NEURODEVELOPMENT

Molecular and cellular reorganizationof neural circuits in thehuman lineageAndré M. M. Sousa,1* Ying Zhu,1* Mary Ann Raghanti,2 Robert R. Kitchen,3,4

Marco Onorati,1,5 Andrew T. N. Tebbenkamp,1 Bernardo Stutz,6 Kyle A. Meyer,1

Mingfeng Li,1 Yuka Imamura Kawasawa,1,7 Fuchen Liu,1 Raquel Garcia Perez,8

Marta Mele,8 Tiago Carvalho,8 Mario Skarica,1 Forrest O. Gulden,1 Mihovil Pletikos,1

Akemi Shibata,1 Alexa R. Stephenson,2 Melissa K. Edler,2 John J. Ely,9

John D. Elsworth,4 Tamas L. Horvath,1,6 Patrick R. Hof,10 Thomas M. Hyde,11

Joel E. Kleinman,11 Daniel R. Weinberger,11 Mark Reimers,12 Richard P. Lifton,13,14,15

Shrikant M. Mane,16 James P. Noonan,13 Matthew W. State,17 Ed S. Lein,18

James A. Knowles,19 Tomas Marques-Bonet,8,20,21 Chet C. Sherwood,22

Mark B. Gerstein,3 Nenad Sestan†1,4,6,13,23

To better understand the molecular and cellular differences in brain organization betweenhuman and nonhuman primates, we performed transcriptome sequencing of 16 regions ofadult human, chimpanzee, and macaque brains. Integration with human single-celltranscriptomic data revealed global, regional, and cell-type–specific species expressiondifferences in genes representing distinct functional categories. We validated and furthercharacterized the human specificity of genes enriched in distinct cell types throughhistological and functional analyses, including rare subpallial-derived interneurons expressingdopamine biosynthesis genes enriched in the human striatum and absent in the nonhumanAfrican ape neocortex. Our integrated analysis of the generated data revealed diversemolecular and cellular features of the phylogenetic reorganization of the human brain acrossmultiple levels, with relevance for brain function and disease.

Although the human brain is about threetimes as large as those of our closest livingrelatives, the nonhuman African great apes(chimpanzee, bonobo, and gorilla), increasedsize and neural cell counts alone fail to ex-

plain its characteristic functionalities (1–5). Thebrain has also undergone microstructural, con-nectional, and molecular changes in the humanlineage (1–5), changes likely mediated by diver-gent spatiotemporal gene expression (6–17 ).Here, we profiled themRNA and small noncod-

ing RNA transcriptomes of 16 adult brain regionsinvolved in higher-order cognition and behaviorof human (H) (Homo sapiens); chimpanzee (C)(Pan troglodytes), our closest extant relative; andrhesus macaque (M) (Macaca mulatta), a com-monly studied nonhuman primate.We integratedthese profiles with single-cell transcriptomic datafrom the human brain (18, 19), histological datafrom adult and developmental brains of theseand other primates (bonobo, gorilla, orangutan,

pig-tailedmacaque, baboon, and capuchin), andmultimodal data from human primary and in-ducedpluripotent stemcell (iPSC)–derived neuralcultures. In doing so, we have investigated theevolutionary, cellular, and developmental frame-work that makes the human brain unique.

Overview of regional transcriptome profiling

We generated transcriptional profiles of 247 tis-sue samples representing hippocampus, amygdala,striatum, mediodorsal nucleus of thalamus, cere-bellar cortex, and 11 areas of the neocortex fromsix humans, five chimpanzees, and five macaques(figs. S1 to S5 and table S1). To minimize biasesin comparative transcriptome analyses, we usedthe XSAnno pipeline to create a common annota-tion set of 26,514 orthologous mRNAs, including16,531 protein-coding genes and 3253 long inter-genic noncoding (linc) RNAs (fig. S2). We reanno-tated all chimpanzee and macaque microRNAs(miRNAs) based on annotated human precursor

sequences (fig. S3). Assessment of global correla-tion between regions and species by unsupervisedhierarchical clustering (fig. S5) revealed clusteringof themiRNA data set primarily by species. In con-trast, cerebellar mRNA samples from all speciesformedadistinct cluster separated fromotherbrainregions (fig. S5), indicating that the various cerebellaaremore similar to each other than to other brainregions within the same species. Within each spe-cies, hierarchical clustering of mRNA ormiRNAdata setswere calculated based on pairwise corre-lationmatrices of brain regions and confirmedbymultiscale bootstrap resampling and intraspeciesgenetic distance measurements (figs. S6 and S7).This revealed a similar pattern of interregional hi-erarchical clustering, reflecting known topograph-ical proximity and functional overlap (11, 14).

