Content uploaded by Jeroen Pasterkamp
Author content
All content in this area was uploaded by Jeroen Pasterkamp on Feb 24, 2023
Content may be subject to copyright.
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 134
nature reviews neuroscience
Review article
https://doi.org/10.1038/s41583-022-00669-3
Check for updates
Development, wiring and function
of dopamine neuron subtypes
Oxana Garritsen 1,2, Eljo Y. van Battum1,2, Laurens M. Grossouw1,2 & R. Jeroen Pasterkamp 1
Abstract
The midbrain dopamine (mDA) system is composed of molecularly and
functionally distinct neuron subtypes that mediate specic behaviours
and are linked to various brain diseases. Considerable progress has
been made in identifying mDA neuron subtypes, and recent work has
begun to unveil how these neuronal subtypes develop and organize
into functional brain structures. This progress is important for
further understanding the disparate physiological functions of mDA
neurons and their selective vulnerability in disease, and will ultimately
accelerate therapy development. This Review discusses recent
advances in our understanding of molecularly dened mDA neuron
subtypes and their circuits, ranging from early developmental events,
such as neuron migration and axon guidance, to their wiringand
function, and future implications for therapeutic strategies.
Sections
Introduction
Identiication of mDA neuron
subtypes
Developing dopamine neuron
diversity
Subtype-speciic wiring and
function
Establishing mDA neuron
subtype connectivity
Selective vulnerability
in disease
Conclusions and future
perspectives
1Department of Translational Neuroscience, UMC Utrecht Brain Center, University Medical Center, Utrecht
University, Utrecht, Netherlands. 2These authors contributed equally: Oxana Garritsen, Eljo Y. van Battum,
Laurens M. Grossouw. e-mail: r.j.pasterkamp@umcutrecht.nl
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 135
Review article
defined by tyrosine hydroxylase gene (Th) expression, and can be sub-
divided into roughly two cell populations: Sox6+ cells or Slc17a6+ cells
(Slc17a6 is also known as Vglut2, which encodes vesicular glutamate
transporter 2). Noteworthy, during development Vglut2 is transiently
expressed in most mDA neurons, while it becomes more restricted at
later stages
38
. Sox6
+
neurons can be further partitioned into an Aldh1a1
+
subtype (subtype 1: Sox6+Aldh1a1+) and an Aldh1a1− subtype (subtype 2:
Sox6+Aldh1a1−). Vglut2+ neurons are Sox6− and can be subdivided
into a subtype located in the substantia nigra pars lateralis (SNpl)
(subtype 3: Vglut2+Aldh1a1−) and one located in the VTA (subtype 4:
Vglut2
+
Aldh1a1
−
). In addition, there are Sox6
−
Aldh1a1
−
(possibly Vglut2
+
)
neurons in the dorsal SNc (dSNc)
39
. It is, however, unclear whether this
is a unique subtype or whether these neurons are similar to subtype 4
VTA neurons (probably SN_4-4 inref.
17
)
17,39
. A third Vglut2
+
subtype
(subtype 5: Vglut2+Vgat+ (Vgat is also known as Slc32a1)) is a subset of
combinatorial neurons located at the border of the SNc and the VTA.
A fourth Vglut2
+
subtype (subtype 6: Vglut2
+
Aldh1a1
+
Otx2
+
) consists
of neurons that are spread throughout the parabrachial pigmented
nucleus, paranigral nucleus, interfascicular nucleus and caudal lin-
ear nucleus. Finally, a subtype of Vglut2
+
Vip
+
neurons (subtype 7; Vip
encodes vasoactive intestinal polypeptide) is present in the caudal
linear nucleus, periaqueductal grey and dorsal raphe nucleus. Undoubt-
edly this elegant classification will be further defined in the future
on the basis of ongoing profiling, tracing and functional studies. For
example, different clustering pipelines have already revealed addi-
tional subtypes17,21,39 within the clusters defined in Fig.1. For instance,
Neurod6+Grp+ neurons (Grp encodes gastrin-releasing peptide) can be
divided into at least three molecularly distinct sub-subtypes
21
(Fig.1a).
In addition, several recently identified mDA neuron subtypes do not
adhere to the classification discussed above; for example, Grp
+
and
Bcl11a
+
mDA neurons are present in several clusters in both the VTA
and the SNc21,40. Recently, scRNA-seq of rat VTA revealed the existence of
two neuron clusters expressing mDA markers
41
. However, comparison
with the mouse data summarized above did not reveal obvious overlap
between neuron subtypes in the two rodent species. This could be
explained by the relatively small number of mDA neurons detected
in the rat study, and emphasizes the fact that larger numbers of mDA
neurons will need to be used in future scRNA-seq studies to clearly
define and compare distinct subtypes between species.
Previous work examining the heterogeneity of the human mDA
system used bulk sequencing of human midbrain samples and/or his-
tological approaches (for example, in situ hybridization)
41–47
. Although
these techniques do not have the same cellular and molecular resolu-
tion as more recent scRNA-seq approaches, these data suggested that
also the human mDA system can be divided into different neuron sub-
types characterized by the expression of specific molecular markers.
Mapping of immunohistochemical expression data of marker genes
(CALB1, PITX3, DCC and GIRK2, which encodes G protein-gated inwardly
rectifying potassium channel2 and is also known as KCNJ6)9,48,49 onto
the human midbrain revealed molecularly distinct neuron subtypes
that in part are similar to those described in the mouse
9,43,50–52
(Fig.1c).
For example, the expression pattern of ALDH1A1, SOX6 and CALB1 in the
human midbrain suggests that humans and mice have similar neuron
subtypes in the SNc (subtypes 1 and 2 in mice) (Fig.1a,c).
Human scRNA-seq studies on mDA neurons confirmed the hetero-
geneity of the mDA system and the presence of subtypes that are found
in other species18,46,47,53. In 2016, the first human scRNA-seq study was
performed on human embryos, and revealed similarities in develop-
mental subsets of mDA neurons between human embryos and mouse
Introduction
Midbrain dopamine (mDA) neurons in the ventral midbrain (vMB) are
involved in the regulation of complex behaviours and locomotion, and
are affected in multiple brain disorders
1
. Historically, mDA neurons
have been classified into three distinct regions primarily on the basis of
anatomical and cellular features: A8 (retrorubral field), A9 (substantia
nigra pars compacta (SNc)) and A10 (ventral tegmental area (VTA))
2
.
mDA neurons in the SNc and VTA arethe most intensively studied and
are the focus of this Review.
SNc mDA axon projections to the striatum form the nigrostriatal
pathway that regulates voluntary movements
2
. A large part of these
axons and eventually their corresponding cell bodies degenerate in
Parkinson disease (PD)2–4. VTA mDA neurons coordinate rewarding
and aversive stimuli, decision-making and working memory, and devia-
tions in VTA mDA signalling have been implicated in disorders such as
addiction and schizophrenia
5–7
. VTA mDA neurons send prominent
projections to the ventral striatum and prefrontal cortex (PFC) in the
mesolimbic pathway and the mesocortical pathway, respectively. Histo-
logical and anatomical studies have revealed a further subdivision of the
SNc and VTA into different subnuclei, both in rodents and in humans
8,9
.
Molecular approaches such aslaser capture microdissection combined
with gene expression profiling helped to establish the general molecular
identity of neurons in the VTA versus the SNc
10–14
. More recently, single-
cell profiling of the vMB has begun to map molecular heterogeneity
within the VTA and SNc, and several molecularly defined subtypes of
mDA neurons (and of other (non-neuronal) subtypes in the vMB) have
been described
15–21
. In parallel to these studies, mDA neuron subsets are
also being defined using axonal tracing or functional approaches
5,21–37
.
Integration of data from these various approaches is still incomplete.
This Review focuses on recent advances in the identification
of mDA neuron heterogeneity and the molecular mechanisms that
dictate mDA neuron subtype specification, positioning and target
innervation during development. We discuss axonal wiring mecha-
nisms and biological functions of mDA neuron subtypes and highlight
opportunities for further understanding and treating diseases related
to mDA neurons. As outlined above, integration of data gathered using
various approaches to determine mDA neuron subtypes is still incom-
plete. Therefore, we focus on mouse mDA neuron subtype classifica-
tion based on single-cell profiling studies for this Review (see Fig.1).
The recent identification of (combinatorial) molecular markers for
distinct mDA neuron subtypes has allowed their select visualization
and has led to progress in our understanding of the molecular and
cellular mechanisms underlying mDA neuron subtype development.
Identiication of mDA neuron subtypes
Advances in transcriptomic profiling of individual cells have facilitated
the identification of molecularly defined subtypes of mDA neurons
in the anatomically defined regions of the mDA system in mice
15–18,20,21
. In
a first study in 2014, a set of 96 genes was examined in single mouse
neurons, which were shown to be differentially expressed between the
VTA and the SNc20. On the basis of these gene expression patterns, six
distinct subtypes of mDA neurons were proposed (four in the VTA and
two in the SNc). These findings were supported and complemented by
data from subsequentsingle-cell RNA sequencing (scRNA-seq) stud-
ies
15–18,21
. As the annotation of cell clusters differed between different
studies, a unifying classification was proposed based on differential
marker gene expression and cellular localization. According to this
proposal, the adult mouse mDA system is composed of at least three
SNc and four VTA subtypes19 (Fig.1a,b). All mDA neuron subtypes are
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 136
Review article
a
b
c
Mouse
Aldh1a1
Sox6
Scna
Vgat
Vglut2
Otx2
Cck
Calb1
Th
Grp
Neurod6
Dat
Vip
ALDH1A1, PITX3, DCC, GIRK2
PITX3, DCC, GIRK2, CALB1
CALB1, ALDH1A1, PITX3, DCC, GIRK2 CALB1, PITX3, DCC, GIRK2
SOX6 , ALDH1A1, PITX3, DCC, GIRK2
SOX6 , CALB1, PITX3, DCC, GIRK2 SOX6 , ALDH1A1, CALB1, PITX3, DCC, GIRK2
Rostral Caudal
Human
Rostral Caudal
SNr
MRN
RN
RR
rust
mlf
PAG
Aq
III
RL
SC
NB PAG
EW
dtd
tspc
mtg
SNc
IF
CLi
Ctx
RN
PBP
mSNc
dSNc
vSNc
SNr
lSNc
VTA
RLi
PN
N1
RN
Ctx
dSNc
lSNc
mSNc
vSNc
vSNc
vSNc
PBP PAP
N5
Marker genes
MM
SNr
Ctx
fr
cpd
mb
ml
MRN
mtg
FF
SPFp
ZI
SNpl
vSNc
dSNc
1. Sox6+Aldh1a1+
2. Sox6+Aldh1a1–
3. Vglut2+Aldh1a1–5. Vglut2+Vgat+
6. Vglut2+Aldh1a1+Otx2+
7. Vglut2+Vip+
Neurod6+Grp+
Neurod6–Grp+Neurod6+Grp–
Bcl11a+
Sox6– (dSNc)4. Vglut2+Aldh1a1–
1. Sox6+Aldh1a1+
2. Sox6+Aldh1a1–
3. Vglut2+Aldh1a1–
4. Vglut2+Aldh1a1–
5. Vglut2+Vgat+
6. Vglut2+Aldh1a1+Otx2+
7. Vglut2+Vip+
PBP
mSNc
dSNc
vSNc
vSNc
vSNc
SNr
Ctx
RRF
ml xscp
PN
N2
N3
N1
SNr
ml
Ctx cpd IPN
RN
MRN
THL INC
ND
APN
vtd
mtg RL
PAG
rust
EW
SNpl
dSNc
vSNc
PN
PIF IF
PBP
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 137
Review article
embryos18. Although mouse embryos and human embryos are generally
comparable, minor differences were highlighted between them, includ-
ing differences in marker expression and the number of immature mDA
cells. Most scRNA-seq studies of adult human mDA neurons or midbrain
have thus far focused on the SNc, because of its relevance for PD46,47,53
(Box1). Interestingly, a very recentsingle-nucleus RNA sequencing
(snRNA-seq) study on NR4A2
+
(also known as NURR1
+
) neurons defined
ten molecularly distinct subtypes in the human SNc, of which six sub-
types were CALB1+ and four were SOX6+ (ref.46). Cross-species analysis
of these data identified eight SOX6
+
or CALB1
+
clusters with a clear
anatomical separation. One cell cluster (CALB_GEM), characterized by
the expression of FAM83B and GEM, was specific for primates. In situ
hybridization and Slide-seq spatial transcriptomic analysis
54
localized
this primate-specific subtype in the dSNc, which is substantially more
elaborated in primates, and confirmed the separation between CALB1
and ALDH1A1 in the dSNc and ventral SNc (vSNc), respectively.
In conclusion, significant progress has been made in the identifica-
tion of molecularly defined subtypes of mDA neurons in both mice and
humans. Future studies are expected to further define this initial clas-
sification in different species on the basis of expression of molecular
markers but also on the basis of wiring patterns and functions assigned
to these subtypes.
Developing dopamine neuron diversity
A plethora of studies has established the developmental programmes
that are necessary for the proper development of the mDA system as
a whole. Given the focus of this Review, we only briefly discuss these
more general developmental events in the midbrain, followed by
a more extensive description of subtype-specific mechanisms. For a
more detailed overview of all key players involved in mDA neuron
development, we refer the reader to the reviews in refs.2,55–57. mDA
neurons are generated in theloorplate, where antagonistic expression
of Otx2 and Gbx2 establishes the midbrain–hindbrain boundary
58–62
.
Expression of Otx2 induces Wnt1 expression, whereas Gbx2 induces
Fgf8 expression58,63–65. Diffusion of Fgf8 in the floorplate region causes
it to obtain a midbrain fate by stimulating Shh expression and subse-
quently the expression of Foxa1/Foxa2 (refs.66–70). Overlap of expres-
sion of Foxa1/Foxa2 and Otx2 leads to the induction of Lmx1a/Lmx1b,
which further specifies mDA neuron progenitors62,69,71–73. Subsequently,
mDA neuron progenitors start to express Ngn2 (also known as Neurog2,
which encodes neurogenin2) and become postmitotic neurons
73,74
.
