Gray Matter Variation in the Posterior Superior Temporal Gyrus Is Associated with Polymorphisms in the KIAA0319 Gene in Chimpanzees (Pan troglodytes)

Abstract Determining the impact that the KIAA0319 gene has on primate brain morphology can provide insight into the evolution of human cognition and language systems. Here, we tested whether polymorphisms in KIAA0319 in chimpanzees account for gray matter volumetric variation in brain regions implicated in language and communication (particularly within the posterior superior temporal gyrus and inferior frontal gyrus). First, we identified the nature and frequencies of single nucleotide variants (SNVs) in KIAA0319 in a sample of unrelated chimpanzees (Pan troglodytes spp.). Next, we genotyped a subset of SNVs (those important for gene regulation or likely to alter protein structure/function) in a sample of chimpanzees for which in vivo T1-structural magnetic resonance imaging scans had been obtained. We then used source-based morphometry (SBM) to test for whole-brain gray matter covariation differences between chimpanzees with different KIAA0319 alleles. Finally, using histologic sections of 15 postmortem chimpanzee brains, we analyzed microstructural variation related to KIAA0319 polymorphisms in the posterior superior temporal cortex. We found that the SNVs were associated with variation in gray matter within several brain regions, including the posterior superior temporal gyrus (a region associated with language comprehension and production in humans). The microstructure analysis further revealed hemispheric differences in neuropil fraction, indicating that KIAA0319 expression may be involved in regulation of processes related to the formation and maintenance of synapses, dendrites, or axons within regions associated with communication.


Introduction
Characterizing the behavioral, neural and genomic systems that distinguish humans from other primates has been a central focus of comparative research. In humans, there are several key connected brain regions implicated in the production and comprehension of language, particularly within the left hemisphere, including the basal ganglia, supplemental motor, inferior frontal, and posterior superior temporal gyri (Goulven and Tzourio-mazoyer, 2004;Friederici, 2015;Skeide et al., 2016). Further, studies have discovered several genes that account for individual and phylogenetic variation in linguistic functions and their underlying neural systems (Fisher and Marcus, 2006;Fisher and Scharff, 2009;Preuss, 2012;Pfenning et al., 2014). Arguably, the most well-known so-called language gene is FOXP2 (Lai et al., 2001;Enard et al., 2002). Individuals with a FOXP2 R553H missense mutation exhibit oro-facial dyspraxia and concomitant decrements in gray matter volume in the basal ganglia, supplemental motor area, anterior cingulate, and inferior frontal gyrus (Vargha-Khadem et al., 2005).
The gene KIAA0319 has similarly been implicated in language functions of humans. KIAA0319 polymorphisms have been repeatedly associated with developmental dyslexia, a reading disorder marked by difficulties learning associations between sounds with letters and/or written words (Cope et al., 2005;Dennis et al., 2009;Scerri et al., 2011;Luciano et al., 2013;Sun et al., 2014). Meta-analyses have revealed that individuals with dyslexia show atypical patterns of gray matter volume, cortical thickness, and white matter connectivity in brain regions implicated in language (Eckert et al., 2016;Mascheretti et al., 2017). Furthermore, KIAA0319 single nucleotide variants (SNVs) are associated with the prevalence of dyslexia and linked to differences in gray and white matter cortical organization, particularly within the posterior superior temporal lobe and parietal junction within the left hemisphere (Darki et al., 2012;Pinel et al., 2012;Jamadar et al., 2013;Eicher et al., 2016;Ozernov-Palchik and Gaab, 2016).
Studying the impact of variation in genes like KIAA0319 on primate brain morphology can provide insight into the evolution of human cognition and language systems by clarifying the functional role of genetic variation in shaping neural circuits. To this end, we tested whether chimpanzee KIAA0319 polymorphisms account for gray matter variation in brain regions implicated in language, particularly the posterior superior temporal and inferior frontal gyri. Comparative studies in nonhuman primates have shown that these regions play a role in the production and processing of species-specific communicative signals and vocalizations (Heffner and Heffner, 1984;Poremba et al., 2003Poremba et al., , 2004Gil-da-Costa et al., 2006;Petkov et al., 2008;Taglialatela et al., 2009). There is also evidence that chimpanzees, and to a lesser extent some monkeys, exhibit population-level morphologic and cellular asymmetries in the inferior frontal gyrus and superior temporal cortex, which overlap with Wernicke's area in the human brain (Hopkins and Nir, 2010;Spocter et al., 2010;Lyn et al., 2011;Marie et al., 2018), one of the main regions influenced by KIAA0319 polymorphisms. What remains unknown is whether individual variation in morphology or function in these brain regions is associated with KIAA0319 polymorphisms in chimpanzees.