Species differences in regionalgene expression

Differentially expressed genes were identified[false discovery rate (FDR) < 0.01] in each regionby comparing generalized linear models with spe-cies as the main factor and batch as a cofactor.We found 25.9% of mRNAs (6866 of 26,514) and40.6% of miRNAs (603 of 1485 mature miRNAsincluded in the analysis) were differentially ex-pressed between at least two species in one ormore regions. A total of 11.9% of mRNAs (3154)and 13.6% of miRNAs (202), representing distinctfunctional categories, exhibited human-specificup-regulation (H > C = M) or down-regulation(H < C =M) (Fig. 1, A and B, and tables S2 to S4),with the highest number of differentially ex-pressed genes observed in striatum followed bythalamus, primary visual cortex, and dorsolateralprefrontal cortex (fig. S8 and table S3). These ob-servations were not attributable to variations inthe ratio ofmajor cell types among species (fig. S9).Among the 3154 mRNA genes with human-

specific differential expression, only 22 were up-regulated and9down-regulated across all analyzedregions (fig. S10). Only 3 geneswere differentiallyexpressedacrossanalyzedneocorticalareas:TWIST1(down-regulated), a transcriptional regulator ofneural genes that is mutated in Saethre-Chotzensyndrome, a disorder associated with intellec-tual disability (20) (Fig. 1C and fig. S11, A and B),and two functionally uncharacterized lincRNAs(RP11-364P22.1 andCTB-78F1.1) (up-regulated). Theremaining 3120 mRNA genes displayed human-specific differential expression in one or a subset ofbrain regions or neocortical areas (Fig. 1C, fig. S10,and table S3). AmongmiRNAs, 10.4% (155) and3.2% (47) were up-regulated or down-regulated,

RESEARCH

Sousa et al., Science 358, 1027–1032 (2017) 24 November 2017 1 of 6

1Department of Neuroscience and Kavli Institute for Neuroscience, Yale School of Medicine, New Haven, CT, USA. 2Department of Anthropology and School of Biomedical Sciences, Kent State University, Kent,OH, USA. 3Program in Computational Biology and Bioinformatics, Departments of Molecular Biophysics and Biochemistry and Computer Science, Yale University, New Haven, CT, USA. 4Department ofPsychiatry, Yale School of Medicine, New Haven, CT, USA. 5Department of Biology, Unit of Cell and Developmental Biology, University of Pisa, Pisa, Italy. 6Program in Integrative Cell Signaling and Neurobiologyof Metabolism, Department of Comparative Medicine, New Haven, CT, USA. 7Departments of Pharmacology and Biochemistry and Molecular Biology, Institute for Personalized Medicine, Pennsylvania StateUniversity College of Medicine, Hershey, PA, USA. 8Institut de Biologia Evolutiva, Consejo Superior de Investigaciones Científicas, Universitat Pompeu Fabra, Barcelona Biomedical Research Park, Barcelona,Catalonia, Spain. 9Alamogordo Primate Facility, Holloman Air Force Base, NM, USA. 10Fishberg Department of Neuroscience and Friedman Brain Institute, Icahn School of Medicine at Mount Sinai, New York, NY,USA. 11Lieber Institute for Brain Development, Johns Hopkins University Medical Campus, Baltimore, MD, USA. 12Neuroscience Program, Michigan State University, East Lansing, MI, USA. 13Department of Genetics,Yale School of Medicine, New Haven, CT, USA. 14Howard Hughes Medical Institute, Yale University, New Haven, CT, USA. 15Laboratory of Human Genetics and Genomics, The Rockefeller University, 1230 York Avenue,New York, NY 10065, USA. 16Yale Center for Genomic Analysis, Yale School of Medicine, New Haven, CT, USA. 17Department of Psychiatry and Langley Porter Psychiatric Institute, University of California, San Francisco,San Francisco, CA, USA. 18Allen Institute for Brain Science, Seattle, WA, USA. 19Department of Psychiatry and Zilkha Neurogenetic Institute, Keck School of Medicine, University of Southern California, Los Angeles,CA, USA. 20Institució Catalana de Recerca i Estudis Avançats, Barcelona, Catalonia, Spain. 21Centro Nacional de Analisis Genomico, Barcelona, Catalonia, Spain. 22Department of Anthropology, The GeorgeWashington University, Washington, DC, USA. 23Program in Cellular Neuroscience, Neurodegeneration, and Repair and Yale Child Study Center, Yale School of Medicine, New Haven, CT, USA.*These authors contributed equally to this work.†Corresponding author. Email: nenad.sestan@yale.edu

on May 25, 2020

http://science.sciencem

ag.org/D

ownloaded from

respectively, in the human brain, with manydisplaying region-specific patterns (Fig. 1B andfig. S12). Independently validated examplesinclude PKD2L1 (up-regulated in neocortical areasexcept primarymotor cortex), a gene encoding anion channel (21); MET (up-regulated in prefrontalcortex), a gene implicated in autism spectrum dis-order (22); ZP2 (up-regulated in cerebellum), agene encoding a protein mediating sperm-eggrecognition (23); and several miRNAs (Fig. 1Cand figs. S11 and S12).

Species differences in genecoexpression patterns

To extract additional biologically relevant infor-mation, we applied weighted gene coexpressioncorrelation network analysis (WGCNA) to gener-ate modules of genes with similar variation acrossregions and/or species. We identified 229 mRNAmodules, many of which exhibited regional and/or species-specific expression patterns (Figs. 2,A and B, and table S5). For example, genes inmodule 92 (M92) and M32 are respectively up-regulated and down-regulated in human neo-cortex, andM130 genes are up-regulated in humanstriatum, hippocampus, and amygdala (fig. S13).M130 includes tyrosine hydroxylase (TH) and

DOPA (3,4-dihydroxyphenylalanine) decarboxylase(DDC), both involved in dopamine biosynthesis(fig. S13F). Human-specific modules were en-riched for genes associated with categories andpathways such as “thrombospondin N-terminal–like domains” and “alternative splicing” (table S5).We also clustered all miRNAs based on their in-

dividual correlations to the average expression pro-file of each mRNAmodule (fig. S14A and table S6).Because the expression of each miRNA mightcorrelate with multiple mRNA modules, modulepairings were refined using a transcriptome-widehigh-throughput sequencing approach combinedwith cross-linking immunoprecipitation (HITS-CLIP) map of miRNA binding sites in the humanbrain (24) (fig. S14B and table S7). We identi-fied 37 stable miRNA modules, with several pairsof miRNA/mRNA modules exhibiting opposingregional and/or species-specific enrichment forpotential miRNA-mRNA target predictions (fig.S14, C to E).