Simultaneously, mDA neurons migrate to their final position and
innervate target structures75,76. Combinatorial expression of Wnt1,
Lmx1a/Lmx1b and Foxa1/Foxa2 induces the upregulation of genes
necessary for mDA neuron maintenance and function, including Nurr1,
Pitx3, Ddc (which encodes DOPA decarboxylase), Th, Slc6a3 (also known
as Dat, which encodes dopamine transporter) and En1 (refs.73,74,77–80).
Generating mDA neuron diversity
Studies investigating subtype-specific mDA neuron development ini-
tially focused on VTA and SNc neurons in general, but more recent work
has also begun to dissect developmental mechanisms instructing the
Fig. 1 | Anatomical distribution of molecular mDA subtypes in the adult
rodent and human midbrain. a, Coronal images showing midbrain dopamine
(mDA) neurons in the mouse midbrain, highlighted in orange, at different
rostrocaudal levels. Below, higher magnifications of the boxed areas in the
coronal sections depict the anatomical distribution of molecularly defined mDA
neuron subtypes in the mDA system. This overview is based on several recent
studies21,39,40 and a review by Poulin et al.19, and serves as a basis for discussion of
mDA neuron subtype-specific development and function in this Review. Subtype 1
neurons are Th+Sox6+Aldh1a1+ (blue) and are located in the ventral substantia
nigra pars compacta (vSNc). Subtype 2 neurons are classified as Th+Sox6+Aldh1a1−
(yellow) and are located in the dorsal substantia nigra pars compacta (dSNc)
and parabrachial pigmented nucleus (PBP). Subtype 3 and subtype 4 neurons
are both categorized as Th+Vglut2+Aldh1a1−, but can be distinguished on the
basis of anatomical location (subtype 3 (green) is located in the substantia nigra
pars lateralis (SNpl), whereas subtype 4 (red) is located in the ventral tegmental
area (VTA)). An additional, Th+Sox6− (possibly Vglut2+) subtype was described
in the dSNc39, but it is unclear whether this is a unique subtype or whether it
represents subtype 4 neurons located in the substantia nigra (probably SN_4-4
inref.17)17,39. Subtype 5 neurons (Th+Vglut2+Vgat+, purple) consist of a small
subset of combinatorial neurons located at the border of the SNc and the VTA,
intermingling with subtype 4 neurons in a medial (subtype 4) to lateral (subtype 5)
gradient (not depicted). Subtype 6 neurons are Th+Vglut2+Aldh1a1+Otx2+ (orange)
and are spread throughout the PBP, paranigral nucleus (PN), interfascicular
nucleus (IF) and caudal linear nucleus (CLi). This group probably represents
multiple subtypes, as most of these neurons express Calb1 and Cck, and sub-
subtypes contain Grp and/or Neurod6 (ref.21). Subtype 7 (Th+Vglut2+Vip+, lilac)
neurons are located in the CLi, periaqueductal grey (PAG) and dorsal raphe
nucleus. An additional subtype of Bcl11a+ neurons, not confined to one
of the other subtypes or anatomical regions, has recently been described40.
It is important to note that although within a particular subtype most neurons
share expression of specific genes (for example, VTA subtypes are Vglut2+ and
subtype 7 neurons are Vip+), individual neurons may (partially) deviate from
this general molecular subtype profile (for example, Th+Vglut2− neurons are
also present in the VTA and Th+Vglut2+Vip− neurons are also present in the CLi,
PAG and dorsal raphe nucleus). Establishing whether this deviation represents a
further subdivision of mDA neuron subtypes requires future study. b, Expression
of marker genes shared between mouse mDA neuron subtypes. c, Schematic
interpretation (similarly to part a) of putative mDA neuron subtypes in coronal
sections of human midbrain based on integration of immunohistochemical
expression data of marker genes (ALDH1A1, SOX6, CALB1, PITX3, DCC and
GIRK2) reported in several bulk profiling studies9,39,48,49 and mapped onto a
schematic human brain atlas. Nigrosomes (N1–N5) are based on ref.52. Molecular
subtypes of mDA neurons in the human midbrain are poorly defined relative
to those in rodents, and in most cases have not been histologically validated.
Although a recent single-nucleus RNA sequencing study confirmed the relative
dorsal–ventral positioning of SOX6+ and CALB1+ subtypes in human substantia
nigra46, these substantia nigra subtypes were not compared with subtypes
emerging from, for example, bulk profiling studies and have therefore been
omitted from the schematic illustration. APN, anterior pretectal nucleus;
Aq, cerebral aqueduct; cpd, cerebral peduncle; Ctx, cortex; dtd, dorsal tegmental
decussation; EW, Edinger–Westphal nucleus; FF, fields of Forel; fr, fasciculus
retroflexus; III, oculomotor nerve; INC, interstitial nucleus of Cajal; IPN,
interpeduncular nucleus; lSNc(human), lateral substantia nigra pars compacta;
mfb, medial forebrain bundle; ml, medial lemniscus; mlf, medial longitudinal
fascicle; MM, medial mammillary nucleus; MRN, midbrain reticular nucleus; mtg,
mammillotegmental tract; mSNc, medial substantia nigra pars compacta; NB,
nucleus of the brachium of the inferior colliculus; ND, nucleus of Darkschewitsch;
PAP (human), parapeduncular nucleus; PIF, parainterfascicular nucleus;RL(i),
rostral linear nucleus of the raphe; RN, red nucleus; RR, midbrain reticular
nucleus, retrorubral area; RRF, retrorubral field; rust, rubrospinal tract; SC,
superior colliculus; SNr, substantia nigra pars reticulata; SPFp, subparafascicular
nucleus, parvicellular part; THL, thalamus; tspc, crossed tectospinal pathway;
vtd, ventral tegmental decussation; xscp (human), decussation of superior
cerebellar peduncle; ZI, zona incerta.
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 138
Review article
development of subtypes within these structures. Here, we discuss
progress for both strategies. Several scRNA-seq studies in mice used
embryonic tissues to study early embryonic mDA progenitor cells
15,16,18
.
Sequencing of embryonic day 11 (E11.5) to E18.5 mouse embryos iden-
tified three embryonic mDA neuron types (mDA0, mDA1 and mDA2)
and one mDAneuroblast type (NbDA), originating from oneventricular
zone-progenitor cluster
18
. Embryonic (immature) mDA0 is charac-
terized by Th expression, whereas the other two subtypes express
additional mDA markers, such as Dat (mDA1) or Lmo3 (encoding LIM
domain-only protein 3) and Aldh1a1 (mDA2)
19,39
. Other studies reported
only two mDA neuroblast clusters16. This apparent discrepancy may
be caused by differences in, for example, tissue quality, sequencing
approaches or sequencing depth. Interestingly, some adult mouse
subtypes are already present as early as E13.5 (ref.
15
), and thus far com-
parable results were observed in scRNA-seq of age-matched human
embryos15,18. As outlined earlier for the adult mDA system, more
extensive scRNA-seq analysis will facilitate further cross-species
comparison and the delineation of developmental trajectories.
Morphogens
The identification of distinct (embryonic) subtypes raises the question
how mDA neuron heterogeneity is established during development,
starting at the time of floorplate specification. Generally, medial pro-
genitor cells in the mouse floorplate generate SNc neurons, whereas
cells in the lateral domain of the floorplate give rise to VTA neurons
81,82
.
One factor contributing to this distinction is the timing of mDA neuron
birth
83
. Administration of radioactive[3H]thymidine at several embry-
onic time points (E11–E15) in rodents revealed that more rostral SNc
mDA neurons are born first around E11.5, whereas caudal VTA neurons
start to appear at E12.5 (ref.
83
). This (relative) temporal regulation of
SNc and VTA neuron birth is conserved in rodents and primates84–87.
As the morphogen Sonic Hedgehog (SHH) displays a biphasic pat-
tern of expression in the floorplate during early mDA specification, it
was proposed that SHH signalling may regulate the temporal differ-
ences between lateral and medial progenitors
81,88,89
. Genetic lineage
tracing using Shh–CreERT2 or Gli1–CreERT2 mice confirmed that medial
progenitors start to respond to SHH at E7.5, whereas lateral progenitors
respond 1 day later81,85,90. Similarly, Shh expression is delayed in lateral
progenitors as compared with medial progenitors in mice (Fig.2a).
These findings were corroborated by conditional knockout of SHH
signalling from E8.5 onwards. In these mice, the SNc is less severely
affected than the VTA
91
. Moreover, inactivation of a negative regulator
of SHH signalling, Cdon, results in an increase in the number of VTA
mDA neurons92.
Alterations in Wnt signalling strength contribute to mDA neuron
subtype specification
90,93,94
(Fig.2a). Various Wnt signalling molecules
are present in the caudolateral floorplate
95
. In the lateral floorplate,
R-spondin2 (RSPO2) enhances Wnt signalling
95–97
, which activates lym-
phoid enhancer-binding factor1 (LEF1). LEF1 induces Lmx1a98, but ham-
persLMX1A in activating Pitx3. Low PITX3 levels seem to be necessary
for the VTA lineage early in development. RSPO2 is also involved in the
regulation of these neurons during migration97. Medial floorplate cells
(giving rise to the rostrolateral SNc) also express Dkk3 (ref.
99
). DKK3
reduces LEF1 expression by downregulating Wnt signalling, allowing
Pitx3 expression through LMX1A and pre-B cell leukaemia transcription
factor 1 (PBX1)
95,99
. Interestingly, the rostrolateral SNc is affected when
Lmx1a is ablated97. Altogether, Wnt signalling plays a complex role in
mDA neuron specification and can also contribute to subtype-specific
processes at later developmental stages. This is underscored by the fact
that the Wnt inhibitor Mest regulates the development and survival of
a specific part of the SNc100,101 (Fig.2a).
Transcription factors
Transcription factors add an extra level of complexity to floorplate
specification. First, inactivation of Pitx3 or Mest in mice leads to the
ablation of mutually exclusive SNc sub-subtypes101–103. The subtype
lost in Pitx3 mutants expresses Aldh1a1 (refs.
104–106
) (Fig.2a). The VTA is
spared upon Pitx3 loss, and only lateral mDA neurons (that is, prospec-
tive SNc) require Pitx3 for Th induction103. This highlights a require-
ment for Pitx3 in determining the mDA phenotype in a subset of SNc
Box 1
Parkinson disease and
dopamine replacement therapy
In 1817, James Parkinson described six patients with “involuntary
tremulous motion, with lessened muscular power, in parts not in
action and even when supported; with a propensity to bend the
trunk forwards, and to pass from a walking to a running pace:
the senses and intellects being uninjured”252. Two centuries later
these symptoms are deined as bradykinesia and muscle rigidity,
low-frequency rest tremor and instable posture, and are called
‘parkinsonism’. Parkinson disease (PD) is distinguished from other
parkinsonian disorders (for example, multiple system atrophy or
progressive supranuclear palsy) by post-mortem conirmation
of intracellular α-synuclein (SNCA) aggregates (Lewy bodies),
selective degeneration of midbrain dopamine neurons within
the nigrostriatal pathway and reactive gliosis253,254. Non-motor
symptoms such as hyposmia, depression, rapid eye movement
sleep behaviour disorder and constipation are increasingly
recognized and often precede motor deicits255.
Ageing is the most important risk factor of sporadic PD,
combined with environmental factors, such as pesticides, and
genetic variants. Genetics explain around 10–15% of PD, and even
monogenic causes are described256, in which patients display
mutations in the genes encoding α-synuclein (SNCA)257, leucine-
rich repeat kinase 2 (LRRK2)258 and glucocerebrosidase (GBA1)259.
Probable cellular causes of selective substantia nigra pars
compacta neurodegeneration are described in Box2.
One of the main strategies for treating PD aims to uphold
midbrain dopamine signalling in the striatum, and pharmacological
dopamine therapies are most frequently usedfor this260. l-DOPA,
COMT or MAOB inhibitors are used to increase dopamine availability,
whereas amantadine induces dopamine release in the synaptic cleft.
Also, dopamine receptor agonists can be used. All these therapies
are accompanied by mild to serious side eects (including l-DOPA-
induced dyskinesiasand dopamine agonist-induced nausea), and
at later stages patients show luctuating responses to dopaminergic
medication due to progressive neurondegeneration. Considerable
progress has been made in both (early) detection and recognition
of PD (genetic) subtypes261, as well as in novel pharmacological
and cellular replacement strategies239,262 to combat symptoms and
disease progression in the future.
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 139
Review article
neurons and shows that VTA neurons are less dependent on Pitx3 for
their identity.
In mice, Otx2 is expressed in all mDA progenitors, whereas in the
adult it becomes restricted to Aldh1a1+Calb1+ VTAneurons (probablypart
ofsubtype6) (refs.19,107) (Figs.1 and 2a). However, Otx2 expression
is stronger in lateral floorplate domains than in medial floorplate
domains
82
. Medial progenitors express Sox6, an adult marker for the
SNc and a subtype of dorsolateral VTA neurons (subtype 2)39,82 (Figs. 1a
and 2a). Additionally, a small subset of lateral Otx2+ mDA progeni-
tors expresses Zfp503 (also known as Nolz1, encoding zinc-finger
a
b
SHH signalling Wnt signalling Transcription factors
E7. 5 E8.5 E10.5–E12.5 E11.5
Gli1+
SHH
Gli1+
SHH
Adult Adult Adult
IFP mFP
RSPO2
Bcat
LEF1 LMX1A
↓ PITX3
DKK3
Bcat Mest
PBX1
LEF1 LMX1A
↑ PITX3
?
Unknown?
