RLIM Is a Candidate Dosage-Sensitive Gene for Individuals with Varying Duplications of Xq13, Intellectual Disability, and Distinct Facial Features

Abstract

Interpretation of the significance of maternally inherited X chromosome variants in males with neurocognitive phenotypes continues to present a challenge to clinical geneticists and diagnostic laboratories. Here we report 14 males from 9 families with duplications at the Xq13.2-q13.3 locus with a common facial phenotype, intellectual disability (ID), distinctive behavioral features, and a seizure disorder in two cases. All tested carrier mothers had normal intelligence. The duplication arose de novo in three mothers where grandparental testing was possible. In one family the duplication segregated with ID across three generations. RLIM is the only gene common to our duplications. However, flanking genes duplicated in some but not all the affected individuals included the brain-expressed genes NEXMIF, SLC16A2, and the long non-coding RNA gene FTX. The contribution of the RLIM-flanking genes to the phenotypes of individuals with different size duplications has not been fully resolved. Missense variants in RLIM have recently been identified to cause X-linked ID in males, with heterozygous females typically having normal intelligence and highly skewed X chromosome inactivation. We detected consistent and significant increase of RLIM mRNA and protein levels in cells derived from seven affected males from five families with the duplication. Subsequent analysis of MDM2, one of the targets of the RLIM E3 ligase activity, showed consistent downregulation in cells from the affected males. All the carrier mothers displayed normal RLIM mRNA levels and had highly skewed X chromosome inactivation. We propose that duplications at Xq13.2-13.3 including RLIM cause a recognizable but mild neurocognitive phenotype in hemizygous males.

Keywords: NEXMIF; RLIM; Tonne-Kalscheuer syndrome; Xq13; autism; chromosomal duplication; chromosomal microarray; dosage sensitive gene; intellectual disability; whole genome sequencing.

Figure 1

Pedigrees Pedigrees of families with Xq13.2–13.3 duplication: affected males with neurodevelopmental disorder, filled boxes; unaffected carrier females, half-filled circles; tested individuals with Xq duplication are marked with a plus sign; tested individuals without Xq duplication are marked with a minus sign; individuals who have not been tested are denoted “NT;”

Figure 2

Photographs Facial features of males with Xq13.2–13.3 duplications where families consented to sharing photographs. Common distinctive facial features include straight, medially flared eyebrows, short palpebral fissures, and flat midface.

Figure 3

Duplication of RLIM Caused by Structural Variants, Revealed by WGS in the Male Probands from Family 1 This panel demonstrates the location of the two duplications relative to the X chromosome and RLIM, i.e., Xq13.1 (red triangle) and Xq13.2–13.3 (blue triangle) in the proband from family 1. The depth of coverage is shown in blue and the phred scaled read mapping quality in gray (Q = −10 log₁₀^∗ P, Q of 30 corresponds to 1 misaligned reads in 1,000). The structural variants (SVs) and their supporting discordant pairs (DP) and split reads (SR) are shown. The WGS data indicate the existence of four breakpoints (labeled Bp1–4) and 3 structural variants (SV1–3). Also shown is a dot matrix view of genome sequence surrounding breakpoint 2 (Bp2) aligned to itself, showing a palindromic sequence that could cause DP and SR reads to misalign. As a result, SV1 and SV3 could represent a single SV. Alongside is shown models of the rearrangement considering depth of coverage change, discordant pairs, and split reads. Model I and II are equally plausible and only involve SV1 and SV2. A model involving SV3 is not plausible as it would require the middle section between both duplications to also be duplicated (model III), which contradicts the observed depth of coverage. The sequencing reads of SV3 at Bp2 are most likely wrongly mapped due to the palindrome and should instead map upstream of Bp2, which then provides further supporting evidence for SV1, thus SV1 and SV2 being the only real SV in this region. Breakpoints are provided in hg38.

Figure 4

Duplication of RLIM Caused by Structural Variants, Revealed by WGS in the Male Probands from Family 5 This panel demonstrates duplication relative to the X chromosome and RLIM in the proband from family 5. Evidence from split reads is shown, consistent with the duplication being a simple tandem duplication, which contains a small deletion within ABCB7. Breakpoints are provided in hg38.

Figure 5

Gene Expression Studies RT-qPCR analysis of FTX, RLIM, and NEXMIF expression in LCLs from the affected males from families 1–5 (upper graphs) and RT-qPCR analysis of FTX, SLC16A2, RLIM, and NEXMIF expression in fibroblasts (FIB) from individual 2 (III:7) from family 2 (lower graphs), with a graphic depiction of the duplicated region in each family tested. Samples were run in triplicate and mean gene expression was normalized to mean expression of either HPRT1 or GAPDH. WT male data is presented as mean ± SD (error bars).

Figure 6

Protein Expression Studies Western blots of LCL lysates probed for RLIM and its downstream targets MDM2 and TP53. ACTB was used as a loading control. RLIM, MDM2, and TP53 western blot signals were quantified using Image Lab software (Bio-Rad) and their normalized values relative to ACTB signal intensity were plotted. ^∗p = 0.029 and ^∗∗p = 0.006, using a two-tailed unpaired t test. Control subjects were four individuals with wild-type RLIM, “RLIM dup” were 7 individuals with chromosomal duplications from the current cohort with all of RLIM duplicated, and “R365C” were two brothers with a previously reported missense variant in RLIM p.Arg365Cys. The data are presented as mean ± SD (error bars).