Cell-type specificity of differentiallyexpressed genes

To investigate differential gene expression pat-terns at the cellular level, we integrated our datasets with single-cell RNA sequencing (RNA-seq)

data generated from the human neocortex (18, 19)and validated findings via immunohistochem-istry or in situ hybridization. We found that manyof the genes displaying species- and/or region-specific patterns also exhibited cell-type–specificexpression. For example, PKD2L1 is enriched inexcitatory projection neurons (Fig. 3, A and C), THis expressed in a subset of somatostatin (SST)–expressing inhibitory interneurons in human andmacaque neocortical deep layers and white matter(Fig. 3, B and D), and ZP2 is up-regulated in thegranule cells of the human cerebellum (fig. S11, Cand D). Additionally, we found cell-type–specificenrichment among WGCNA modules, includinghuman-specific M81 and M162, which were com-posed of genes enriched in a subset of neocorti-cal excitatory projection neurons (Fig. 2B, rightpanels, and table S5).

Species differences in neurotransmitterreceptor gene expression

The species- and region-specific expression pat-terns of several genes associated with neurotrans-mission prompted us to investigate whether therewere broad interspecies differences in the co-expression networks and genomic sequences ofgenes encoding receptors underlying excitatory,

Sousa et al., Science 358, 1027–1032 (2017) 24 November 2017 2 of 6

−10

−5

0

5

10

−10 −5 0 5 10Expression ratio (log2 [H/C])

Exp

ress

ion

ratio

(lo

g2 [H

/M])

H=C>M: 178

H=M<C: 85

H<C=M: 47 H=C<M: 94

H=M>C: 33

H>C=M: 155

H=C=M: 446

OFCDFC

VFCMFC M1C

S1CIPC

A1CSTC

ITCV1C

HIPAMY

STRMD

CBC

Neocortex (NCX)

Neocortex (NCX)

Differentiallyexpressed genes

H=C>M

H=M>C

H<C=M

H=C<M

H=M<C

H>C=M

H=C=M

(0 - 100]

(100 - 300]

(300-500]

(500 - 700]

(700 - 900]

(900 - 1,100]

(13,000 - 16,000]

OFCDFC

VFCMFC M1C

S1CIPC

A1CSTC

ITCV1C

HIPAMY

STRMD

CBC

PARM1ZP2

KCNK1CPXM2

DRD1C5orf4

TRPM3HNRNPA1

C1QL2ADORA2A

METTWIST1PKD2L1

CTB−78F1.1RBM8A

AC066593.1

−3 0 3

Human (log2 [RPKM+1])Human-specific

differential expressionlog2-fold difference (H/avg. of C & M)

Fig. 1. Interspecies differential gene expression across 16 brainregions. (A) Bubble matrix showing the number of mRNA genes withconserved expression (gray circles), species-specific up-regulation(filled circles), or down-regulation (open circles). Post hoc comparisonsare described in table S2. H, human; C, chimpanzee; M, macaque.(B) Interspecies patterns of normalized miRNA expression across allregions. Guidelines indicate ± 2-fold difference. (C) Examples of protein-coding and noncoding genes exhibiting global and regional human-specificup-regulation (red circles with black borders) or down-regulation(blue circles with black borders). Additional information and validationsare provided in figs. S10 to S12. MFC, medial prefrontal cortex;OFC, orbital prefrontal cortex; DFC, dorsolateral prefrontal cortex;VFC, ventrolateral prefrontal cortex; M1C, primary motor cortex; S1C,primary somatosensory cortex; IPC, inferior posterior parietal cortex;A1C, primary auditory cortex; STC, superior temporal cortex; ITC, inferiortemporal cortex; V1C, primary visual cortex; HIP, hippocampus; AMY,amygdala; STR, striatum; MD, mediodorsal nucleus of the thalamus;CBC, cerebellar cortex.

RESEARCH | RESEARCH ARTICLEon M

ay 25, 2020

http://science.sciencemag.org/

Dow

nloaded from

inhibitory, or modulatory signaling (fig. S15, Ato D). Gene coexpression networks of the cholin-ergic and serotonergic systems differed amongthe three species (figs. S15, E and F). Althoughthe dopaminergic system did not have enough

genes for reliable network construction, we foundthat DRD1, DRD2, and DRD3, genes encodingdopamine receptors, exhibited human-specificdown-regulation in striatum (fig. S10). By con-trast, excitatory glutamatergic and inhibitory

g-aminobutyric acid (GABA)–ergic systems’ genesexhibited conserved networks among species,and their coding sequences were more conservedthan the coding sequences of genes with similarexpression levels (figs. S15 to S17 and table S8).