To test this hypothesis, we initially identified the frequencies of KIAA0319 SNVs in a sample of unrelated chimpanzees. Next, we selected SNVs of interest for further genotype-phenotype study. We focused on SNVs present in the 59 upstream region of the gene (likely important for gene regulation as they are in or near the promoter region and transcription start site of the gene), as well as nonsynonymous SNVs in the coding region of the gene (these cause amino acid changes and therefore are more likely to alter structure and/or function of the protein). Next, we genotyped these SNVs in a sample of chimpanzees for which in vivo T1-structural magnetic resonance imaging scans had previously been obtained. We used source-based morphometry (SBM) to examine whole-brain gray matter covariation differences across KIAA0319 SNVs. Additionally, we analyzed microstructural variation related to KIAA0319 polymorphisms in the posterior superior temporal cortex from histologic sections of postmortem chimpanzee brains.

Subjects
In the current study, we used archival structural T1weighted magnetic resonance imaging scans obtained from 136 adult captive chimpanzees (79 females and 57 males). Captive chimpanzees in the United States descended from a founder population that was .95% from the western subspecies (Pan troglodytes verus; Ely et al., 2005). The subjects were socially housed at the National Center for Chimpanzee Care (NCCC), which is part of The University of Texas MD Anderson Cancer Center. The NCCC facilities have indoor/outdoor enclosures with climbing structures, bedding, and daily environmental enrichment (Neal Webb et al., 2018). All chimpanzees had 24-h access to both indoor and outdoor areas except during cleaning. Care staff fed the chimpanzees a diet of commercially available primate chow and fresh produce multiple times per day, as well as several daily foraging opportunities and ad libitum access to water. All work was approved by the Institutional Animal Care and Use Committee at NCCC, followed the guidelines of the Institute of Medicine on the use of chimpanzees in research, and complied with the Society for Neuroscience Policy on Ethics.
Next, to confirm the presence and allele frequencies of the genetic variants in the population of chimpanzees for which we have matching neuroanatomical data, we genotyped the SNVs that showed variation in genome data for P. troglodytes verus in an additional set of 25 presumed unrelated chimpanzees. Only variants that showed minor allele frequencies .5% were then genotyped for the complete population. To infer the putative functional consequences of the resulting coding variants, we used SNAP2 (Hecht et al., 2015), a trained classifier based on a machine-learning device called a "neural network." It distinguishes between effect and neutral variants/nonsynonymous SNVs by taking a variety of sequence and variant features into account. The effect of a variant is believed to be of importance to the native protein function if the SNAP2 score exceeds 50, is neutral if the score is below À50 and is unreliable when between 50 and À50.

DNA extraction and genotyping
For the chimpanzees with matching neuroanatomical data, genomic DNA was extracted from 200-ml blood samples using the QIAampDNA Mini kit automated on a QiaCube (QIAGEN), following the manufacturer's instructions. The DNA extract was recovered in 200 ml of elution buffer and kept frozen at -20°C. Negative extraction controls showed no evidence of DNA contamination. DNA concentrations were quantified using a Nanodrop 2000 (Thermo-Fisher Scientific) spectrophotometer. Subsequent genotyping of the SNVs of interest was done using either Sanger sequencing or High-Resolution Melt Analysis (Smith et al., 2010). For exons where multiple SNVs were present in relatively short fragments (,500 bp), exons were amplified using PCR. Primers were developed flanking the region with the SNVs (exon 4, forward CTT GGA AAC CAC CAG AAT CAT GCG G and reverse CTT TCT CCA ACA CCT CTC CTG AAG; exon 18, forward CTC AGC CCT ACA CAC CTC TTT G and reverse AAG CCT CCT CTA TAC CAG AAC TC). The 