• BCL-11A
• NEUROD6
• UNCX4.1
• EBF1
• TCF4
• PHOXB2
SOX6 +
• Otx2+
• Nolz1+
• Otx2+
• Nolz1+
Adult
PITX3
Mest
E13.5–E15.5
Radial migration Tangential migration Population-dependent positioning
CXCL12
Radial glia-like cell
NTN1+
dSNc mDA neuron
vSNc mDA neuron
VTA mDA neuron
• CXCR4+
• DCC+
• CXCR4+
• DCC+
SNpl mDA neuron
GABAergic neuron
Meninges
RELN
Meninges
DAB1+
DAB1+NTN1
NTN1+
striatal
fibre
• CXCR4+
• DCC+
• SIX3+
Ventricle
mFP
IFP
SNc
VTA
VTA
SNc
Aqueduct
Aqueduct
Fig. 2 | Molecular mechanisms that regulate thedifferentiation and positioning
of mDA neuron subtypes. a, Temporally controlled Sonic Hedgehog (SHH)
signalling, Wnt signalling and various transcription factors dictate the generation
of distinct midbrain dopamine (mDA) neuron subtypes in the adult mDA system.
At embryonic day 7.5 (E7.5), early SHH signalling to Gli1+ precursors in the medial
floorplate (mFP; red) gives rise to cells that eventually form the substantia nigra
pars compacta (SNc). At E8.5, Gli1 expression (red) extends to the lateral floorplate
(lFP) and SHH signalling then also initiates a ventral tegmental area (VTA) fate.
At E10.5–E12.5, R-spondin2 (RSPO2) in the lFP (orange) enhances Wnt signalling,
resulting in β-catenin (Bcat) expression, whereas DKK3 in the mFP (blue) inhibits
Wnt signalling, resulting in decreased Bcat expression. This variation in Wnt
signalling strength causes differential downstream gene expression, culminating
in different PITX3 levels, which regulate the specification of different mDA neuron
subtypes in the SNc and VTA. From E11.5 onwards, the transcription factors OTX2,
NOLZ1 and SOX6 mark floorplate progenitors for future SNc and VTA regions.
Additional transcription factors (listed on the right side of the panel) have been
implicated in the survival and further specification of mDA neuron subtypes.
However, their temporal expression, downstream molecular targets, anatomical
distribution and precise function and the processes that they mediate are
incompletely understood and require further studies. b, At least three different
mechanisms are involved in the migration and positioning of (subtypes of) mDA
neurons between E13.5 and E15.5. The radial migration of Cxcr4+ mDA neurons is
regulated by attraction to C-X-C motif chemokine 12 (CXCL12) expressed in the
meninges76,85 (left panel). The tangential migration of Dab1+ neurons (SNc and lateral
VTA) is controlled by reelin (RELN) gradients originating in the red nucleus85,136
(middle panel). Netrin1 (NTN1)-mediated migration of a Six3+ GABA population
into the substantia nigra par reticulata repositions SNc mDA neurons dorsally
away from the pial surface130 (right panel). BCL-11A, B cell chronic lymphocytic
leukaemia/lymphoma 11A; dSNc, dorsal substantia nigra pars compacta; EBF1,
early B cell factor 1; LEF1, lymphoid enhancer-binding factor 1; PBX1, pre-B cell
leukaemia transcription factor 1; SNpl, substantia nigra pars lateralis; UNCX4.1,
UNC homeobox 4.1; vSNc, ventral substantia nigra pars compacta.
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 140
Review article
protein 503)82. High levels of OTX2 inhibit the expression of Sox6, restrict-
ing this transcription factor to the medial domain. Loss of Otx2 therefore
causes the Sox6+ domain to be expanded, resulting in the generation of
more rostral SNc neurons. By contrast, loss of Sox6 does not influence
Otx2 expression but increases the number of Calb1+Calb2+ cells (Calb2
encodes calretinin), suggesting that even with limited Otx2 expression
mouse floorplate cells are programmed to become VTA neurons. As Otx2
and Sox6 expression is not limited to mDA precursors, but also occurs in
postmitotic mDA neurons, a role for the transcription factorsencoded
by these genes during mDA specification at later stages is likely. The
observation that not all mDA progenitor cells are completely converted
to a different cell fate upon Sox6 or Otx2 ablation hints at the existence of
additional molecular regulators of mDA neuron subtype specification
108
.
At E14.5, Otx2 levels are lower in rostral regions than in caudal
regions of the Sox6
+
domain
39
. This gradient has been suggested to
cause rostral progenitors to become dSNc and SNpl neurons (sub-
types 2 and 3, both Otx2− in the adult), whereas caudal progenitors
are thought to become Otx2
+
ventromedial VTA neurons (subtype 6
in Fig.1a). Most Sox6+ progenitors contribute to Aldh1a1+ SNc neu-
rons and to a lesser extent to Calb1+ neurons. Sox6+ progenitors also
contribute to Otx2
+
VTA neurons after downregulation of Sox6 (ref.
39
).
Interestingly, the SNc is not completely derived from Sox6
+
progeni-
tors as some neurons upregulate Sox6 only postmitotically (ref.
39
).
This suggests that subtype specification exhibits a certain degree of
developmental flexibility and that additional cues (for example, from
target tissues) influence the final molecular signature of postmitotic
mDA neurons. Also in humans, SOX6 expression is restricted to the
medial floorplate, while OTX2 is more broadly expressed
109
, suggesting
that early specification may follow similar patterns as in mice.
After E15.5, a Th
+
Neurod6
+
subtype is detected in the mouse ven-
tral VTA110. Neurod6, together with Neurod1, regulates the survival of
this subtype. In addition, Bcl11a is present in a region below the pro-
genitor domain and in some differentiated cells across the VTA and
SNc from E12.5 onwards40. This could indicate a role for Bcl11a in the
differentiation of mDA precursors into a Bcl11a+ subtype40. In addition
to the aforementioned transcription factors in this section, additional
genes encodingtranscription factors and other proteins have been
implicated in the early development of mouse mDA neuron subsets: Dcc
(SNc)111, Phox2b (dorsocaudal VTA)112, Tcf4 (VTA and Aldh1a1+ subsets)113,
Ebf1 (which encodes early B cell factor 1; SNc)114 and Uncx4.1 (also known
as Uncx, which encodes UNC homeobox 4.1; Calb1
+
subset and SNpl)
115
.
Although further studies are needed to examine the role of these cues in
more detail, it is becoming evident that the combinatorial actions of differ-
ent intrinsic (and probably extrinsic) molecular regulators is required to
dictate the birth and subsequent differentiation of mDA neuron subtypes.
mDA neuron subtype development in vitro
Interestingly, the in vitro differentiation of DA neurons fromembry-
onic stem cells (ES cells) or induced pluripotent stem cells (iPS cells)
relies on pathways similar to those described earlier herein. In the early
steps of one of the first differentiation protocols, OTX2 is detected in
undifferentiated human embryoid bodies, whereas midbrain markers
such as WNT1, EN1 and NURR1 first appear during early differentiation
in NES
+
neuronal precursors
116
. Treatment with SHH and FGF8 then
induces TH expression and mDA neuron fate in a subset (around 30%)
of cells
116,117
. Other cells display serotonergic or GABAergic fates. Dif-
ferent methods that force expression of LMX1A in cultured neuronal
progenitors can increase the percentage of mDA cells to 75–95%
73,118–120
.
Although differences in the timing and the concentration of SHH
and/or FGF8 have no effect on TH
+
cell number
116
, they seem to influ-
ence subtype specification73. In addition, neuronal precursors can be
directed towards more ventral or caudal fates by tweaking expression
of WNT1 together with SHH121,122.
As ES cell- and iPS cell-generated mDA neurons are valuable tools
for both further understanding and treating diseases such as PD,
it is important to establish exactly which mDA neurons are created
by different differentiation protocols. Analyses of iPS cell-derived
mDA neurons created by (variations of) the Kriks protocol
120
show
the presence of four neuronal clusters (of which one is SNc-like and
another is VTA-like) and two progenitor subsets
123,124
, which somewhat
resemble molecular subtypes described by Saunders et al.
17
and La
Manno et al.
18
, respectively, but not Poulin et al.
19
. Although low levels
of SOX6 can be detected in iPS cell-derived mDA neurons
123
, clear SNc
SOX6
+
ALDH1A1
+
cells appear to be lacking. The original Lee protocol
116
seems to favour VTA-like cells (high levels of OTX2 and CALB1 expres-
sion), but changes in the timing of SHH and FGF8 administration and/or
forced SOX6 expression is able to facilitate the generation of more
vSNc-like mDA neurons (resembling the SOX6
+
ALDH1A1
+
subtype 1)
109
.
mDA neurons can also be produced in vitro by co-culture of ES cells
with a stromal cell line
125,126
, or by treatment with stromal factors
127
.
However, it is unclear which developmental programmes are activated
by these treatments and which subtypes are generated. Although
further work is needed to refine the currently available protocols,
stem cell-based modelling of mDA neurons and their development
holds great promise for drug discovery and therapeutic application.
Further insight into which mDA neuron subtypes exist and how these
develop in vivo will assist the future generation of more accurate mDA
neuron subtypes in vitro.
Migration and positioning of mDA neuron subtypes
Following their generation in the floorplate, differentiating mDA
neurons migrate to their final positions, occupying the SNc and
VTA. Genetic fate mapping studies show that SNc and VTA neurons
are intermingled in medial regions of themarginal zone from E11.5
to E13.5 (ref.
85
). Between E15.5 and E18.5, most SNc neurons start to
move laterally bytangential migration. Interestingly, at more rostral
levels this lateral shift can be detected already at E13.5 (refs.
39,85
). In
addition, it was found that Sox6+ progenitors have distinct migratory
routes depending on their destination (that is, VTA, dSNc or vSNc).
Sox6−Otx2+Nolz1+ progenitors mainly migrate radially, whereas Sox6+
neurons also migrate tangentially39. Some Sox6− progenitors in more
rostral regions also show tangential migration when forming the dSNc
and SNpl.
Radially migrating mDA neurons use radial glia for structural sup-
port and molecular guidance
128,129
.Radial migration of mDA neurons is
in part regulated by the chemokine receptor gene Cxcr4 at E12.5 (ref.
76
)
(Fig.2b). With use of explant cultures and in vivo electroporation
experiments, meningeal expression of Cxcl12 was shown to attract
Cxcr4
+
neurons to the pial surface
76
. Analysis of Cxcr4-knockout or
Cxcl12-knockout mice confirmed impairments in radial migration,
and revealed defects in radial glia processes and aberrantly localized
Th+ neurons76,85.
Theaxon guidance cue (AGC) netrin1 (NTN1) is another regula-
tor of radial migration130. It was shown that the Ntn1–Dcc pathway
coordinates SNc neuron localization131,132. More recent work has
expanded on this observation by showing that NTN1 from radial glia
instructs the migration of dSNc neurons (Sox6
+
Aldh1a1
−
Otx2
−
, prob-
ablysubtype 2)
130
. Loss of NTN1 leads to aberrantly localized dSNc
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 141
Review article
neurons in the red nucleus
130
. This effect is mediated by DCC
132
(Fig.2b).
Because only a subset of dSNc neurons is mislocalized in the absence
of NTN1, it is likely that other guidance cues regulate the migration of
other SNc neuron subtypes57,130. Interestingly, scRNA-seq has revealed
the existence of multiple subtypes ofradial glia-like cells during mouse
and human embryonic development, each expressing different axon
guidance factors
18,57
. As these radial glia-like cells show a distinct locali-
zation in the embryonic vMB
18
, it is tempting to speculate that different
subtypes of mDA neurons use different migratory routes depending
on their interactions with radial glia subtypes.
Tangential migration of SNc neurons starts when these neurons
reach the pial surface and is in part regulated by reelin (RELN) signal-
ling85. RELN is expressed in the red nucleus and diffuses into the SNc
area. A subtype of SNc neurons expressing DAB1, a RELN signalling
molecule, is sensitive to the RELN gradient and orients their processes
laterally85 (Fig.2b). Inhibition of RELN signalling leads to mislocaliza-
tion of SNc neurons positive for Lmo3, Sox6 and Girk2 in the lateral
VTA and retrorubral field, whereas Calb1
+
VTA neurons are unaffec
ted
76,85,133–135
. This defect is caused by a shift in migration speed, specifi-
cally affecting moderate-to-fast tangential migration movements
85,136
.
As RELN influences only a subtype of SNc neurons, additional cues must
exist to direct tangential migration of other SNc neurons.
NTN1 has multiple distinct roles during mDA neuron develop-
ment, and was recently also implicated in the indirect positioning of
SNc neurons130. At E14.5, SNc mDA neurons are located at the pial sur-
face. However, from E16.5 onwards, Six3+ GABAergic neurons migrate
ventrally to the presumptive substantia nigra pars reticulata region
attracted by NTN1, thereby pushing SNc mDA neurons more dorsally
to their final position. This effect of NTN1 is dependent on DSCAM, and
NTN1 is deposited in the vMB by axons of striatal neurons that traverse
this region (Fig.2b). This suggests that correct positioning of SNc mDA
neurons in part depends on the migration of other neuron types into
the vMB. Moreover, this study provides evidence that axon guidance
molecules can be provided to vMB neurons by long-range projections
originating at distant sites in the brain. Another example is RELN, which
could possibly be deposited in the vMB by striatal fibres to affect SNc
neuron migration135.
Conclusively, the molecular pathways underlying mDA neuron
subtype migration are only starting to be elucidated. Our current
knowledge highlights a diverse set of molecular pathways regulating
VTA- or SNc-specific migration events. Advances inlineage tracing
and intersectional genetics approaches are expected to provide
more mechanistic insights into mDA neuron subtype migration and
positioning in the future.