Sousa et al., Science 358, 1027–1032 (2017) 24 November 2017 3 of 6

Fig. 2. Conserved and species-specific gene coexpression mod-ules. (A) Number of WGCNAmodules (numbers on gray background;see table S5) clustered by differen-tial expression across brain regions,species, and interspecies differen-ces across regions (interaction).Analysis of variance of eigengeneBonferroni-adjusted P < 0.01, solidline; ≥ 0.01, dashed line. (B) (Left)Enrichment of gene expressionfor modules (columns) in severalcell types (rows) based on humansingle-cell transcriptome data(18, 19), sorted by unsupervisedhierarchical clustering to show sim-ilarities among modules. (Right)Species-specific modules showinghuman (red), chimpanzee (blue), ormacaque (green) up-regulation(normal font) or down-regulation(italics) relative to the other twospecies exhibit distinct patterns ofcell-type–associated geneexpression.

39

1

26

1

113

12

22

15

229

162

67

40

27

125

37

Region

Species

Interaction

All modules

MicrogliaAstrocyteOligodendrocyteInterneuronPyramidal neuron

Excitatory projection neuron subtypes 1-8

Interneuron subtypes 1-8

0 2 4 6P-value (-log10)

M10

8

M14

M85

M22

3

M80

M10

M25

M10

3M

139

M18

4M

180

M96

M10

7M

115

M14

4M

147

M18

1M

208

M47

M79

M82

M92

M13

0M

215

M32

M86

M22

M11

6M

125

M19

6M

214

M30

M10

5M

152

M17

0M

175

M81

M16

2M

229

M69

M51

M14

0M

48M

149

M63

M10

6M

126

M17

8M

34M

89M

133

M14

3

Species-specific modules

log2

(fo

ld-c

hang

e)

FEZF2+excitatory neurons

VIP+ inhibitory interneurons

CUX1+excitatory neurons

SST+ inhibitory interneurons

0.5

1.5

2.5

0.5

1.5

2.5

0.3

0.6

0.9

0.2

0.4

0.6

ii

iii

iii

iii

iii

iii

iii

iii

i

COMT

CLCN4KCNN1

KCNN2

KCNJ2

KCNJ11

PKD2L1KCNC1

KCND3

GRIN1GRM3

GLRA3HRH3

AGAP1AHI1 UBE2HUBE3A

TSC2NLGN3

METNRGN

HCN4

KCND3

CLCN4KCNN2

KCNC1ADORA2A

CHRNA2GAD1GRIN1

HRH3HTR5AAGAP1

AHI1

COMTUBE2HUBE3A

TSC2

NLGN3NRGN

GLS

SHMT2

P2RX5

KCNT1

CACNA1DCACNA1G

GABRA5

GRIA1

INPP1ABAT

PRKCB

PPP1R1B

PGBD1GLS

CLCN2

KCNT1

TRPC6KCNK2

CACNA1D

CACNA1G

GABRA5

GRIA1

GLRA2

HTR1EABAT

PRKCBPGBD1

TH

PK

D2L

1

Human Chimpanzee Macaque

TH

Human Chimpanzee Macaque

Fig. 3. Cellular specificity of neocortical human and chimpanzee-specific differential expression. (A and B) Radar plots depicting neocorticalneuron cell-type enrichments of (A) human- or (B) chimpanzee-specificdifferences of genes associated with (i) neuropsychiatric disorders; (ii)neurotransmitter biosynthesis, degradation, and transport proteins; and (iii)encoding ion channels (table S10). Only genes expressed in the respectivecell type are plotted. The distance of each gene from the center representsdifferential expression between human and the average of chimpanzee

and macaque (red) or between chimpanzee and the other two species (blue).The direction of triangles denotes up- or down-regulation; filled trianglesrepresent cell-type–specific expression (Pearson correlations > 0.5). (C) In situhybridization shows that PKD2L1 is expressed in pyramid-shaped cell bodiesof excitatory projection neurons of human, but not chimpanzee or macaque,neocortex. (D) TH-immunopositive interneurons (filled arrowheads) are presentin neocortex of human and macaque, but not chimpanzee, where only TH+

midbrain dopaminergic axons (open arrowheads) are present. Scale bar, 30 mm.

RESEARCH | RESEARCH ARTICLEon M

ay 25, 2020

http://science.sciencemag.org/

Dow

nloaded from

Species differences in dopaminebiosynthesis gene expressionWe next investigated dopamine biosynthesis andsignaling genes. TH and DDC displayed human-specific (H > C = M) up-regulation in the striatum(Fig. 4A). TH also displayed chimpanzee-specificdown-regulation (C < H = M) in the neocortex(Fig. 4A). An extended analysis of RNA-seq data(25) independently validated the down-regulationof TH mRNA in chimpanzee neocortex comparedwith human, as well as the down-regulation ofTH expression in the neocortex of bonobo andgorilla, but not of orangutan (fig. S18A).Analyses of cis-regulatory elements active near

theTH gene in the adult human, chimpanzee, andmacaque brain (26) revealed no differences thatwould explain observed TH expression patterns.We hypothesized that the species-specific TH ex-pression patternsmight be explained by changesin the number and distribution of TH-expressinginterneurons, which have been previously iden-tified in telencephalic regions and shown to varyin distribution across species (27–29), includingdepletion in the prefrontal cortex of nonhumangreat apes (28). Therefore, we quantified TH-immunopositive (TH+) interneurons (fig. S19, Ato C) on an independent set of 45 adult brainsfrom nine primate species (table S9). Consistentwith our transcriptome data, humans have ahigher number (Tukey’s honest significance test;all P < 0.05) of TH+ interneurons in both thedorsal caudate nucleus and putamen (striatum)when compared with all other analyzed non-human primates (Fig. 4C and fig. S20A). Further-more, we found neocortical TH+ interneurons inall analyzed areas of human, all monkey species,and orangutan (Fig. 4C and fig. S20, B andD), butonly TH+ fibers in all analyzed neocortical areasof chimpanzee, bonobo, and gorilla (Fig. 4C andfig. S20B).We foundnodifferences in the numberof TH+ interneurons in human, chimpanzee, go-rilla, and macaque olfactory bulbs (fig. S19D).