25-ml PCR mix contained 10.5 ml MasterMix (QIAGEN; HotStarTaq plus DNA polymerase, deoxynucleotides, and MgCl 2 ), 2.0 ml primer mix (10 mM concentration of combined forward and reverse primers) and ;5-20 ng of genomic DNA (quantified via a NanoDrop microvolume spectrophotometer). PCR started with an initial activation at 95°C (5 min), followed by 35 cycles of denaturing at 95°C (10 s), annealing at 54°C (30 s) and extension at 72°C (1 min), with a final extension period of 10 min at 72°C. All reactions included positive and no-template controls. The amplicons were then Sanger sequenced on an Applied Biosystems platform following manufacturer's specifications at the Yale DNA Analysis Facility. For ,100-bp exon sequences where only one SNV site was present, High-Resolution Melt Analysis (as in Jacobs et al., 2016) was used to assess genotypes. Primer pairs flanking the SNVs were designed to target a short segment (;100 bp) containing the single polymorphic site (chr6:24899739, forward CTG ACT GAG ACT GGG AAC and reverse CTG TGG TTT AGA TTA TCA CG; chr6:24900865, forward GCA ATT CGT GAG GTT GGG and reverse GCC CTG AGT TCA GAA GG; chr6:24828663, forward GAA TGG AAA CCA GAG CAG TG and reverse ATG GAG AAC TTG CCT GCA TGA). All qPCRs and melting curves were generated on a Rotor-Gene Q (QIAGEN) platform. The 25 mL qPCR mix contained 12.5 mL HRM MasterMix (QIAGEN; HotStarTaq plus DNA polymerase, EvaGreen dye, Q-Solution, deoxynucleotides, and MgCl 2 ), 1.75 ml primer mix (10 mM concentration forward and reverse primers) and ;20 ng of genomic DNA (quantified via a NanoDrop microvolume spectrophotometer). qPCR started with an initial activation at 95°C (5 min), followed by 40 cycles of denaturing at 95°C (10 s), annealing at 54°C (30 s) 72°C (40 s) and a final extension period of 10 min at 72°C. In each reaction, no-template controls were included. Data on melting curves and temperature of qPCR products were generated immediately following amplification by increasing temperatures from 65°C to 95°C, rising by 0.1°C/2 s. Fluorescence data were plotted as a function of temperature during DNA denaturation (melting) and visualized and compared using the Rotor-Gene Q HRM software package (QIAGEN). The resulting melting temperatures and curve shapes were assigned to different genotypes based on previously established reference curves (as in Jacobs et al., 2016) of known variants. Genotypes were confirmed using multiple independent qPCR and high-resolution melt analyses (mean = 2.76 replicates per sample). Additionally, for each melting curve-based genotype, we Sanger sequenced 10% of the samples to validate and confirm the genotype (Applied Biosystems Genetic Analyzer, DNA Analysis Facility at Yale University). Multiple alignments of the resulting DNA sequences were performed using Geneious (version 6.0.6).

Magnetic resonance image acquisition
We previously obtained magnetic resonance images using methods described elsewhere (Bogart et al., 2014;Bard and Hopkins, 2018;Mulholland et al., 2020). No new magnetic resonance images were collected for this study. Briefly, subjects were initially immobilized by injection of ketamine (10 mg/kg) or telazol (2-6 mg/kg) and subsequently anesthetized with propofol (40-60 mg/kg) per standard institutional guidelines. The subjects were then transported to the imaging facility and remained anesthetized for the duration of the scan (40-60 min depending on brain size), as well as during transportation back to a recovery cage (;75-120 min in total). The subjects were placed in the scanner chamber in a supine position with their head inside the human-head coil. The chimpanzees were scanned using a 1.5-Tesla scanner and T1weighted images were collected in the transverse plane using a gradient echo protocol (pulse repetition = 19 ms, echo time = 8.5 ms, number of signals averaged = 8, and a 256 Â 256 matrix). After completing the magnetic resonance imaging scan collection, the subjects were temporarily housed in a single cage for several hours to allow them to recover from the anesthesia, after which they were returned to their social group.