Subtype-speciic wiring and function
Viral tracing and intersectional genetics have identified a multitude
of subset-specific circuits within the nigrostriatal, mesolimbic and
mesocortical pathways
21–24,28,29,32,39,40,137
. Although this extensive body
of experimental work has led to the identification of numerous novel
mDA neuron subtypes, integration of these findings with subtype infor-
mation based on specific functions or molecular markers is rather
incomplete. Therefore, for this Review, we focus on the most impor-
tant circuits and associated behaviours described for the subtypes
defined in Fig.1.
Axonal targets of SNc subtypes
Data showing anatomical and functional subdivisions of the nigrostri-
atal pathway
22,29
and results categorizing SNc neurons into different
molecular subtypes (of which at least two display distinct electrophysiol-
ogical and wiring properties) underscore the heterogenous nature of
the SNc
19,138,139
. Most SNc neurons are Sox6
+
, and projections of Sox6
+
neurons target all rostrocaudal levels of the dorsal striatum. Phasic
mDA signalling in the dorsal striatum has been linked to start–stop
signalling and acceleration during locomotion
140,141
. Sox6
+
Aldh1a1
+
SNc ventral tier neurons (subtype 1, Fig.1) project to the dorsolateral
striatum in a dorsal to ventral, rostral to caudal and lateral to medial
gradient
137,142,143
(Fig.3). Rostral Aldh1a1
+
neurons in the vMB project to
the caudal striatum, whereas caudal Aldh1a1
+
neurons project to the
rostral striatum
143
. Selective ablation of Aldh1a1
+
SNc neurons elicits
a moderate reduction in high-speed walking and absence of motor
improvement in accelerating rotarod tests, but does not affect the
maintenance of acquired motor skills
143
. In addition, signalling from
these neurons in the dorsal striatum was time-locked with accelera-
tions while running
39
. Sox6
+
Aldh1a1
−
SNc neurons (subtype 2, Fig.1) are
defined on the basis of the absence of Aldh1al mRNA, but their func-
tional role remains to be elucidated
19
. These neurons mainly target the
dorsomedial to ventral striatum from rostral to caudal levels
39
(Fig.3).
Sox6−Aldh1a1−Calb1+ SNc dorsal tier neurons (Fig.1a) primarily target
the ventromedial striatum at intermediate and caudal levels, and the
medial striatum at rostral levels137 (Fig.3). Interestingly, these projec-
tions are activated upon reward or reward-predicting cues
34,39,141,144
,
but not during movement39,141.
Most mDA axons in the striatum tail originate from
Vglut2
+
Calb1
+
Sox6
−
neurons in the SNpl
39,137
(subtype 3, Figs.1 and 3)
and are implicated in general salience and novelty signalling
33–35
. In
addition, Vglut2+ mDA neurons of the SNpl contribute to innervation
of the amygdala137, suggesting that they are involved in affective behav-
iours. Together, these observations suggest that within the nigrostriatal
pathway, classically implicated in locomotion, only the ventral tier
Sox6
+
subtype exerts this function, whereas dorsal and lateral tier Sox6
−
neurons have mesolimbic projection targets and functions.
Axonal targets of VTA subtypes
Although the VTA is anatomically divided into subdomains, the rela-
tionship between these subdomains, molecularly defined subtypes and
their projection targets is less well defined5,19 as VTA subtypes are highly
heterogeneous and spatially intermingled19,22. VTA mDA neurons play
a role in incentive-based behaviour, motivation and cognition
5,145,146
,
and they mainly project to the nucleus accumbens (NAc) and olfac-
tory tubercle in the ventral striatum, and the amygdala (mesolimbic
pathway) and the PFC (mesocortical circuit)
147,148
. In addition, mDA VTA
projections have been found in the lateral septum and entorhinal cor
-
tex137. With use of various combinations of viral tracing, optogenetics
and electrophysiology, different VTA subcircuits with different roles
have been identified5,21–27,149,150.
The VTA contains neurons that co-transmit glutamate and GABA,
of which only a small subset is also Th+ (refs.5,151) Vglut2+ neurons are
found in all four molecular VTA subtypes. Vglut2+ neurons project to
the medial shell of the NAc, the olfactory tubercle, the lateral septum,
the entorhinal cortex and the PFC137. VTA Vglut2+Vgat+ neurons (subtype
5, Fig.1) have not been traced directly but may project to the lateral
habenula (Fig.3) and differentially respond to errors in the prediction
or omission of reward152.
Ventromedial Vglut2+Aldh1a1+Otx2+ neurons (subtype 6, Figs.1 and 3)
target the lateral septum, the medial shell of the NAc and specific
olfactory tubercle columns, whereas Sox6+ neurons at the VTA–SNc bor-
der project to the NAc core and NAc lateral shell
23,26,137
. VTA projections
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 142
Review article
1
b
4
LS
BNST
FS
OT
STR
NAcc
SI
OT
NAcsh
STR
2
SI
OT
NAcsh
NAcc
STR
5
IHB
mHB
GPe
GPi
BMA
CeA
BLA
STRt
3
NAcsh
OT
STR
Neurod6+Grp+Neurod6–Grp+Neurod6+Grp–Bcl11a+
Sox6– (dSNc)
4
LS
BNST
FS
OT
STR
a
1
NAcc
SI
OT
NAcsh
STR
2
SI
OT
NAcsh
NAcc
STR
5
IHB
mHB
GPe
GPi
BMA
CeA
STRt
BLA
3
NAcsh
OT
STR
1. Sox6+Aldh1a1+
2. Sox6+Aldh1a1–
3. Vglut2+Aldh1a1–5. Vglut2+Vgat+
6. Vglut2+Aldh1a1+Otx2+
7. Vglut2+Vip+
Cck+
4. Vglut2+Aldh1a1–
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 143
Review article
to the NAc core, but not the medial shell of the NAc, are involved in
acquiring motivational value for a reward
153
, whereas aversion-coding
VTA neurons project to the medial shell of the NAc, but not the NAc
core154. Although VTA-to-NAc-projecting neurons are classically heavily
studied, their heterogeneity has made it difficult to appoint specific
functions to molecular subtypes (for discussion, seeref.155).
Neurons located in the supposed VTA periaqueductal grey and
dorsal raphe nucleus that project to the central nucleus of the amygdala
and the bed nucleus of the stria terminalis have recently been shown
to consist of two distinct subtypes: Dat
+
Th
+
and Dat
+
Vip
+
(of which
most also express Vglut2)
156
. Presumably these subtypes are part of
molecular subtypes 4 and 7, respectively (Figs.1 and 3). Projections
from the dorsal raphe nucleus to the central nucleus of the amygdala
are mainly involved in incentive salience, reward memory, sleep–wake
cycle and opioid addiction
157,158
. Interestingly, projections towards the
bed nucleus of the stria terminalis reduce nociceptive signalling in male
mice, whereas in female mice activation of this pathway produces loco-
motor behaviour
159
, pointing at possible sex-dependent differences in
subtype specificity. By process of elimination, subtype 4 could account
for Vglut2+ projections to the basolateral amygdala (BLA), entorhinal
cortex and PFC, although this has not been confirmed137.
Axonal targets of additional mDA neuron subsets
Additional molecular subsets of mDA neurons that arelikely tofurther
subdivide the proposed subtypes in Fig.1 are often not anatomically
confined to either the SNc or the VTA. For example, a subset of subtype 6
(Fig.1) consists of Cck
+
neurons (Cck encodes cholecystokinin) that are
also found in all defined VTA subtypes, as well as subtypes 1 and 3 in
the SNc and SNpl. Cck+ mDA neurons have been suggested to strongly
contribute to projections to the olfactory tubercle and the medial
shell of the NAc
149
, and may also contribute to projections to the NAc
core, NAc lateral shell, lateral septum, bed nucleus of the stria termi-
nalis and PFC
137
. In addition, the most prominent innervation of the
BLA derives from Cck
+
VTA neurons, but not Vip
+
or Vglut2
+
neurons
137
.
Although dopamine signalling in the BLA was previously linked to fear
processing
160–162
, more recent work has shown that VTA projections
towards the striatum and BLA are connected to fear learning and anxi-
ety
28,163
. So far, ventral and medial VTA connections to the PFC have been
described for Cck+ and Vglut2+ neurons25,137. Interestingly, also for these
connections, dual functions have been described in stress–response
signalling164. Furthermore, VTA projections were found in the motor
and somatosensory cortex165.
The recently described subsets of Bcl11a
+
and Neurod6
+
Grp
+
mDA
neurons have unique projection patterns. Bcl11a+ mDA neurons project
to the olfactory tubercle, ventral striatum tail, lateral septum, caudal
dorsomedial striatum and ventral NAc lateral shell, and mice lacking
Bcl11a show deficits in skilled motor learning
40
. Neurod6
+
Grp
+
and
Neurod6+Grp− VTA neurons project to the NAc and lateral septum21,166;
Neurod6
−
Grp
+
neurons also project to the dorsomedial striatum. Their
roles in behaviour are still unknown.
Thus, whereas molecular mDA VTA subtypes are often scattered
throughout the VTA and SNc subtypes are more anatomically con-
fined, projections from mDA neuron subtypes in both regions have
distinct targets and functions. This suggests that specific molecular
programmes must be in place to establish these complex connectivity
patterns.
Establishing mDA neuron subtype connectivity
In mice, the first mDA neurons extend their axons into the medial fore-
brain bundles (MFBs) as early as E11.5, and simultaneously with mDA
migration167,168. From E14.5 onwards, mDA neurons in the MFB reach
their target areas, thereby initiating formation of the nigrostriatal and
mesolimbic pathways. Chemoattractive and chemorepulsive proteins
expressed en route or by target structures provide spatial and temporal
guidance at these stages168–170.
At early stages, AGCs expressed in the vMB either attract or repel
mDA axons, enforcing their dorsolateral extension171–178 (reviewed
inref.169). The complex interplay between FGF8, semaphorin3F
(SEMA3F), Slit homologue 1 protein (SLIT1) and SLIT3 initially attracts
and then repels mDA axons away from the vMB171,173,175,177, whereas the
combinatorial effect of WNT5A, WNT7B, SHH and NTN1 contrib-
utes to the rostrally oriented growth of mDA axons
172,173,176,179
. In the
Fig. 3 | Subtype-specific wiring of rodent mDA neurons. The wiring of the
molecularly defined mouse midbrain dopamine (mDA) neuron subtypes
presented in Fig.1 has been defined in multiple studies and is integrated in
this figure. Their specific projection patterns are complex and cross multiple
anatomical borders. An overview is presented in five coronal hemisections
along the rostrocaudal axis. Colour gradients represent distinct projections
of molecular mDA neuron subtypes blending and overlapping in their target
areas. a, Axonal targets of the molecular mDA neuron subtypes identified in19.
Subtype 1 (blue) projects to the dorsolateral striatum (STR), throughout the
dorsal to ventral, rostral to caudal and lateral to medial axes39,137,143,251. Rostral
Aldh1a1+ neurons in the ventral midbrain project to the caudal STR, whereas
caudal Aldh1a1+ neurons project to the rostral striatum143. Subtype 2 (yellow)
mainly targets from rostral to lateral the dorsomedial to ventral striatum39.
Subtype 3 (green) sends axons towards the tail of the striatum (STRt) and
the amygdala33,137. Vglut2+ neurons are found in all four ventral tegmental area
(VTA) subtypes, as well as subtype 3 in the substantia nigra pars lateralis(SNpl).
Vglut2+ neurons project to the medial shell of the nucleus accumbens (NAc),
olfactory tubercle (OT), lateral septum (LS) and prefrontal cortex (PFC)137.
Subtype 6 (orange) encompasses multiple subsets. Cck+ neurons (turquoise)
are found in all four VTA subtypes as well as in subtypes 1 and 3 in the substantia
nigra pars compacta(SNc) and SNpl, and project to the NAc, OT, basolateral
amygdala (BLA), LS, bed nucleus of the stria terminalis (BNST) and PFC137.
Vglut2+Aldh1a1+Otx2+ neurons project to the medial shell of the NAc, LS and
specific columns in the OT137. Sometimes projections to the dorsolateral and
dorsomedial striatum are also attributed to Aldh1a1+ VTA neurons39,143, although
this may be due to challenges in disentangling Aldh1a1+ VTA and SNc neurons
after viral labelling. Vglut2+Calb1+ neurons also project to medial shell of the
NAc137. By process of elimination, subtype 4 (red) could account for Vglut2+
projections to BLA, entorhinal cortex (not shown) and PFC, although this has
not been confirmed137. Subtype 5 (purple) projections have not been directly
traced, but on the basis of their GABAergic profile could project to the medial
shell of the NAc, PFC and lateral habenula (lHb)152. Subtype 7 (lilac) primarily
projects to the lateral part of the central nucleus of the amygdala (CeA) and
BNST137,156,157,159. b, Axonal targets of the molecular mDA neuron subtypes
identified in21,39,40. Sox6−Aldh1a1− neurons (turquoise circle) project to the
ventromedial striatum andmight also contribute toprojections to the lateral
shell of the NAc and the NAc core (NAcc)39. Neurod6+Grp+ VTA neurons (orange
circle) and Neurod6+Grp− neurons (turquoise square) project to the NAc and
LS21,166. A subtype of Neurod6−Grp+ neurons (light turquoise circle) also projects
to the dorsomedial striatum. Bcl11a+ mDA neurons (blue square) project to
the OT, ventral STRt, LS, caudal dorsomedial striatum and ventral NAc lateral
shell40. BMA, basomedial amygdala; FS, fundus of striatum; GPe, external globus
pallidus; GPi, internal globus pallidus; mHb, medial habenula; NAcsh, nucleus
accumbens shell; SI, substantia innominata.