Molecular profiling of humanTH+ interneurons

To further explore the phenotype of adult humanneocortical TH+ interneurons, we performedimmunohistochemistry and in situ hybridization.TH+ interneurons expressedGAD1, the GABA syn-thesis enzyme (Fig. 4B), but were lacking canon-ical markers of neocortical interneuron subtypessuch as SST, PVALB, NPY, NOS1, CALB2, andVIP (fig. S21), as well as ETV1, which is requiredfor differentiation of dopaminergic neurons inmultiple species (30), or its homolog, ETV5 (figs.S21, H and I). Most TH+/GAD1+ interneurons co-expressed DDC (62.54 ± 1.01%) (Fig. 4B), the en-zyme that converts L-DOPA to dopamine, but notdopamine-b-hydroxylase (DBH) (fig. S18) (31), theenzyme that converts dopamine to noradrenaline,indicating that a subset of TH+ interneurons areable to produce dopamine but not noradrenaline.

Developmental origin of humanTH+ interneurons

To gain insight into the development of TH+ in-terneurons, we analyzed the regional expression

of TH across human brain development usingthe BrainSpan RNA-seq data set (www.brainspan.org). The highest TH expression is observed instriatum and increases steadily from early fetaldevelopment [period 2, as defined in (11)] toyoung adulthood (period 13) (Fig. 5A). LowerTH expression is observed in neocortex, hippo-campus, and amygdala and increases perina-tally [periods 7 (late fetal development) and 8(early infancy)] and remains stable in neocortex.In addition, TH expression increases from early

childhood (period 10) to young adulthood in theamygdala and hippocampus (Fig. 5A).Using immunohistochemistry,we detectedTH+

axons in striatum as early as late midfetal devel-opment (fig. S22), and occasional bipolar TH+

interneurons were first observed in the externalcapsule and neocortical white matter in the new-born human (Fig. 5B). The neonatal chimpanzeebrain displayed the samepattern of unmyelinatedTH+ fibers in the external capsule (Fig. 5B), but noTH+ interneuronswere detected in the neocortex.

Sousa et al., Science 358, 1027–1032 (2017) 24 November 2017 4 of 6

TH

(lo

g2 [R

PK

M+

1])

OFCDFC

VFCM

FCM

1C S1C IPC

A1C STCIT

CV1C HIP

AMYSTR

MDCBC

Neocortex (NCX)

0.0

0.5

1.0

1.5

DD

C (

log2

[RP

KM

+1]

)

*

0.0

0.5

1.0

1.5

2.0

2.5

3.0 Human

Chimpanzee

Macaque

**

DDC GAD1 Merge + TH DAPI

MFC

% o

f TH

+ in

tern

euro

ns

Caudate

Putamen

Human

Chimpa

nzee

Gorilla

Rhesu

s m.

Pig-ta

iled

m.

Baboo

n

Capuc

hin

0

0.2

0.4

0.6

0.8

* *

0

0.5

1

1.5

2

2.5

3 *

*

0

0.5

1

1.5

2

2.5

3

M1C

* *0

0.2

0.4

0.6

0.8

0

STC

*0.2

0.4

0.6

Fig. 4. Human-specific expression of genes encoding dopamine biosynthesis enzymes. (A) THand DDC, respectively, showing higher expression in the human striatum (STR). TH is also down-regulated in the chimpanzee neocortex. Boxes represent quartiles and whiskers 1.5 times interquartilerange. Red and blue asterisks represent human-specific differential expression in striatum andchimpanzee-specific differential expression combining all neocortical areas, respectively (FDR < 0.01).(B) Immunofluorescence shows colocalization of TH, DDC, and GAD1 in adult human neocorticalinterneurons (arrowheads). Scale bar, 10 mm. (C) STR (caudate and putamen) shows an enrichmentof TH+ interneurons in human. MFC, M1C, and STC show a complete depletion of TH+ interneuronsin chimpanzee and gorilla. Asterisk represents Tukey’s honest significance test P < 0.05, comparinghuman or chimpanzee/gorilla with all other species.

RESEARCH | RESEARCH ARTICLEon M

ay 25, 2020

http://science.sciencemag.org/

Dow

nloaded from

To identify the birthplace of TH+ interneurons,we prepared primary cell cultures from 17 to 18postconceptional week (pcw)–old human brains(Fig. 5C) of lateral, medial, and caudal ganglioniceminences (LGE, MGE, and CGE) of the ventralforebrain (subpallium), known to generate inter-neurons (32–34), and neocortical proliferativezones, which may also generate interneurons inhumans (13). We found TH+ interneurons coex-pressing canonical markers of distinct progenitorlineages within ganglionic eminences (NKX2-1,NR2F2, or SP8) (Fig. 5D). BrdU (bromodeoxyuri-dine) birth-dating confirmed that TH+ interneu-rons are generated by ganglionic eminence, butnot neocortical, progenitors (Fig. 5, D and E),indicating that TH+ interneurons are derivedfrom diverse subpallial lineages and are devel-opmentally heterogeneous. Similar to adult neo-cortical TH+ interneurons, subpallial-derivedTH+ interneurons also coexpressed GAD1 andDDC (Fig. 5D). Neocortical TH+ interneurons weremainly SP8+ (77.78 ± 12.11%), with a smallerNR2F2+ (22.22 ± 8.78%) subpopulation and were

all BrdU–, indicating that they began to migrateinto neocortex before 17 pcw but express TH pro-tein later in development (Fig. 5, D and E).

