr4:3208331	NM_002111:exon43: c.5827G>A	Het	Huntington disease, [MIM:143100]; Lopes-Maciel-Rodan syndrome, [MIM:617435]	AD/AR	NA	Phenotype not match
	PPP2R1A	chr19:52724277	NM_014225:exon12: c.1409T>C(p.V470A)	Het	Houge-Janssen syndrome 2  [MIM:616362]	AD	NA	Uncertain significance variant matches the patient's phenotype
	VCAN	chr5:82837289	NM_004385:exon8: c.8467C>G(p.Q2823E)	Het	Wagner syndrome 1, [MIM:143200]	AD	NA	Phenotype not match
Patient 5	DDX41	chr5:176938856	NM_016222:exon17: c.1786_1804del (p.P596Rfs*)	Het	{Myeloproliferative/lymphoproliferative neoplasms, familial (multiple types), susceptibility to}, [MIM:616871]	AD	NA	Phenotype not match
	GJB4	chr1:35227495	NM_153212:exon2: c.640A>G(p.M214V)	Het	Erythrokeratodermia variabilis et progressiva 2, [MIM:617524]	AD	NA	Phenotype not match
	PPP2R1A	chr19:52724361	NM_014225:exon12: c.1493G>T(p.R498L)	Het	Houge-Janssen syndrome 2 [MIM:616362]	AD	De novo	Likely pathogenic variant matches the patient's phenotype
	RHOBTB2	chr8:22864996	NM_001160036:exon7: c.1304T>C(p.V435A)	Het	Developmental and epileptic encephalopathy 64, [MIM:618004]	AD	NA	9 individuals without intellectual disability in the in-house dataset
	TRIO	chr5:14498209	NM_007118:exon52: c.8059T>A(p.F2687I)	Het	Intellectual developmental disorder, autosomal dominant 44, with microcephaly, [MIM:617061]; Intellectual developmental disorder, autosomal dominant 63, with macrocephaly, [MIM:618825]	AD	Paternal	From no feature father

3. Comments on the Quality of English Language

Editing for language needs to done. 

Response: Thanks for the reviewer’s comment. We had sent our manuscript for English languange editing.

Author Response File: Author Response.docx

Reviewer 2 Report

Qian et al. report yet unreported three PPP2R1A variants with functional tests and make a molecular and clinical review of the literature. However, I do not think that this work adds significant material to what is already known. Indeed, of the three novel variants, to my mind only the c.1493G>T, p.Arg498Leu seems relevant. It may be classified ACMG class 4 as the authors propose but I would use PM2 moderate, PP2 supporting, and PS2 strong arguments. The issue is that there are three pathogenic missense variants clusters and that this variant is not localized in any of them, in the 3’ side of the protein where several missense variants are reported in gnomAD. However, missense 3D predictions are in favor of a deleterious impact (Structural damage detected, Buried charge replaced Buried salt bridge breakage) but it may act as a loss of function mechanism whereas the known mechanism is dominant-negative type of gain-of-function effect (see GeneReviews).

The c.1409T>C, p.Val470Ala is clearly a class 3 variant (PM2 moderate, and PP2 supporting arguments ; authors use PP3 argument but in silico predictions are clearly not in favor of a deleterious impact). Missense 3D predictions are: No structural damage detected.

Last, the c.843dupA, p.Asp282Argfs*14 is a truncating variant. ACMG classification could be class 4 (PM2 moderate, and PS2 strong). PVS1 can not be used here because loss of function is not a known mechanism, see GeneReviews ("It should be noted that PPP2R1A loss-of-function changes (e.g., nonsense variants or deletions) are not known to cause PPP2R1A-NDD, even though the gene is quite intolerant to loss-of-function changes (gnomAD pLI is 0.98)."). To my mind, this variant is rather a class 3 variant. Besides, authors do not explain why functional tests are not in favor of NMD which should occur because of the localization of the variant.

Functional tests are not in accordance of what is known. In Figure 2B, it seems that well known pathogenic M180T variant impacts PPP2R5D binding which should not be the case, see GeneReviews (“The most important B-subunit for brain function appears to be the PP2A B56-delta subunit PPP2R5D, and pathogenic missense variants in this gene are the cause of a phenotypically overlapping condition, PPP2R5D-related neurodevelopmental disorder. Remarkably, all pathogenic PPP2R1A variants retain binding to PPP2R5D [Lenaerts et al 2021]”).

I also think that the clinical review does not add significant information compared to GeneReviews.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Qian et al. identified the Novel variants in PPP2R1A and also compared the genotype-phenotype relations in NDD patients having PPP2R1A variants. Overall, the study is well-written and concise.

1. Did the author check novel variants in "Bravo" and "all of US" genome browsers?

2. What was the cut-off value for allelic frequency during exome analysis?

3. Did the author check compound het variants in exome data?

4. Are all the variants mentioned in Supplementary Table 1, de novo? if yes please mention it in the table or legend.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The authors have responded to my comments and questionings.