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 144
Review article
diencephalon, mDA projections fasciculate and form the MFB. Several
transcription factors, AGCs and receptors cooperate to maintain the
fasciculation and ipsilateral trajectory of the MFB, including NOLZ1,
DCC, homeobox protein Nkx-2.1 (NKX2-1), SLIT1/2 and plexinA3
(PLXNA3)132,180–184. Similarly, the presence of repulsive ephrin A5,
SEMA3A and SEMA3F in the dorsal diencephalon ensures MFB fascicula-
tion and its ventral trajectory177,183,185,186. mDA axons also rely on descend-
ing fibre tracts for their pathfinding, such as Gad65
+
descending fibres
(Gad65 is also known as Gad2) from the mammillotegmental tract or
LAMP1+ fibres in the fasciculus retroflexus187,188. Thus, the growth of
mDA axons from the dorsocaudal mesencephalon to the diencephalon
and rostral target structures requires complex molecular programmes.
Target area innervation
mDA neurons innervate large but specific target areas with relatively
limited collateralization as compared with other neuron types (for
example, serotonergic or noradrenergic neurons), while forming
complex and highly compartmentalized axonal arborizations in the
striatum and PFC
29,189–197
. It has been proposed that mDA axons initially
collateralize non-specifically throughout the striatum starting at E13.5
and branch extensively, potentially due to cues secreted by striatal
cells198. Subsequently, mistargeted collaterals are pruned between E15.5
and postnatal day0 (ref.
199
). Gradients of AGCs have been proposed
to guide axons to their targets200, and examples of guidance cues that
regulate striatum innervation and axonal branching include glial cell
line-derived neurotrophic factor (GDNF)201, ephrin A5 (ref.186), NTN1
(refs.
131,132
), ephrin A1 and ephrin A4 (ref.
202
). NTN1 is particularly inter-
esting as it segregates SNc-derived and VTA-derived axons, which differ
in their responsiveness to NTN1 (ref.131). The mechanisms underlying
this differential NTN1 responsiveness are unknown.
A subset of VTA neurons projects to the PFC. These neurons are pos-
sibly born late compared with other mDA neurons and mainly localize
in the rostromedial VTA, and around half of these neurons co-express
Vglut2 in the rat VTA
137,177,203,204
(Figs.1 and 3). To establish their connec-
tions, these axons follow two distinct routes: (1) through the striatum,
stalling in the NAc and later innervating the PFC or (2) growing ventral
of and around the striatum to the PFC. Th
+
axons are first seen in the PFC
at E15.5, and after a delay of 2 days start to innervate the cortical plate177.
Attraction towards the cortical plate is mediated by SEMA3F detected by
the receptor NRP2 (ref.
177
). Interestingly, whereas a subset of VTA mDA
projections arrive in the PFC during embryonic development, more VTA
neurons start to innervate the PFC at the onset of adolescence
205,206
.
More specifically, Dcc
low
VTA axons reroute themselves from the NAc to
the PFC, whereas Dcc
hi
VTA axons remain in the NAc
205–207
. This example
nicely illustrates how differential AGC expression can aid in the axonal
segregation of neurons within a VTA subtype.
Subtype-specific axon guidance
The expression of AGCs contributes to the ability of different subtypes
of mDA neurons to innervate specific target areas. Most of our current
knowledge concerns the differential guidance of axons derived from the
SNc versus the VTA. For example, Sema7a mRNAand the mRNA forits
receptor Plxnc1 are differentially expressed in the mDA neuron pool and
striatum208. Sema7a mainly marks the SNc, whereas Plxnc1 is confined to
subdomains of the VTA (probably the parabrachial pigmented nucleus
and paranigral nucleus region). Sema7a is primarily expressed in the
dorsal striatum and acts as an axon repellent to direct VTA projections
to the ventral striatum, probably mediated by PLXNC1. OTX2 promotes
Plxnc1 expression and is specifically expressed in a ventromedial VTA
subset (Fig.1, probably subtype 6)209. Therefore, SEMA7A–PLXNC1-
mediated innervation of the ventral striatum may be specific to the
Otx2
+
Aldh1a1
+
mDA neuron subtype 6. LMX1A represses Plxnc1 (ref.
209
),
and a combination of LMX1A and Wnt signalling is required for Pitx3+
SNc development. This suggests that LMX1A may specifically repress
Plxnc1 in the SNc. OTX2 also increases expression of Nrp1/Nrp2, which
is important for several different aspects of VTA wiring57,210. Together
these observations support a model in which interplay between LMX1A
and OTX2 aids in establishing the specific wiring patterns of SNc and
ventromedial VTA subtypes.
The idea that different subtypes of mDA neurons respond differ-
ently to environmental cues to establish precise patterns of axonal
connectivity is supported by the expression of several AGCs and recep-
tors in undefined (that is, with respect to wiring or function) mDA
subsets
173,174,178
. For example, microdissection of the mDA system in
embryonic Pitx3–GFP mice followed by proteomic analysis revealed
subregion-specific expression of guidance cues and receptors130.
Further, several studies showed differential responsiveness of mDA
neurons in distinct parts of the midbrain to AGCs in vitro173,174 ,178.
However, thus far, the specific guidance cues required to establish
axonal connectivity from molecularly (or functionally) defined mDA
neuron subtypes remain unknown.
scRNA-seq can provide specific gene expression information
that will help dissect subtype-specific molecular mechanisms, such
as axon guidance pathways. For example, data published by Tiklová
et al.
15
show that several AGCs are simultaneously expressed in mul-
tiple subtypes (for example, Ntn1, Dcc or Sema3f). This observation
probably represents the fact that distinct mDA neuron subtypes share
common axon guidance mechanisms to mediate generic pathfinding
decisions such as rostrally oriented growth and fasciculation. Never-
theless, AGCs also show subtype specificity. For example, Cntn4 and
Unc5b are specifically expressed in the Vip
+
mDA neuron subcluster.
Moreover, various other AGCs, such as Ntn1 and Flrt2, are expressed in
all subtypes except for Vip
+
neurons
15
. Ntrk1, Sema3d and Sema3e are all
expressed in a Pitx3+Th− subset, and Epha10 (which encodes Ephrin type
A receptor 10), Nrp1 and Wnt3 are enriched in Aldh1a1+ mDA subtypes15.
Intriguingly, loss of Nrp1 causes mild disorganization of mDA axons,
supporting the idea that it may be involved in the wiring of a subset of
axons
178
. Expression of Robo1 (which encodes Roundabout guidance
receptor 1), Plxna4 and Ntf3 (which encodes Neurotrophin3) is likely
to bedetected in Sox6
+
Aldh1a1
−
subtype 2 (ref.
19
). Particularly interest-
ing are Robo1 and Plxna4, as the ligands for the proteins encoded by
these genes (SLIT proteins and semaphorins) are broadly expressed.
Subtype-specific expression of receptors could create an additional
layer of regulation for broadly expressed AGCs. Plxnc1 and Unc5b are
most likelyto be specific for postnatal Sox6
+
Aldh1a1
−
neurons and may
be involved in their target innervation and/or survival
19
. Conclusively,
most AGCs are expressed in specific parts of mDA subtypes, or a limited
number of cells in a subtype, indicating that AGCs can also function
within subsets of the currently defined mDA subtypes.
Our understanding of the molecular mechanisms that help estab-
lish mDA subtype-specific connectivity is rather rudimentary. Analysis
of scRNA-seq and/or snRNA-seq data will help to dissect these mecha-
nisms. However, these techniques are still limited in their ability to
detect all transcripts in a given cell, and transcripts may be translated
locally (that is, in the dendrite or axon instead of the soma). Therefore,
techniques such asRibo-tagging211 and intersectional genetics are
needed to establish how mDA neuron subtypes connect with their
targets during development.
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 145
Review article
Selective vulnerability in disease
In addition to anatomical location, projection area, electrophysiologi-
cal properties and gene expression profiles, mDA neurons can be sub-
divided on the basis of their selective involvement in psychiatric or
neurological disorders. For example, specific projections to the rostral
and dorsal caudate putamen show increased dopamine transmission
in patients with schizophrenia212,213. In patients with substance abuse,
distinct pathological phenotypes are mediated by specific mesolimbic
and/or nigrostriatal pathways
214
. For example, heroine induces specific
activation of a subset of medial VTA neurons projecting to the mouse
medial shell of the NAc during initial reinforcement
215
, whereas human
cocaine addicts show higher neuromelanin MRI signal specifically in
their ventrolateral substantia nigra
216
. The best-characterized disease-
related mDA neuron subtype is neurons in the SNc that degenerate in
PD (Box1) and are particularly vulnerable to cellular stress (Box2).
Vulnerable mDA neuron subtypes in PD
mDA neurons in the ventral tier of the human SNc (Fig.1c) contain less
neuromelanin than their dorsal counterparts and are selectively lost
in PD
217,218
. Not long after this discovery it was shown that the human
SNc ventral tier contains five specific subnuclei, called ‘nigrosomes’
9
.
Nigrosome 1 seems to be affected most, with approximately 98%
cell loss
52
, calling for studies into more specific characterization of
degenerating SNc neuron subtypes. In 2014, it was shown that sub-
types of human vSNc neurons are ALDH1A1+ (Fig.1c), and that there is
a link with ALDH1A1 expression and vulnerability to degeneration219.
In parallel, a cluster of Sox6
+
Aldh1a1
+
vSNc neurons (subtype 1) was
identified in mice that is more vulnerable than other SNc neuronsto
the neurotoxin1-methyl-4-phenyl-1,2,3,5-tetrahydropyridine
20
. Later
it was reported that these Sox6
+
Aldh1a1
+
neurons correspond to the
subsetthat degenerates specifically in patients with PD
39
. A recent
study reported a more specific SOX6
+
AGTR1
+
subtype (AGTR1 encodes
angiotensin II receptor type 1a) in the ALDH1A1+ region that selec-
tively degenerates in patients with PD
46
. In addition, this study also
shows that PD-related genetic factors contribute to degeneration by
intensifying the cell-intrinsic aspects of selective vulnerability that
are described in Box2. However, how gene expression patterns of the
most vulnerable molecular subtype (ALDH1A1 and AGTR1) correspond
to human nigrosome-matrix anatomy is not clear. As nigrosomes can
be recognized using high-resolution MRI technology220,221, further
advances in correlating these structures with molecular markers and
phenotypes could help to improve diagnostics related to disease onset
and progression.
In addition to defining the mechanistic basis of neuronal vulner-
ability, it is valuable to understand why some neurons are more resil-
ient. Within VTA and dSNc subtypes, multiple protective mechanisms
are at play. First, high neuromelanin level-containing cells in the dorsal
tier of the SNc (Fig.1c) are relatively spared in PD
218
. Neuromelanin is a
pigment formed from dopamine quinones or aminochrome in a pro-
cess catalysed by iron
222
. It sequesters dopamine quinones and chelates
iron and therefore initially protects against their toxicity
223
. Second,
calbindin expression in VTA and dSNc subtypes
224–226
is thought to pro-
tect against Ca
2+
-mediated toxic dopamine levels, α-synuclein (SNCA)
aggregation and mitochondrial dysfunction
227,228
. OTX2 expression in
VTA neurons is protective against1-methyl-4-phenylpyridinium neuro-
toxicity by regulating neurotrophic factor expression
210
, whereas VTA
Neurod6 plays a role in mitochondrial maintenance and reactive oxygen
species resistance21,229. Bcl11a expression within some SNc mDA neurons
protects against SNCA toxicity
40
. Finally, mDA neurons that co-release
glutamate (for example, subtypes 3 and 4, Fig.1a) may increase their
Vglut2 expression under toxicconditions
230,231
and are spared from
degeneration in patients with PD38,232. Vglut2 expression is protective
against many forms of induced SNc degeneration by lowering oxidative
stress and preserving mitochondrial health (reviewed in231).
A further understanding of the molecular determinants of the
development of mDA neuron subtypes and of the mechanisms under-
lying their selective degeneration or resilience in PD is key to the
development of novel therapeutic strategies.
Box 2
Mechanisms underlying
neuronal vulnerability
in Parkinson disease
Compared with ventral tegmental area neurons, neuromelanin-
containing substantia nigra pars compacta (SNc) neurons are very
large with extremely branched axonal bushes263. The size of SNc
neuron axonal bushes directly correlates with their vulnerability to
neurotoxins264,265. Combined, the burst spiking iring pattern22and
size of these neurons make it clear that SNc neurons have very
high bioenergetic needs. Their many axon-targeted mitochondria
display a higher basal respiratory state than those in ventral
tegmental area (VTA) neurons and produce high levels of reactive
oxygen species264. Their limited capacity to increase mitochondrial
respiratory activity in the case of stress renders them vulnerable
to stress-induced damage264,266. In addition, age-dependent
increases in oxidative and nitrative stress are selectively high in SNc
neurons267,268, and mitochondrion-protective mechanisms decline
with age269.
Dopamine itself also adds to SNc vulnerability. Within the
acidic milieu of neurotransmitter vesicles, dopamine is protected
from oxidation270. However, when in the cytoplasm, dopamine
and its metabolites form cytotoxic dopamine quinones and
aminochrome. Alone, this build-up is not enough to cause neuron
degeneration271. However, together with high levels of calcium and
iron (inherent to dopamine production223) and also α-synuclein
(SNCA) accumulation, it is thought to lead to a multiple-hit
mechanism that triggers degeneration228,272–275. The iring pattern
and L-type Ca2+ channels characteristic of SNc neurons cause inlux
of high levels of Ca2+ (refs.276,277). Free Ca2+ increases dopamine
production and thus cytoplasmic dopamine levels. Quinones can
form adducts with SNCA aggregates278 that are stabilized by free
Ca2+. Higher cytoplasmic dopamine levels directly correlate with
toxicity and vulnerability to SNCA279,280. Conversely, the presence
of SNCA makes midbrain dopamine neurons more vulnerable to
oxidative stress, probably due to mitochondrial and proteasomal
dysfunction228,272. Moreover, a vicious circle of more oxidative
stress, iron accumulation, SNCA aggregation and degradation
of accumulated neuromelanin is thought to spread pathology to
neighbouring neurons274,275,281. It is noteworthy that degeneration is
thought to start at axonal endings and concludes with neuronal loss
in the SNc282–284.