In vitro characterization of humanTH+ interneurons

To further characterize TH+ interneuron devel-opment and properties, we asked whether TH+

interneurons could be generated from humaniPSCs using a differentiation protocol for corti-cal excitatory projection neurons and inhibitoryinterneurons (fig. S23) (see the supplementarymaterials). Immunofluorescence confirmed thepresence of TH+ cells coexpressing GAD1 andSP8, but not SST, PVALB, NR2F2, or NKX2-1,confirming that iPSC-derived TH+ interneu-rons display a similar molecular profile to TH+

interneurons from the adult neocortex and neo-cortical primary culture (fig. S24). Complementaryanalysis of a single-cell RNA-seq data set fromhuman embryonic stem cell–derived cortical in-terneurons (35) revealed that many TH-expressingcells coexpress GABAergic marker genes GAD1/2

at all time points, as well as SST, ETV1, and ETV5transiently at early time points (fig. S25).We characterized human iPSC-derived TH+ inter-

neurons by assessing their ability to produceand transport dopamine using immunofluores-cence, a monoamine uptake assay, and high-performance liquid chromatography. We foundthat 72.14 ± 10.02% of 80 days in vitro (DIV) TH+

interneurons that had taken up a monoamine-imitating fluorophore were DDC+ (Fig. 6, A and B)and consequently could produce and transportdopamine in vitro. Commensurate with these ob-servations, we detected dopamine in conditionedculture media from iPSC-derived and LGE pri-mary neural cultures, both of which containedTH+/DDC+ interneurons, but not in control culturemedia (Fig. 6C).

Discussion

Our analysis of transcriptomic data revealedglobal, regional, and cell-type–specific speciesexpression differences in protein-coding andnoncoding genes. Genes with human-specificdifferential expression patterns include thoseencoding transcription factors, ion channels,and neurotransmitter biosynthesis enzymes andreceptors. Changes in the regional and cellularexpression patterns of these genes could affectfunction of neural circuits by altering transcrip-tion of other genes, intrinsic electrophysiologicalproperties, or synaptic transmission.Neuromodulatory systems show broad expres-

sion differences between species. One exampleincludes a rare and molecularly heterogeneoussubpopulation of interneurons expressing dopa-mine biosynthesis genes TH and DDC, which areenriched in the human striatum andneocortex ascompared with nonhuman African apes. Thesecells originate in the subpallial ganglionic emi-nences and likely migrate into the striatum andneocortex during late prenatal and early post-natal development.We also observed an increasein TH expression during postnatal developmentand young adulthood, suggesting thatTH expres-sion and/or the migration of TH+ interneuronsmay be dynamically regulated and protracted.The absence of TH+ interneurons from the

cortex of nonhuman African apes [see also (28)],and their decreased density in the striatum ofnonhuman primates, may result from severalmechanisms. First, these cells could have beenlost due to genetic disruptions affecting interneu-ron migration, differentiation, or survival (32–34).These disruptions may have occurred in the com-mon ancestor of African apes before being re-versed in the human lineage (homoplasy) or, ina less-likely scenario, may have occurred inde-pendently in the Gorilla and Pan lineages. Asecond possibility is that these interneurons arepresent in the nonhuman African ape cortexbut do not express TH, do so only transiently, ordie before our ability to detect them. Commen-surate with this possibility, the molecular profileof mouse cortical SST-positive interneurons ismalleable (36), and sensory stimuli can cause aswitch from the production of TH and dopamineto SST in rat hypothalamic interneurons (37).

Sousa et al., Science 358, 1027–1032 (2017) 24 November 2017 5 of 6

C’ C’’

0

1

2

50 100 200 500 2000 10000Postconceptional days (log10)

TH

(lo

g2 [R

PK

M+

1])

AMYHIPNCX STR

Birth

Human (TH)Period: 3456 7 8 9 10 11 12 1321

C’

LGEMGE

NCX

C’’

LGE

CGE

NCX

STR

NCX

Chimpanzee (TH)

*STR

NCX

Ext

erna

l cap

sule

ST

R

THTHMBP

*

Ext

erna

l cap

sule

ST

R

THTHMBP

1

2

3

4

5

CGELG

EM

GENCX

% o

f TH

+/B

rdU

+ c

ells

0

NKX2-1TH

TH NR2F2

TH SP8

TH DDC BrdU

TH GAD1 BrdU

TH SP8 BrdU

Fig. 5. Human telencephalic TH+ interneurons are of subpallial origin and start to express THprotein perinatally. (A) TH expression in human neocortex (NCX), HIP, AMY, and STR throughoutdevelopment.The shaded area corresponds to a confidence interval of 50%. (B) Immunohistochemistryreveals TH+ axons in external capsule (arrowheads), STR, and NCX of newborn (38 pcw) humanand chimpanzee brains. Bipolar TH+ interneurons (filled arrowhead) are present in parallel withmyelin basic protein (MBP)+/TH– (arrows) and TH+/MBP– (open arrowheads) fibers in the externalcapsule. No TH+ cells were detected in chimpanzee external capsule. Scale bar, 1 cm. (C) Schematic ofdissection of ganglionic eminences [lateral (LGE), medial (MGE), and caudal (CGE)] and neocorticalproliferative zones (NCX) from mid-fetal brain for primary cell culture. (D) TH+ cells from ganglioniceminences also express NKX2-1, NR2F2, or SP8, and are BrdU+, DDC+, and GAD1+. TH+ interneurons inthe neocortical culture are SP8+, but BrdU– (bottom right). Scale bar, 20 mm. (E) Percentage ofTH+/BrdU+ cells in culture from MGE, LGE, CGE, and NCX. Error bars, SEM. Pairwise t tests wereperformed and corrected for multiple testing using Bonferroni correction. *P < 0.05; **P < 0.01.