Postimage processing and analyses
The archival magnetic resonance imaging data were processed on a Macintosh computer using previously described methods (Mulholland et al., 2020). First, we imported the raw DICOM files into 3D Slicer 4 (www. slicer.org) and converted each into NIfTI format (Fedorov et al., 2012;Kikinis et al., 2014). Next, we used the BET function in FSL for skull-stripping, using fractional intensity thresholds ranging between 0.35 and 0.80 (Smith, 2002;Jenkinson et al., 2005). The skull-stripped brains were then imported into 3D Slicer for N4ITK bias correction (spline distance = 50, bias field = 0.15, convergent threshold = 0.001; Boyes et al., 2008;Tustison and Gee, 2009;Tustison et al., 2010Tustison et al., , 2014. Using the MRI Denoising Package for MATLAB (R2015b; MathWorks), the scans were then denoised using an optimized nonlocal means denoising filter (Coupe et al., 2008). The scans were then resampled at 0.625 isotropic voxels and aligned in radiologic space in Analyze 11.0 (AnalyzeDirect), using a capsule placed during the imaging process as a left-right orientation marker. Finally, we used the FLIRT function in FSL to perform a 12-parameter affine linear registration (Jenkinson and Smith, 2001;Jenkinson et al., 2002) of the processed scan to a chimpanzee template brain (Hopkins and Avants, 2013).
We used the FSL-VBM pipeline (http://fsl.fmrib.ox.ac. uk/fsl/fslwiki/FSLVBM) to process each preprocessed scan. The FSL-VBM pipeline included (1) segmentation of each scan into gray and white matter, (2) linear registration of each scan to a standard chimpanzee template (Hopkins and Avants, 2013), (3) creation of a study-specific gray matter template (Smith et al., 2004;Andersson et al., 2007;Douaud et al., 2007), (4) nonlinear registration of each subject's gray matter images to the study-specific template, (5) modulation of the gray matter volume by use of a Jacobian warp to correct for local expansion or contraction of gray matter within each voxel, and (6) smoothing with an isotropic Gaussian kernel with a s of 2 mm.

SBM and statistical analyses
Following previously described methods (Hopkins et al., 2019;Mulholland et al., 2020), we subjected the individual modulated gray matter volume produced in the FSL-VBM pipeline to perform SBM using the software program GIFT (https://trendscenter.org/software/gift/ ). SBM is a relatively new method used to characterize gray and white matter structural covariation in samples of magnetic resonance imaging scans (Alexander-Bloch et al., 2013;Bard and Hopkins, 2018). Unlike univariate analytic methods, such as voxel-based morphometry, SBM is a multivariate, data-driven analytic approach that utilizes information about relationships among voxels to group voxels carrying similar information across the brain. Without requiring prior determination of regions of interest, the resulting components or sources are identified based on the spatial information between voxels grouped in a natural manner and represent similar covariation networks between subjects; thus, this approach has been described as a multivariate version of voxel-based morphometry (Xu et al., 2009). At the individual level, GIFT outputs weighted scores for each component that reflect the relative contribution of each subject's gray matter volume to their generation. For this sample of chimpanzees, the SBM analysis yielded eight components. The weighted scores for each of the eight components were the dependent measures in the subsequent multivariate analysis of covariance (MANCOVA). Sex and KIAA0319 SNV alleles were the independent variables, while genetic relatedness was the covariate.

Microstructural analyses of the posterior superior temporal cortex (area Tpt)
In addition to the in vivo SBM analysis, we selected area Tpt for histologic analysis because it overlaps with one of the larger gray matter regions that differed between chimpanzees in the SBM analysis. For histologic sectioning, tissue blocks were frozen on dry ice, and cut at 40 mm on a freezing sliding microtome. Every 10th section (spaced 400 mm apart) from each block was mounted on chromalum-subbed slides, stained using 0.5% cresyl violet to visualize cytoarchitecture, and coverslipped with DPX.
We quantified the neuropil fraction from the posterior superior temporal cortex (area Tpt) in the postmortem brains of 15 chimpanzees (ages 12-48 years, mean 32.6 years; seven female, eight male). The neuropil is defined as the space between neuronal and glial cell bodies, which is comprised of dendrites, axons, synapses, and microvasculature (Spocter et al., 2012). The neuropil fraction data came from previously published data in 12 individuals from Spocter et al. (2012), supplemented by an additional three brains measured for the current study, which had been processed more recently for histology. Neuropil fraction was measured using high-resolution images of these Nissl-stained sections. Imaging of the regions of interest was performed with a Zeiss Axioplan 2 photomicroscope (Zeiss) equipped with a Ludi XY motorized stage (Ludl Electronics), Heidenhain z-axis encoder, and an Optronics MicroFire color video camera coupled to a Dell PC using Stereoinvestigator software (MBF Bioscience). For each chimpanzee brain, three evenly spaced coronal sections were sampled throughout the region of interest. For each section, regions of interest were contoured under low magnification (2.5Â objective lens) underneath a representative portion of the area (for cortical regions this was ;3 mm in length along the cortical surface). At least 30 systematically random sampled images were taken within the contours of each section using a 20Â objective lens, resulting in images at 0.53 pixels/mm resolution. Each image was imported into ImageJ (v.1.32j) and subjected to background subtraction with a rolling ball radius of 50 pixels and then converted to binary by an automated threshold routine (Spocter et al., 2012). Before calculation of the neuropil fraction, images that contained artifacts were removed from the batch and the remaining images were blind-coded to avoid observer bias.