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 146
Review article
Clinical implications of mDA neuron diversity
The idea that implantation of new mDA neurons as a source of dopa-
mine in the striatum could compensate for the degeneration of the
nigrostriatal pathway arose in the 1970s and 1980s
233–235
. Since then,
many different strategies using a plethora of engrafted material (from
fetal midbrain and stromal cells to ES cell- or iPS cell-derived mate-
rial) have been tested both preclinically and in patients with PD, with
differing success (for reviews, seerefs.236–241). Interestingly, there are
clear indications that the success of this therapeutic strategy strongly
correlates with the neuronal (sub)type used. For example, only fetal
mDA neurons but not olfactory DA neurons grafted into the SNc of a
rat PD model could partially reconstitute the nigrostriatal pathway and
relieve motor symptoms
242
. In addition, although definite subtypes
still need to be confirmed with scRNA-seq, human ES cell-derived mDA
grafts with a caudal ventromesencephalic profile characterized by
markers for the midbrain–hindbrain border were most successful in
reinnervating the nigrostriatal pathway
243
. It has alsobeen shown that
regenerative connections from grafted neurons are subtype specific
244
.
Monosynaptic tracing studies revealed that SNc-like subsets project to
the dSTR, whereas VTA-like cells within the graft send axons to limbic
areas and the PFC245.
In addition to the potential to recreate subtype specificity using
iPS cell-derived mDA cells or tissues
109,124
, it has been confirmed that
monkeys and human patients with PD engrafted with major histo-
compatibility complex-matched or autologous iPS cell-derived
mDA cells suffer from fewer or even no graft-related immune prob-
lems246–249. Therefore, understanding mDA molecular subtype
identity (including the development and wiring diagrams of these
subtypes) in health and disease and establishing the most efficient
iPS cell differentiation protocols to recreate relevant subtypes in vitro
are fundamental steps for improving therapeutic regenerative
strategies.
Glossary
1-Methyl-4-phenyl-1,2,3,5-
tetrahydropyridine
Neurotoxin that upon intracerebral
injection causes rapid degeneration of
the substantia nigra and parkinsonian
symptoms, a method used for model-
ling (late-stage) Parkinson disease in
animal models.
1-Methyl-4-phenylpyridinium
A toxic metabolite of 1-methyl-4-phenyl-
1,2,3,5-tetrahydropyridine.
[3H]Thymidine
Radioactive thymidine analogue that is
taken up when DNA is synthesized, used
as a marker for cell proliferation.
Assembloid
A fused region-speciic organoid used
to model interactions between dierent
tissue types or organs.
Axon guidance
Process during which extrinsic molec-
ules instruct the orientation of axo nal
growth through attraction and/or
repulsion of the axon tip.
Embryonic stem cells
(ES cells). Pluripotent stem cells derived
from the inner cell mass of blastocyst-
stage embryos.
Floorplate
A ventral organizer region along
the midline of the neural tube that
regulates neuronal dierentiation and
positioning.
Genetic fate mapping
Genetic labelling of ancestor cells
and their descendants to map the
anatomical and cellular origin of cells
of interest.
Induced pluripotent stem
cells
(iPS cells). Pluripotent stem cells that are
generated through the reprogramming
of somatic cells by expression of a set
of transcription factors.
Intersectional genetics
Selective targeting of cells by
exploiting the combinatorial expression
of two or more genes to express
genetically encoded recombinases
that results in the activation of proteins
to label or manipulate cells.
Laser capture microdissection
Laser- and microscope-assisted
cutting thatenables precise dissection
of microregions within the tissue of
interest.
Lineage tracing
The identiication of cellular progeny
at subsequent developmental stages
and processes by labelling an ancestor
(progenitor) cell.
Major histocompatibility
complex
Cell surface proteins that present
self-antigens to prevent an autoimmune
response.
Marginal zone
Cell-sparse, outermost zone of the
neural tube or brain containing primarily
axons and glial cells.
Neuroblast
An undierentiated precursor cell
in the central nervous system that
will eventually develop into a fully
dierentiated neural cell.
Organoids
Stem cell-derived and self-assembled
3D cultures that represent key features
of the represented organ.
Radial glia-like cells
Cells that are positive for radial glia
markers in single-cell RNA sequencing
datasets.
Radial migration
Migration of cells along radial glia ibres
away from the ventricular zone.
Ribo-tagging
Tagging of ribosomal subunits to enable
immunopuriication and downstream
processing of ribosomes and attached
mRNAs.
Single-cell RNA sequencing
(scRNA-seq). Dissociation and
isolation of individual cells followed by
sequencing of the RNA transcriptome
per cell.
Single-nucleus RNA
sequencing
(snRNA-seq). Dissociation and isolation
of individual nuclei followed by
sequencing of the RNA transcriptome
per nucleus.
Slide-seq
Processing of tissue sections on an
indexed slide to label RNA transcripts
so as to preserve their spatial origin.
Spatial transcriptomics
Methods to assign cell types (based
on mRNA readouts) to their anatomical
location in tissue sections.
Tangential migration
Migration of cells along the medial–
lateral axis, parallel to the ventricular
surface and orthogonal to radial
glia ibres.
Ventricular zone
A transient layer of tissue lining the
ventricles of the central nervous system
that contains neural stem cells.
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 147
Review article
Conclusions and future perspectives
During the past several decades, the diverse physiological roles of
mDA neurons and their association with a broad spectrum of psychi
-
atric and neurological disorders have prompted intense investigation
into how mDA neurons develop and function. It has become clear that
the functional heterogeneity of mDA neurons is mirrored by a high
level of molecular heterogeneity, as visualized by recent single-cell
profiling studies. Our knowledge of mDA neuron diversity at specific
developmental stages or in the human midbrain is limited, but this void
is likely to be filled soon. This will allow us to predict developmental
trajectories and transcriptional programmes specific to subtypes of
mDA neurons and to assess their evolutionary conservation. An initial
profiling study of human mDA neurons identified a primate-specific
subtype
46
, which is an important observation in the context of human
mDA neuron functions and disease. Single-cell profiling studies using
larger numbers of cells and greater sequencing depths, different tech-
niques (for example, snRNA-seq) and integrated with advancedspatial
transcriptomics will aid in matching mDA neuron subtypes in different
species. This will highlight subtypes that can be studied in model organ-
isms, but probably also those requiring human-specific approaches,
such as iPS cell-derived cultures or organoids. The study of mDA neuron
subtypes requires specific tools that facilitate manipulation dependent
on combinatorial marker expression. Several studies have constructed
and applied intersectional genetics approaches that target mDA neu-
ron subtypes defined by a combination of genes39,130,137,166. However,
especially for developmental studies, further approaches have to be
devised
57
as, for example, viral approaches are less suitable and cur-
rent Th- or Dat-focused strategies do not capture all subtypes or early
developmental events. These technical advances will allow us to address
key unresolved questions, such as the following: When is mDA neuron
diversity established and how do projection targets or extrinsic stimuli
affect molecularly defined subtyping? But they will also allow us to
address questions such as the following: How are subtype-specific syn-
aptic inputs established and do differences exist at the level of dendritic
processes? Initial observations hint at a high level of cellular hetero-
geneity in connecting brain regions such as the developing habenula,
and several populations of GABAergic neurons have been found in the
vMB17,18,41,250. This raises the question of whether the heterogeneity of
mDA neurons is dependent on equally diverse populations of associated
cell types in the vMB (for example, GABAergic neurons or glial cells)
or in brain regions that receive efferent connections or send afferent
connections, especially during development.
The identification of molecularly defined mDA neurons has served
as a great starting point for dissection of the mechanisms underlying
the subtype-specific development of mDA neurons, including their
differentiation, migration and wiring. In parallel, an increasing number
of studies have identified functionally defined subtypes of neurons
that may represent subsets of these molecular subtypes or even define
novel subtypes. Our ultimate goal should be to integrate these different
observations into a single framework. Defining functional subtypes
at the molecular level and, vice versa, establishing connectivity pat-
terns and biological roles of molecular subtypes will accelerate this
integration.
It is evident that subtypes of mDA neurons are differentially
affected in disease. A more detailed understanding of their develop-
ment and function will enable the development of improved in vitro
human disease models and could refine cell replacement strategies,
allowing transplantation of relevant subtypes only. Valuable data
derived from profiling grafts after transplantation could be further
exploited to assess, and subsequently improve, integration of iPS
cell-derived neurons in neural circuits. The application of human
iPS cell-derived neurons, both experimentally and therapeutically, is
advancing at a rapid pace, and the development of advanced organoid
or assembloid models will provide unprecedented opportunities for
studying human mDA neuron development, maturation and circuitry.
Published online: 18 January 2023
References
1. Björklund, A. & Dunnett, S. B. Dopamine neuron systems in the brain: an update.
Trends Neurosci. 30, 194–202 (2007).
2. Hegarty, S. V., Sullivan, A. M. & O’Keee, G. W. Midbrain dopaminergic neurons: a review
of the molecular circuitry that regulates their development. Dev. Biol. 379, 123–138
(2013).
3. Maiti, P., Manna, J., Dunbar, G. L., Maiti, P. & Dunbar, G. L. Current understanding of
the molecular mechanisms in Parkinson’s disease: targets for potential treatments.
Transl. Neurodegener. 6, 1–35 (2017).
4. Kalia, L. V. & Lang, A. E. Parkinson’s disease. Lancet 386, 896–912 (2015).
5. Morales, M. & Margolis, E. B. Ventral tegmental area: cellular heterogeneity, connectivity
and behaviour. Nat. Rev. Neurosci. 18, 73–85 (2017).
6. Meyer-Lindenberg, A. et al. Reduced prefrontal activity predicts exaggerated
striatal dopaminergic function in schizophrenia. Nat. Neurosci. 5, 267–271
(2002).
7. Milton, A. L. & Everitt, B. J. The persistence of maladaptive memory: addiction,
drug memories and anti-relapse treatments. Neurosci. Biobehav. Rev. 36, 1119–1139
(2012).
8. Fu, Y. H. et al. A cytoarchitectonic and chemoarchitectonic analysis of the dopamine cell
groups in the substantia nigra, ventral tegmental area, and retrorubral ield in the mouse.
Brain Struct. Funct. 217, 591–612 (2012).
9. Damier, P., Hirsch, E. C., Agid, Y. & Graybiel, A. M. The substantia nigra of the human
brain: I. Nigrosomes and the nigral matrix, a compartmental organization based on
calbindin D28K immunohistochemistry. Brain 122, 1421–1436 (1999).
10. Grimm, J., Mueller, A., Hefti, F. & Rosenthal, A. Molecular basis for catecholaminergic
neuron diversity. Proc. Natl Acad. Sci. USA 101, 13891–13896 (2004).
11. Greene, J. G., Dingledine, R. & Greenamyre, J. T. Gene expression proiling of rat midbrain
dopamine neurons: implications for selective vulnerability in parkinsonism. Neurobiol.
Dis. 18, 19–31 (2005).
12. Chung, C. Y. et al. Cell type-speciic gene expression of midbrain dopaminergic neurons
reveals molecules involved in their vulnerability and protection. Hum. Mol. Genet. 14,
1709–1725 (2005).
13. Brochier, C. et al. Quantitative gene expression proiling of mouse brain regions
reveals dierential transcripts conserved in human and aected in disease models.
Physiol. Genomics 33, 170–179 (2008).
14. Brichta, L. et al. Identiication of neurodegenerative factors using translatome–regulatory
network analysis. Nat. Neurosci. 18, 1325–1333 (2015).
15. Tiklová, K. et al. Single-cell RNA sequencing reveals midbrain dopamine neuron
diversity emerging during mouse brain development. Nat. Commun. 10, 1–12
(2019).
16. Hook, P. W. et al. Single-cell RNA-Seq of mouse dopaminergic neurons informs
candidate gene selection for sporadic parkinson disease. Am. J. Hum. Genet. 102,
427–446 (2018).
17. Saunders, A. et al. Molecular diversity and specializations among the cells of the adult
mouse brain. Cell 174, 1015–1030.e16 (2018).
18. La Manno, G. et al. Molecular diversity of midbrain development in mouse, human, and
stem cells. Cell 167, 566–580.e19 (2016).
19. Poulin, J.-F., Gaertner, Z., Moreno-Ramos, O. A. & Awatramani, R. Classiication of
midbrain dopamine neurons using single-cell gene expression proiling approaches.
Trends Neurosci. 43, 155–169 (2020).
20. Poulin, J. F. et al. Deining midbrain dopaminergic neuron diversity by single-cell gene
expression proiling. Cell Rep. 9, 930–943 (2014).
21. Kramer, D. J., Risso, D., Kosillo, P., Ngai, J. & Bateup, H. S. Combinatorial expression of Grp
and Neurod6 deines dopamine neuron populations with distinct projection patterns and
disease vulnerability. eNeuro 5, ENEURO.0152-18.2018 (2018).
22. Farassat, N. et al. In vivo functional diversity of midbrain dopamine neurons within
identiied axonal projections. Elife 8, e48408 (2019).
23. Beier, K. T. et al. Circuit architecture of VTA dopamine neurons revealed by systematic
input-output mapping. Cell 162, 622–634 (2015).
24. Beier, K. T. et al. Topological organization of ventral tegmental area connectivity
revealed by viral-genetic dissection of input-output relations. Cell Rep. 26, 159–167
(2019).
25. Lammel, S., Ion, D. I., Roeper, J. & Malenka, R. C. Projection-speciic modulation of
dopamine neuron synapses by aversive and rewarding stimuli. Neuron 70, 855–862
(2011).
26. Lammel, S. et al. Unique properties of mesoprefrontal neurons within a dual
mesocorticolimbic dopamine system. Neuron 57, 760–773 (2008).
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 148
Review article
27. de Jong, J. W. et al. A neural circuit mechanism for encoding aversive stimuli in the
mesolimbic dopamine system. Neuron 101, 133–151.e7 (2019).