RESEARCH | RESEARCH ARTICLEon M

ay 25, 2020

http://science.sciencemag.org/

Dow

nloaded from

Finally, TH+ interneurons of nonhuman Africanapes may have lost their ability to deviate to thecortex from the rostral migratory stream. Indeed,some human TH+ interneurons migrating via therostral migratory stream to the olfactory bulbdivert to the prefrontal cortex (38), and our ob-servation of SP8+/TH+ coexpression is consistentwith a rostral migratory stream origin. However,other routes of migration are possible, as sug-gested by our observation of TH+ interneuronsin the external capsule of newborn human brain.Neuromodulatory transmitters, in particular

dopamine, are involved in distinctly human as-pects of cognition and behavior, such as workingmemory, reasoning, reflective exploratory behav-ior, and overall intelligence. By analyzing brainregions involved in these processes, we show thatevolutionarymodifications in gene expression andthe distribution of neurons associatedwithneuro-modulatory systems may underlie cognitive andbehavioral differences between species. CorticalTH+ interneurons are depleted in patients af-fected by Parkinson’s disease (39) or dementiawith Lewy bodies (40), and these alterationsmaycontribute to cognitive impairments.

As these results demonstrate, the resource wepresent here may aid future studies on the evo-lution and neuroscience of primates.

REFERENCES AND NOTES

1. T. M. Preuss, in The Cognitive Neurosciences, M. S. Gazzaniga,Ed. (MIT Press, Cambridge, MA, 2004), pp. 5-22.

2. C. C. Sherwood, F. Subiaul, T. W. Zawidzki, J. Anat. 212,426–454 (2008).

3. K. Teffer, K. Semendeferi, Prog. Brain Res. 195, 191–218(2012).

4. M. Gabi et al., Proc. Natl. Acad. Sci. U.S.A. 113, 9617–9622(2016).

5. A. M. M. Sousa, K. A. Meyer, G. Santpere, F. O. Gulden,N. Sestan, Cell 170, 226–247 (2017).

6. M. C. King, A. C. Wilson, Science 188, 107–116 (1975).7. M. Cáceres et al., Proc. Natl. Acad. Sci. U.S.A. 100,

13030–13035 (2003).8. M. Uddin et al., Proc. Natl. Acad. Sci. U.S.A. 101, 2957–2962

(2004).9. P. Khaitovich et al., Science 309, 1850–1854 (2005).10. C. C. Babbitt et al., Genome Biol. Evol. 2, 67–79 (2010).11. H. J. Kang et al., Nature 478, 483–489 (2011).12. G. Konopka et al., Neuron 75, 601–617 (2012).13. D. H. Geschwind, P. Rakic, Neuron 80, 633–647 (2013).14. M. Pletikos et al., Neuron 81, 321–332 (2014).15. B. I. Bae, D. Jayaraman, C. A. Walsh, Dev. Cell 32, 423–434 (2015).16. D. L. Silver, BioEssays 38, 162–171 (2016).17. E. S. Lein, T. G. Belgard, M. Hawrylycz, Z. Molnár, Annu. Rev.

Neurosci. 40, 629–652 (2017).

18. S. Darmanis et al., Proc. Natl. Acad. Sci. U.S.A. 112, 7285–7290(2015).

19. B. B. Lake et al., Science 352, 1586–1590 (2016).20. T. D. Howard et al., Nat. Genet. 15, 36–41 (1997).21. A. L. Huang et al., Nature 442, 934–938 (2006).22. D. B. Campbell et al., Proc. Natl. Acad. Sci. U.S.A. 103,

16834–16839 (2006).23. T. Hoodbhoy, J. Dean, Reproduction 127, 417–422 (2004).24. R. L. Boudreau et al., Neuron 81, 294–305 (2014).25. D. Brawand et al., Nature 478, 343–348 (2011).26. M. W. Vermunt et al., Nat. Neurosci. 19, 494–503 (2016).27. B. Berger, C. Verney, P. Gaspar, A. Febvret, Brain Res. 23,

141–144 (1985).28. M. A. Raghanti et al., Neuroscience 158, 1551–1559 (2009).29. R. Benavides-Piccione, J. DeFelipe, J. Anat. 211, 212–222

(2007).30. N. Flames, O. Hobert, Nature 458, 885–889 (2009).31. P. Gaspar, B. Berger, A. Febvret, A. Vigny, J. P. Henry, J. Comp.