We also analyzed previously published data on regional volume, total neuron number, and neuron density from area Tpt in 12 chimpanzees (Spocter et al., 2010). In brief, stereologic methods were used, including Cavalieri point counting to calculate the regional volumes of area Tpt, and optical fractionator sampling to estimate neuron numbers. In the cresyl violet Nissl-stained sections, neurons were distinguished from glial cells by the presence of a large, lightly stained nucleus and a distinct nucleolus, accompanied by lightly stained dendritic processes. Stereologic counts of neuron numbers based on cresyl violet Nissl-staining have been shown to produce results that are highly correlated with other methods of labeling neurons, such as NeuN immunostaining (Zhu et al., 2015).
We ran a series of multivariate general linear models using Type III sum of squares with each dependent measure of cortical area Tpt neuropil fraction, regional volume, neuron number, and neuron density in the left and right hemisphere, with sex, age, and brain mass as covariates, and all three KIAA0319 genotypes as fixed factors. We also analyzed the asymmetry quotient (AQ; rightleft/average of hemispheres) of neuropil fraction using general linear models with sex, age, and brain mass as covariates.

KIAA0319 genetic variation
We identified 33 SNVs in chimpanzee KIAA0319 59 flanking region and coding regions combined (see Table 1). Out of 33 SNVs, four were in the upstream flanking region and 29 were in the amino acid coding region. A total of 18 out of 29 SNVs were nonsynonymous variants, and only two of these were present in P. troglodytes verus: chr6:24841281: C/A in exon 4 resulting in a Val295Phe substitution, and chr6:24828663: A/C in exon 10 resulting in a Glu563Asp substitution. Next, we genotyped these two SNVs in our population of chimpanzees with matching neuroanatomical data and found that only chr6:24828663: A/C showed minor allele frequencies .5% and was thus suitable for further genotype-phenotype analysis. Genotype frequencies for this SNV (hereafter referred to as KIAA0391_Glu563Asp) were in Hardy Weinberg equilibrium (x 2 = 0.52, df = 1, p = 0.47) and SNAP2 prediction of its functional consequences resulted in a modest SNAP2 effect score of 35 (expected accuracy 66%).
Out of the four SNVs present in the promoter flanking region, two were in complete linkage disequilibrium in our chimpanzee population (chr6:24899739:T/A and chr6:24899753:C/A), therefore only one was included in further genotype-phenotype analysis. A third SNV (chr6:24899741:C/G) did not exceed the .5% minor allele threshold and was also excluded from further analysis. For the two promoter region flanking region SNVs that were included in the analysis, genotype frequencies were in Hardy Weinberg equilibrium (chr6:24899739:T/ A x 2 = 3.79, df = 1, p = 0.05; chr6:24900865:A/G x 2 = 2.84, df = 1, p = 0.09; hereafter referred to as KIAA0391rsP1 and KIAA0391rsP4, respectively). The specific SNVs used in the subsequent analyses are highlighted in Table 1.