28. Tang, W., Kochubey, O., Kintscher, M. & Schneggenburger, R. A VTA to basal amygdala
dopamine projection contributes to signal salient somatosensory events during fear
learning. J. Neurosci. 40, 3969–3980 (2020).
29. Lerner, T. N. et al. Intact-brain analyses reveal distinct information carried by SNc
dopamine subcircuits. Cell 162, 635–647 (2015).
30. Bimpisidis, Z. et al. The NeuroD6 subtype of VTA neurons contributes to psychostimulant
sensitization and behavioral reinforcement. eNeuro 6, ENEURO.0066-19.2019 (2019).
31. Viereckel, T. et al. Midbrain gene screening identiies a new mesoaccumbal glutamatergic
pathway and a marker for dopamine cells neuroprotected in Parkinson’s disease. Sci. Rep.
6, 35203 (2016).
32. Watabe-Uchida, M., Zhu, L., Ogawa, S. K., Vamanrao, A. & Uchida, N. Whole-brain
mapping of direct inputs to midbrain dopamine neurons. Neuron 74, 858–873 (2012).
33. Menegas, W. et al. Dopamine neurons projecting to the posterior striatum form an
anatomically distinct subclass. Elife 4, e10032 (2015).
34. Menegas, W., Babayan, B. M., Uchida, N. & Watabe-Uchida, M. Opposite initialization
to novel cues in dopamine signaling in ventral and posterior striatum in mice. Elife 6,
e21886 (2017).
35. Menegas, W., Akiti, K., Amo, R., Uchida, N. & Watabe-Uchida, M. Dopamine neurons
projecting to the posterior striatum reinforce avoidance of threatening stimuli.
Nat. Neurosci. 21, 1421–1430 (2018).
36. Engelhard, B. et al. Specialized coding of sensory, motor and cognitive variables in VTA
dopamine neurons. Nature 570, 509–513 (2019).
37. Roeper, J. Dissecting the diversity of midbrain dopamine neurons. Trends Neurosci. 36,
336–342 (2013).
38. Steinkellner, T. et al. Role for VGLUT2 in selective vulnerability of midbrain dopamine
neurons. J. Clin. Invest. 128, 774–788 (2018).
39. Pereira Luppi, M. et al. Sox6 expression distinguishes dorsally and ventrally biased
dopamine neurons in the substantia nigra with distinctive properties and embryonic
origins. Cell Rep. 37, 109975 (2021).
40. Tolve, M. et al. The transcription factor BCL11A deines distinct subsets of midbrain
dopaminergic neurons. Cell Rep. 36, 109697 (2021).
41. Phillips, R. A. et al. An atlas of transcriptionally deined cell populations in the rat ventral
tegmental area. Cell Rep. 39, 110616 (2022).
42. Aguila, J. et al. Spatial RNA sequencing identiies robust markers of vulnerable and
resistant human midbrain dopamine neurons and their expression in Parkinson’s disease.
Front. Mol. Neurosci. 14, 699562 (2021).
43. Monzón-Sandoval, J. et al. Human-speciic transcriptome of ventral and dorsal midbrain
dopamine neurons. Ann. Neurol. 87, 853–868 (2020).
44. Cantuti-Castelvetri, I. et al. Eects of gender on nigral gene expression and Parkinson
disease. Neurobiol. Dis. 26, 606–614 (2007).
45. Zheng, B. et al. PGC-1α, a potential therapeutic target for early intervention in Parkinson’s
disease. Sci. Transl. Med. 2, 52ra73 (2010).
46. Kamath, T. et al. Single-cell genomic proiling of human dopamine neurons identiies
a population that selectively degenerates in Parkinson’s disease. Nat. Neurosci. 25,
588–595 (2022).
47. Smajic, S. et al. Single-cell sequencing of human midbrain reveals glial activation and a
Parkinson-speciic neuronal state. Brain 145, 964–978 (2022).
48. Reyes, S. et al. GIRK2 expression in dopamine neurons of the substantia nigra and ventral
tegmental area. J. Comp. Neurol. 520, 2591–2607 (2012).
49. Reyes, S. et al. Trophic factors dierentiate dopamine neurons vulnerable to Parkinson’s
disease. Neurobiol. Aging 34, 873–886 (2013).
50. Reyes, S., Cottam, V., Kirik, D., Double, K. L. & Halliday, G. M. Variability in neuronal
expression of dopamine receptors and transporters in the substantia nigra. Mov. Disord.
28, 1351–1359 (2013).
51. Afonso-Oramas, D. et al. Dopamine transporter glycosylation correlates with the
vulnerability of midbrain dopaminergic cells in Parkinson’s disease. Neurobiol. Dis. 36,
494–508 (2009).
52. Damier, P., Hirsch, E. C., Agid, Y. & Graybiel, A. M. The substantia nigra of the human
brain: II. Patterns of loss of dopamine-containing neurons in Parkinson’s disease. Brain
122, 1437–1448 (1999).
53. Agarwal, D. et al. A single-cell atlas of the human substantia nigra reveals cell-speciic
pathways associated with neurological disorders. Nat. Commun. 11, 1–11 (2020).
54. Rodriques, S. G. et al. Slide-seq: a scalable technology for measuring genome-wide
expression at high spatial resolution. Science 363, 1463–1467 (2019).
55. Arenas, E., Denham, M. & Villaescusa, J. C. How to make a midbrain dopaminergic
neuron. Development 142, 1918–1936 (2015).
56. Blaess, S. & Ang, S. L. Genetic control of midbrain dopaminergic neuron development.
Wiley Interdiscip. Rev. Dev. Biol. 4, 113–134 (2015).
57. Brignani, S. & Pasterkamp, R. J. Neuronal subset-speciic migration and axonal wiring
mechanisms in the developing midbrain dopamine system. Front. Neuroanat. 11, 55 (2017).
58. Rhinn, M. et al. Sequential roles for Otx2 in visceral endoderm and neuroectoderm
for forebrain and midbrain induction and speciication. Development 125, 845–856 (1998).
59. Millet, S. et al. A role for Gbx2 in repression of Otx2 and positioning the mid/hindbrain
organizer. Nature 401, 161–164 (1999).
60. Wassarman, K. M. et al. Speciication of the anterior hindbrain and establishment of a
normal mid/hindbrain organizer is dependent on Gbx2 gene function. Development 124,
2923–2934 (1997).
61. Joyner, A. L., Liu, A. & Millet, S. Otx2, Gbx2 and Fgf8 interact to position and maintain
a mid-hindbrain organizer. Curr. Opin. Cell Biol. 12, 736–741 (2000).
62. Ono, Y. et al. Dierences in neurogenic potential in loor plate cells along an
anteroposterior location: midbrain dopaminergic neurons originate from mesencephalic
loor plate cells. Development 134, 3213–3225 (2007).
63. Wilkinson, D. G., Bailes, J. A. & McMahon, A. P. Expression of the proto-oncogene int-1
is restricted to speciic neural cells in the developing mouse embryo. Cell 50, 79–88
(1987).
64. Rhinn, M., Dierich, A., Meur, Mle & Ang, S. Cell autonomous and non-cell autonomous
functions of Otx2 in patterning the rostral brain. Development 126, 4295–4304 (1999).
65. Brodski, C. et al. Location and size of dopaminergic and serotonergic cell populations
are controlled by the position of the midbrain-hindbrain organizer. J. Neurosci. 23,
4199–4207 (2003).
66. Basson, M. A. et al. Speciic regions within the embryonic midbrain and cerebellum
require dierent levels of FGF signaling during development. Development 135, 889–898
(2008).
67. Sasaki, H., Hui, C. C., Nakafuku, M. & Kondoh, H. A binding site for Gli proteins is essential
for HNF-3beta loor plate enhancer activity in transgenics and can respond to Shh in
vitro. Development 124, 1313–1322 (1997).
68. Roelink, H. et al. Floor plate and motor neuron induction by dierent concentrations
of the amino-terminal cleavage product of sonic hedgehog autoproteolysis. Cell 81,
445–455 (1995).
69. Lin, W. et al. Foxa1 and Foxa2 function both upstream of and cooperatively with Lmx1a
and Lmx1b in a feedforward loop promoting mesodiencephalic dopaminergic neuron
development. Dev. Biol. 333, 386–396 (2009).
70. Ang, S. L. et al. The formation and maintenance of the deinitive endoderm lineage
in the mouse: involvement of HNF3/forkhead proteins. Development 119, 1301–1315
(1993).
71. Omodei, D. et al. Anterior-posterior graded response to Otx2 controls proliferation and
dierentiation of dopaminergic progenitors in the ventral mesencephalon. Development
135, 3459–3470 (2008).
72. Deng, Q. et al. Speciic and integrated roles of Lmx1a, Lmx1b and Phox2a in ventral
midbrain development. Development 138, 3399–3408 (2011).
73. Andersson, E. et al. Identiication of intrinsic determinants of midbrain dopamine
neurons. Cell 124, 393–405 (2006).
74. Kele, J. et al. Neurogenin 2 is required for the development of ventral midbrain
dopaminergic neurons. Development 133, 495–505 (2006).
75. Kawano, H., Ohyama, K., Kawamura, K. & Nagatsu, I. Migration of dopaminergic
neurons in the embryonic mesencephalon of mice. Brain Res. Dev. Brain Res. 86, 101–113
(1995).
76. Yang, S. et al. Cxcl12/Cxcr4 signaling controls the migration and process orientation of
A9-A10 dopaminergic neurons. Development 140, 4554–4564 (2013).
77. Sacchetti, P., Mitchell, T. R., Granneman, J. G. & Bannon, M. J. Nurr1 enhances
transcription of the human dopamine transporter gene through a novel mechanism.
J. Neurochem. 76, 1565–1572 (2001).
78. Chung, S. et al. Wnt1-lmx1a forms a novel autoregulatory loop and controls midbrain
dopaminergic dierentiation synergistically with the SHH-FoxA2 pathway. Cell Stem Cell
5, 646–658 (2009).
79. Prakash, N. et al. A Wnt1-regulated genetic network controls the identity and fate of
midbrain-dopaminergic progenitors in vivo. Development 133, 89–98 (2006).
80. Ferri, A. L. M. et al. Foxa1 and Foxa2 regulate multiple phases of midbrain dopaminergic
neuron development in a dosage-dependent manner. Development 134, 2761–2769
(2007).
81. Blaess, S. et al. Temporal-spatial changes in Sonic Hedgehog expression and signaling
reveal dierent potentials of ventral mesencephalic progenitors to populate distinct
ventral midbrain nuclei. Neural Dev. 6, 29 (2011).
82. Panman, L. et al. Sox6 and Otx2 control the speciication of substantia nigra and ventral
tegmental area dopamine neurons. Cell Rep. 8, 1018–1025 (2014).
83. Bayer, S. A., Wills, K. V., Triarhou, L. C. & Ghetti, B. Time of neuron origin and gradients
of neurogenesis in midbrain dopaminergic neurons in the mouse. Exp. Brain Res. 105,
191–199 (1995).
84. Bye, C. R., Thompson, L. H. & Parish, C. L. Birth dating of midbrain dopamine neurons
identiies A9 enriched tissue for transplantation into parkinsonian mice. Exp. Neurol. 236,
58–68 (2012).
85. Bodea, G. O. et al. Reelin and CXCL12 regulate distinct migratory behaviors during the
development of the dopaminergic system. Development 141, 661–673 (2014).
86. Levitt, P. & Rakic, P. The time of genesis, embryonic origin and dierentiation of the brain
stem monoamine neurons in the rhesus monkey. Brain Res. 256, 35–57 (1982).
87. Altman, J. & Bayer, S. A. Development of the brain stem in the rat. V. Thymidine-
radiographic study of the time of origin of neurons in the midbrain tegmentum.
J. Comp. Neurol. 198, 677–716 (1981).
88. Ribes, V. et al. Distinct Sonic Hedgehog signaling dynamics specify loor plate and
ventral neuronal progenitors in the vertebrate neural tube. Genes Dev. 24, 1186–1200
(2010).
89. Mavromatakis, Y. E. et al. Foxa1 and Foxa2 positively and negatively regulate Shh
signalling to specify ventral midbrain progenitor identity. Mech. Dev. 128, 90–103 (2011).
90. Hayes, L., Zhang, Z., Albert, P., Zervas, M. & Ahn, S. The timing of Sonic hedgehog and
Gli1 expression segregates midbrain dopamine neurons. J. Comp. Neurol. 519, 3001
(2011).
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 149
Review article
91. Kabanova, A. et al. Function and developmental origin of a mesocortical inhibitory
circuit. Nat. Neurosci. 18, 872–882 (2015).
92. Verwey, M. et al. Mesocortical dopamine phenotypes in mice lacking the Sonic
Hedgehog receptor Cdon. eNeuro 3, ENEURO.0009-16.2016 (2016).
93. Joksimovic, M. et al. Spatiotemporally separable Shh domains in the midbrain
deine distinct dopaminergic progenitor pools. Proc. Natl Acad. Sci. USA 106, 19185
(2009).
94. Brown, A., Machan, J. T., Hayes, L. & Zervas, M. Molecular organization and timing
of Wnt1 expression deine cohorts of midbrain dopamine neuron progenitors in vivo.
J. Comp. Neurol. 519, 2978 (2011).
95. Nouri, P. et al. Dose-dependent and subset-speciic regulation of midbrain dopaminergic
neuron dierentiation by LEF1-mediated WNT1/b-catenin signaling. Front. Cell Dev. Biol.
8, 587778 (2020).
96. Gyllborg, D. et al. The matricellular protein R-spondin 2 promotes midbrain
dopaminergic neurogenesis and dierentiation. Stem Cell Rep. 11, 651 (2018).