Neurol. 279, 249–271 (1989).32. D. V. Hansen, J. H. Lui, P. R. L. Parker, A. R. Kriegstein, Nature

464, 554–561 (2010).33. T. Ma et al., Nat. Neurosci. 16, 1588–1597 (2013).34. A. Kepecs, G. Fishell, Nature 505, 318–326 (2014).35. J. L. Close et al., Neuron 93, 1035–1048.e5 (2017).36. N. Dehorter et al., Science 349, 1216–1220 (2015).37. D. Dulcis, P. Jamshidi, S. Leutgeb, N. C. Spitzer, Science 340,

449–453 (2013).38. N. Sanai et al., Nature 478, 382–386 (2011).39. T. Fukuda, J. Takahashi, J. Tanaka, Neuropathology 19, 10–13

(1999).40. W. Marui, E. Iseki, M. Kato, K. Kosaka, Neurosci. Lett. 340,

185–188 (2003).

ACKNOWLEDGMENTS

Data are available at National Center for Biotechnology InformationBioProjects (accession number PRJNA236446). We thankA. Bauernfeind, M. Horn, D. Singh, G. Terwilliger, I. B. Toxopeus,B. Wicinski, and S. Wilson for assistance with tissue acquisitionand processing, and the Alamogordo Primate Facility and thePrimate Brain Bank, Netherlands Institute for Neuroscience, forproviding primate tissue. Data was generated as part of thePsychENCODE Consortium, supported by MH103339, MH106934,and MH110926. Additional support was provided by NIH grantsMH109904, MH106874, AG048918, DK111178, and NS092988(National Chimpanzee Brain Resource), the Kavli Foundation, theJames S. McDonnell Foundation, NSF grant BCS-1316829, MINECOBFU2014-55090-P (FEDER), Howard Hughes International EarlyCareer, and Secretaria d’Universitats i Recerca del Departamentd’Economia i Coneixement de la Generalitat de Catalunya. Thesupplementary materials contain additional data.

SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/358/6366/1027/suppl/DC1Materials and MethodsFigs. S1 to S25Tables S1 to S11References (41–77)

30 March 2017; accepted 17 October 201710.1126/science.aan3456

Sousa et al., Science 358, 1027–1032 (2017) 24 November 2017 6 of 6

DDC+

DDC-

TH DDC TH DDC DAPI

0.20

0.15

0.10

0.0520

40

60

Neu

rons

(%

) 80

Fluorophore

Dop

amin

e (p

g/µl

)

Ctrl

iPSC

LGE

100

00

***

Fig. 6. Human telencephalic TH+ interneurons synthesize and transport dopamine in vitro.Human iPSC-derived neurons were incubated with a fluorophore-labeled synthetic monoamine.(A) TH+ (red) and DDC+ (blue) immunolabeled interneurons (arrowheads) that transported monoamine-imitating fluorophore (green) in vitro. Scale bar, 10 mm. (B) Percentage of neurons that took up thefluorophore and were positive for both the uptake assay and TH. This population is composed of DDC+

(blue) or DDC– (red) interneurons. (C) Concentration of dopamine detected by high-performanceliquid chromatography in the unused [control (Ctrl)] cell culture medium and the conditioned mediafrom LGE and iPSC-derived cultures. Error bars, SEM. Dunnett’s test, ***P < 0.001.

RESEARCH | RESEARCH ARTICLEon M

ay 25, 2020

http://science.sciencemag.org/

Dow

nloaded from

Molecular and cellular reorganization of neural circuits in the human lineage

Marques-Bonet, Chet C. Sherwood, Mark B. Gerstein and Nenad SestanRichard P. Lifton, Shrikant M. Mane, James P. Noonan, Matthew W. State, Ed S. Lein, James A. Knowles, TomasD. Elsworth, Tamas L. Horvath, Patrick R. Hof, Thomas M. Hyde, Joel E. Kleinman, Daniel R. Weinberger, Mark Reimers, Mario Skarica, Forrest O. Gulden, Mihovil Pletikos, Akemi Shibata, Alexa R. Stephenson, Melissa K. Edler, John J. Ely, JohnStutz, Kyle A. Meyer, Mingfeng Li, Yuka Imamura Kawasawa, Fuchen Liu, Raquel Garcia Perez, Marta Mele, Tiago Carvalho, André M. M. Sousa, Ying Zhu, Mary Ann Raghanti, Robert R. Kitchen, Marco Onorati, Andrew T. N. Tebbenkamp, Bernardo

DOI: 10.1126/science.aan3456 (6366), 1027-1032.358Science 

, this issue p. 1027Scienceneurodevelopmental disorders.nonhuman African apes. Such differences in neuronal transcriptional programs may underlie a variety ofsystems. Furthermore, the dopaminergic interneurons found in the human neocortex were absent from the neocortex of those encoding transcription factors, could alter transcriptional programs. Others were associated with neuromodulatorywhat makes human brains different from those of nonhuman primates. Various differentially expressed genes, such as

overlaid transcriptome and histological analyses to seeet al.tells us that surface similarity is not the whole story. Sousa Although nonhuman primate brains are similar to our own, the disparity between their and our cognitive abilities

The makings of the primate brain

ARTICLE TOOLS http://science.sciencemag.org/content/358/6366/1027

MATERIALSSUPPLEMENTARY http://science.sciencemag.org/content/suppl/2017/11/21/358.6366.1027.DC1

REFERENCES

http://science.sciencemag.org/content/358/6366/1027#BIBLThis article cites 70 articles, 13 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.ScienceScience, 1200 New York Avenue NW, Washington, DC 20005. The title (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

Copyright © 2017, American Association for the Advancement of Science

on May 25, 2020

http://science.sciencem

ag.org/D

ownloaded from