KIAA0319 and gray matter covariation
For KIAA0319rsP1, the MANCOVA revealed a significant main effect for genotype F (16,246) = 2.086, p = 0.01. Subsequent univariate F-tests revealed significant differences between the alleles for SBM component 1 (C1) F (2,129) = 4.186, p = 0.017 and component 2 (C2) F (2,129) = 6.394, p = 0.002 (Fig. 1A). Post hoc analysis for C1 indicated that chimpanzees with the AA allele had higher weighted scores compared with AT and TT individuals, but the AT and TT means did not differ significantly from each other. For C2, post hoc analysis indicated that the means for the AA, AT, and TT groups all differed significantly from each other (Fig. 1B). No other SBM components differed across the KIAA0319rsP1 alleles. For KIAA0319rsP4, the results were very similar. The MANCOVA revealed a significant main effect for KIAA0319 genotype F (16,244) = 2.790, p = 0.007. Univariate F-tests revealed significant main effects of the SNV genotype on C1 F (2,128) = 3.810, p = 0.025 and C2 F (2,128) = 5.884, p = 0.004 (Fig. 1A). Post hoc analysis for component C1 indicated that chimpanzees with the minor AA allele had lower weighted scores compared with GG, but not GT individuals. By contrast, for C2, post hoc analysis indicated that AA had significantly higher values than GG but did not differ from AG apes (Fig. 1C). No other SBM components differed across the KIAA0319rsP4 alleles. For the KIAA0319 Glu563Asp variant, the MANCOVA revealed no significant main effects or interactions. Notwithstanding, the univariate F-tests revealed a significant main effect for the genotype on C7 F (2,1298) = 3.980, p = 0.046 ( Fig. 2A). Chimpanzees with the major AC allele had lower weighted scores than individuals with the CC but not AA allele (Fig. 2B). No other SBM components differed across the KIAA0319 Glu563Asp alleles. (Note, see Fig. 3 for a chimpanzee atlas map; Vickery et al., 2020.) KIAA0319 and microstructure in posterior superior temporal cortex (area Tpt) We ran a series of multivariate general linear models with each dependent measure of cortical area Tpt microstructure in the left and right hemisphere with sex, age, and brain mass as covariates, and all three KIAA0319 genotypes as fixed factors. Of these models, the only effect that approached a conventional statistical threshold was the KIAA0319rsP1 genotype on right hemisphere neuropil   fraction; F (2,10) = 10.266, p = 0.089) (see Table 2). The covariates sex, age, and brain mass did not have significant between-subjects effects. To further examine lateralization in cortical area Tpt microstructure, we also analyzed the AQ of each variable using general linear models with sex, age, and brain mass as covariates. Of these models, the only effect that approached a conventional statistical threshold was the KIAA0319rsP1 genotype on AQ of neuropil fraction; F (2,10) = 14.553, p = 0.064. The covariates sex, age, and brain mass did not have significant between-subjects effects. Photomicrographs and results can be seen in Figure 4.

Discussion
Exploration of polymorphic variation in KIAA0319 revealed a relatively variable coding sequence in chimpanzees. Across all four sub-species included in this study, a total of 29 SNVs were found spread across 12 of the 20 exons of the gene. Of these, a majority were nonsynonymous, thus causing amino acid changes, indicating low structural stability of the receptor protein in chimpanzees. Only a small proportion of SNVs were present in the flanking region (N = 4). This is in line with previous research showing that the most pronounced conservation across species in KIAA0319 extends from the transcription start Figure 1. A, 3D renderings of SBM C1 and C2 (from left to right: right, left, top, and bottom views), and mean (6SE) SBM weighted scores for each allele of (B) KIAA0319rsP1 and (C) KIAA0319rsP4. B, Chimpanzees with the AA allele had higher weighted SBM scores on C1 compared with AT and TT individuals but the AT and TT means did not differ significantly from each other. For component two, AA, AT, and TT groups all differed significantly from each other. C, For KIAA0319rsP4, chimpanzees with the minor AA allele had lower weighted SBM scores for C1 compared with GG but not GT individuals. By contrast, for C2, chimpanzees with AA had significantly higher values than GG but did not differ from AG chimpanzees. site to ;1 kb upstream (Dennis et al., 2009). In humans, several SNVs in the upstream region of the gene have been shown to have strong associations with reading disorders and dyslexia (Cope et al., 2005;Harold et al., 2006;Dennis et al., 2009;Elbert et al., 2011) and similarly, in our study, the strongest association with brain phenotypic differences was found for variants KIAA0319rsP1 and KIAA0319rsP4, despite a lack of overlap between human and chimpanzee promoter region SNVs. It thus appears for our chimpanzee population, that differential gene expression patterns are a likely driving factor behind this phenotypic variation, but further studies using gene expression assays are required to confirm this conclusion.
Findings from the SBM analysis showed that two SNVs in the promotor region of the KIAA0319 gene (KIAA0319rsP1 and KIAA0319rsP4), were associated with differences in gray matter volume of several brain regions. As hypothesized, these polymorphisms were associated with differences in gray matter covariation of the posterior superior temporal gyrus (including area Tpt and Heschl's gyrus, HG). This brain region is associated with a variety of cognitive functions including language comprehension and production and previous studies have reported that atypical language processing is associated with SNVs in KIAA0319 in humans (Hirayasu et al., 2000;Goulven and Tzourio-mazoyer, 2004;Dorsaint-Pierre et al., 2006;Warrier et al., 2009;Darki et al., 2012;Pinel et al., 2012;Jamadar et al., 2013;Friederici, 2015;Cardin et al., 2016;Eicher et al., 2016;Ozernov-Palchik and Gaab, 2016;Skeide et al., 2016).
The evidence that SNVs in KIAA0319 are associated with variation in gray matter within the posterior superior temporal gyrus in chimpanzees is a somewhat paradoxical finding in light of the evidence of its role in language and reading, aptitudes that are arguably uniquely human. In fact, Dumas et al. (2021) recently reported that KIAA0319 is under strong positive selection in humans compared with primate ancestors, and that this gene (and other communication-related genes) only diverged within the hominin lineage. At the most parsimonious level, the findings reported here suggest that although SNVs in KIAA0319 are directly implicated in the morphology of the posterior superior temporal gyrus, their link to the mapping of linguistic functions onto these brain regions is less direct. In other words, SNVs in KIAA0319 may be associated with both Tpt and HG morphology and specific kinds of language impairments, but the association is not causal, but rather mediating of other factors related to experience. Alternatively, the results reported here may indicate that there are more common neuropsychological processes that underlie individual variation in auditory-visual learning in human and nonhuman primates that are similarly influenced by SNV variation in KIAA0319 in both taxa. For instance, studies on statistical learning have shown that human and nonhuman primates are sensitive to temporal regularities in auditory and visual signals, which some have argued is the foundation for speech as well as syntactical processing (Hauser et al., 2001;Newport et al., 2004;Petkov and Jarvis, 2012;Fagot, 2017;Heimbauer et al., 2018). In great apes, several studies have reported impressive human speech comprehension abilities, including the processing of sentences presented in English only (Savage-Rumbaugh et al., 1993;Sevcik and Savage-Rumbaugh, 1994;Brakke and Savage-Rumbaugh, 1995). Finally, neuroimaging and lesion studies have shown that area Tpt, HG, and a number of other regions are involved in processing of species-specific vocalizations (Heffner and Heffner, 1984;Poremba et al., 2003Poremba et al., , 2004Gil-da-Costa et al., 2006;Petkov et al., 2008;Taglialatela et al., 2009). Based on these collective findings, it is intriguing to speculate that SNVs in KIAA0319 in nonhuman primates may underlie individual variation in auditory-visual learning and associated brain regions, although this hypothesis awaits additional research.
In addition, our microstructure analysis showed that KIAA0319rsP1 variants are associated with hemispheric differences in neuropil fraction of chimpanzees. This suggests that KIAA0319 expression may be involved in regulation of processes related to the formation and maintenance of synapses, dendrites, or axons in the microstructure of chimpanzee brains within a region directly involved in communication. Studies in rodents have shown that KIAA0319 plays a role in neural migration during formation of the cerebral neocortex, the development of dendritic arbors and spines, axon outgrowth, and has functional impacts on auditory processing (Peschansky et al., 2010;Platt et al., 2013;Centanni et al., 2014;Franquinho et al., 2017;Guidi et al., 2017). However, there are limited data examining KIAA0319 expression in the brains of humans or other primates Muntané et al., 2017). It should be noted that our findings are based on a relatively small sample size, as is common for postmortem histologic studies. Nevertheless, in combination with the SBM results, these microstructure data further support the conclusion that KIAA0319 influences neuroanatomy in a manner that is likely to be functionally significant. Future studies are needed to address whether the hemispheric differences in neuropil fraction are because of variation in synapses, dendritic morphology, or axon distributions. . Hemispheric differences in neuropil fraction from posterior superior temporal cortex (area Tpt) of chimpanzees genotyped for KIAA0319rsP1. Photomicrographs of Nissl-stained sections from Layer III of area Tpt in an AA genotype chimpanzee's left (A) and right (B) hemisphere, and a TT genotype chimpanzee's left (C) and right (D) hemisphere. After background correction, the image frames are converted to binary for measurement of neuropil fraction (E-H) where the black pixels represent cell profiles and white pixels represent the neuropil space. Results from the sample are displayed as box and whisker plots of the distribution of neuropil fraction values between hemispheres (I) and the AQ (J) according to KIAA0319 rsP1 genotype.