97. Hoekstra, E. J. et al. Lmx1a encodes a rostral set of mesodiencephalic dopaminergic
neurons marked by the Wnt/B-catenin signaling activator R-spondin 2. PLoS ONE 8,
e74049 (2013).
98. Zhang, J. et al. A WNT1-regulated developmental gene cascade prevents dopaminergic
neurodegeneration in adult En1+/- mice. Neurobiol. Dis. 82, 32–45 (2015).
99. Fukusumi, Y. et al. Dickkopf 3 promotes the dierentiation of a rostrolateral midbrain
dopaminergic neuronal subset in vivo and from pluripotent stem cells in vitro in the
mouse. J. Neurosci. 35, 13385 (2015).
100. Jung, H., Lee, S. K. & Jho, E. H. Mest/Peg1 inhibits Wnt signalling through regulation
of LRP6 glycosylation. Biochem. J. 436, 263–269 (2011).
101. Mesman, S., van Hooft, J. A. & Smidt, M. P. Mest/Peg1 is essential for the development
and maintenance of a SNc neuronal subset. Front. Mol. Neurosci. 9, 166 (2017).
102. Smidt, M. P. et al. Early developmental failure of substantia nigra dopamine neurons
in mice lacking the homeodomain gene Pitx3. Development 131, 1145–1155 (2004).
103. Maxwell, S. L., Ho, H. Y., Kuehner, E., Zhao, S. & Li, M. Pitx3 regulates tyrosine hydroxylase
expression in the substantia nigra and identiies a subgroup of mesencephalic
dopaminergic progenitor neurons during mouse development. Dev. Biol. 282, 467–479
(2005).
104. Jacobs, F. M. J. et al. Retinoic acid counteracts developmental defects in the substantia
nigra caused by Pitx3 deiciency. Development 134, 2673–2684 (2007).
105. Jacobs, F. M. J. et al. Retinoic acid-dependent and -independent gene-regulatory
pathways of Pitx3 in meso-diencephalic dopaminergic neurons. Development 138,
5213–5222 (2011).
106. Veenvliet, J. V. et al. Speciication of dopaminergic subsets involves interplay of En1 and
Pitx3. Development 140, 3373–3384 (2013).
107. di Giovannantonio, L. G. et al. Otx2 selectively controls the neurogenesis of speciic
neuronal subtypes of the ventral tegmental area and compensates En1-dependent
neuronal loss and MPTP vulnerability. Dev. Biol. 373, 176–183 (2013).
108. di Salvio, M. et al. Otx2 controls neuron subtype identity in ventral tegmental area and
antagonizes vulnerability to MPTP. Nat. Neurosci. 13, 1481–1489 (2010).
109. Oosterveen, T. et al. Pluripotent stem cell derived dopaminergic subpopulations model
the selective neuron degeneration in Parkinson’s disease. Stem Cell Rep. 16, 2718–2735
(2021).
110. Khan, S. et al. Survival of a novel subset of midbrain dopaminergic neurons
projecting to the lateral septum is dependent on NeuroD proteins. J. Neurosci. 37,
2305 (2017).
111. Lo, P. S., Rymar, V. V., Kennedy, T. E. & Sadikot, A. F. The netrin-1 receptor DCC promotes
the survival of a subpopulation of midbrain dopaminergic neurons: relevance for ageing
and Parkinson’s disease. J. Neurochem. 161, 254–265 (2022).
112. Hoekstra, E. J., von Oerthel, L., van der Linden, A. J. A. & Smidt, M. P. Phox2b inluences
the development of a caudal dopaminergic subset. PLoS ONE 7, e52118 (2012).
113. Mesman, S., Wever, I. & Smidt, M. P. Tcf4 is involved in subset speciication of
mesodiencephalic dopaminergic neurons. Biomedicines 9, 317 (2021).
114. Yin, M. et al. Ventral mesencephalon-enriched genes that regulate the development of
dopaminergic neurons in vivo. J. Neurosci. 29, 5170–5182 (2009).
115. Rabe, T. I. et al. The transcription factor Uncx4.1 acts in a short window of midbrain
dopaminergic neuron dierentiation. Neural Dev. 7, 1–16 (2012).
116. Lee, S., Lumelsky, N., Studer, L., Auerbach, J. M. & McKay, R. D. Eicient generation of
midbrain and hindbrain neurons from mouse embryonic stem cells. Nat. Biotechnol. 18,
675–679 (2000).
117. Ye, W., Shimamura, K., Rubenstein, J. L. R., Hynes, M. A. & Rosenthal, A. FGF and Shh
signals control dopaminergic and serotonergic cell fate in the anterior neural plate.
Cell 93, 755–766 (1998).
118. Friling, S. et al. Eicient production of mesencephalic dopamine neurons by Lmxla
expression in embryonic stem cells. Proc. Natl Acad. Sci. USA 106, 7613–7618 (2009).
119. Panman, L. et al. Transcription factor-induced lineage selection of stem-cell-derived
neural progenitor cells. Cell Stem Cell 8, 663–675 (2011).
120. Kriks, S. et al. Dopamine neurons derived from human ES cells eiciently engraft in
animal models of Parkinson’s disease. Nature 480, 547–551 (2011).
121. Kirkeby, A. et al. Generation of regionally speciied neural progenitors and functional
neurons from human embryonic stem cells under deined conditions. Cell Rep. 1,
703–714 (2012).
122. Kim, T. W. et al. Biphasic activation of WNT signaling facilitates the derivation of midbrain
dopamine neurons from hESCs for translational use. Cell Stem Cell 28, 343–355.e5 (2021).
123. Sandor, C. et al. Transcriptomic proiling of puriied patient-derived dopamine
neurons identiies convergent perturbations and therapeutics for Parkinson’s disease.
Hum. Mol. Genet. 26, 552–566 (2017).
124. Fernandes, H. J. R. et al. Single-cell transcriptomics of Parkinson’s disease human
in vitro models reveals dopamine neuron-speciic stress responses. Cell Rep. 33, 108263
(2020).
125. Kawasaki, H. et al. Induction of midbrain dopaminergic neurons from ES cells by stromal
cell-derived inducing activity. Neuron 28, 31–40 (2000).
126. Vazin, T., Chen, J., Lee, C.-T., Amable, R. & Freed, W. J. Assessment of stromal-derived
inducing activity in the generation of dopaminergic neurons from human embryonic
stem cell. Stem Cells 26, 1517–1525 (2008).
127. Vazin, T. et al. A novel combination of factors, termed SPIE, which promotes
dopaminergic neuron dierentiation from human embryonic stem cells. PLoS ONE 4,
e6606 (2009).
128. Shults, C. W., Hashimoto, R., Brady, R. M. & Gage, F. H. Dopaminergic cells align along
radial glia in the developing mesencephalon of the rat. Neuroscience 38, 427–436
(1990).
129. Marín, O., Valiente, M., Ge, X. & Tsai, L. H. Guiding neuronal cell migrations. Cold Spring
Harb. Perspect. Biol. 2, a001834 (2010).
130. Brignani, S. et al. Remotely produced and axon-derived netrin-1 instructs GABAergic
neuron migration and dopaminergic substantia nigra development. Neuron 107,
684–702.e9 (2020).
131. Li, J. et al. Evidence for topographic guidance of dopaminergic axons by dierential
Netrin-1 expression in the striatum. Mol. Cell Neurosci. 61, 85–96 (2014).
132. Xu, B. et al. Critical roles for the netrin receptor deleted in colorectal cancer in
dopaminergic neuronal precursor migration, axon guidance, and axon arborization.
Neuroscience 169, 932–949 (2010).
133. Nishikawa, S., Goto, S., Yamada, K., Hamasaki, T. & Ushio, Y. Lack of Reelin causes
malpositioning of nigral dopaminergic neurons: evidence from comparison of normal
and Relnrl mutant mice. J. Comp. Neurol. 461, 166–173 (2003).
134. Kang, W.-Y. et al. Migratory defect of mesencephalic dopaminergic neurons in
developing reeler mice. Anat. Cell Biol. 43, 241 (2010).
135. Sharaf, A., Bock, H. H., Spittau, B., Bouché, E. & Krieglstein, K. ApoER2 and VLDLr are
required for mediating reelin signalling pathway for normal migration and positioning
of mesencephalic dopaminergic neurons. PLoS ONE 8, 71091 (2013).
136. Vaswani, A. R. et al. Correct setup of the substantia nigra requires Reelin-mediated fast,
laterally-directed migration of dopaminergic neurons. Elife 8, e41623 (2019).
137. Poulin, J. F. et al. Mapping projections of molecularly deined dopamine neuron
subtypes using intersectional genetic approaches. Nat. Neurosci. 21, 1260–1271
(2018).
138. Evans, R. C., Zhu, M. & Khaliq, Z. M. Dopamine inhibition dierentially controls excitability
of substantia nigra dopamine neuron subpopulations through T-type calcium channels.
J. Neurosci. 37, 3704–3720 (2017).
139. Carmichael, K. et al. Function and regulation of ALDH1A1-positive nigrostriatal
dopaminergic neurons in motor control and Parkinson’s disease. Front. Neural Circuits
15, 644776 (2021).
140. Jin, X. & Costa, R. M. Start/stop signals emerge in nigrostriatal circuits during sequence
learning. Nature 466, 457–462 (2010).
141. Howe, M. W. & Dombeck, D. A. Rapid signalling in distinct dopaminergic axons during
locomotion and reward. Nature 535, 505–510 (2016).
142. Sgobio, C. et al. Aldehyde dehydrogenase 1–positive nigrostriatal dopaminergic ibers
exhibit distinct projection pattern and dopamine release dynamics at mouse dorsal
striatum. Sci. Rep. 7, 5283 (2017).
143. Wu, J. et al. Distinct connectivity and functionality of aldehyde dehydrogenase
1a1-positive nigrostriatal dopaminergic neurons in motor learning. Cell Rep. 28,
1167–1181.e7 (2019).
144. Matsumoto, M. & Hikosaka, O. Two types of dopamine neuron distinctly convey positive
and negative motivational signals. Nature 459, 837–841 (2009).
145. Hauser, T. U., Eldar, E. & Dolan, R. J. Separate mesocortical and mesolimbic pathways
encode eort and reward learning signals. Proc. Natl Acad. Sci. USA 114, E7395–E7404
(2017).
146. Halbout, B. et al. Mesolimbic dopamine projections mediate cue-motivated reward
seeking but not reward retrieval in rats. Elife 8, e43551 (2019).
147. Ioanas, H.-I., Saab, B. J. & Rudin, M. Whole-brain opto-fMRI map of mouse VTA
dopaminergic activation relects structural projections with small but signiicant
deviations. Transl. Psychiatry 12, 60 (2022).
148. Ikemoto, S. Dopamine reward circuitry: two projection systems from the ventral
midbrain to the nucleus accumbens–olfactory tubercle complex. Brain Res. Rev. 56,
27–78 (2007).
149. Heymann, G. et al. Synergy of distinct dopamine projection populations in behavioral
reinforcement. Neuron 105, 909–920.e5 (2020).
150. Lammel, S. et al. Input-speciic control of reward and aversion in the ventral tegmental
area. Nature 491, 212–217 (2012).
151. Miranda-Barrientos, J. et al. Ventral tegmental area GABA, glutamate, and glutamate-
GABA neurons are heterogeneous in their electrophysiological and pharmacological
properties. Eur. J. Neurosci. 54, 4061–4084 (2021).
152. Root, D. H. et al. Distinct signaling by ventral tegmental area glutamate, GABA, and
combinatorial glutamate-GABA neurons in motivated behavior. Cell Rep. 32, 108094
(2020).
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 150
Review article
153. Saunders, B. T., Richard, J. M., Margolis, E. B. & Janak, P. H. Dopamine neurons create
Pavlovian conditioned stimuli with circuit-deined motivational properties. Nat. Neurosci.
21, 1072–1083 (2018).
154. Brischoux, F., Chakraborty, S., Brierley, D. I. & Ungless, M. A. Phasic excitation of
dopamine neurons in ventral VTA by noxious stimuli. Proc. Natl Acad. Sci. USA 106,
4894–4899 (2009).
155. de Jong, J. W., Fraser, K. M. & Lammel, S. Mesoaccumbal dopamine heterogeneity: what
do dopamine iring and release have to do with it? Annu. Rev. Neurosci. 45, 109–129
(2022).
156. Zhao, Q. et al. Histochemical characterization of the dorsal raphe-periaqueductal
grey dopamine transporter neurons projecting to the extended amygdala. eNeuro
https://doi.org/10.1523/ENEURO.0121-22.2022 (2022).
157. Lin, R. et al. The raphe dopamine system controls the expression of incentive memory.
Neuron 106, 498–514.e8 (2020).
158. Lin, R., Liang, J. & Luo, M. The raphe dopamine system: roles in salience encoding,
memory expression, and addiction. Trends Neurosci. 44, 366–377 (2021).
159. Yu, W. et al. Periaqueductal gray/dorsal raphe dopamine neurons contribute to sex
dierences in pain-related behaviors. Neuron 109, 1365–1380.e5 (2021).
160. Darvas, M., Fadok, J. P. & Palmiter, R. D. Requirement of dopamine signaling in the
amygdala and striatum for learning and maintenance of a conditioned avoidance
response. Learn. Mem. 18, 136–143 (2011).
161. Fadok, J. P., Dickerson, T. M. K. & Palmiter, R. D. Dopamine is necessary for cue-dependent
fear conditioning. J. Neurosci. 29, 11089–11097 (2009).
162. Fadok, J. P., Darvas, M., Dickerson, T. M. K. & Palmiter, R. D. Long-term memory for
pavlovian fear conditioning requires dopamine in the nucleus accumbens and
basolateral amygdala. PLoS ONE 5, e12751 (2010).
163. Morel, C. et al. Midbrain projection to the basolateral amygdala encodes anxiety-like
but not depression-like behaviors. Nat. Commun. 13, 1–13 (2022).
164. Ball, K. T., Bennardo, G. M., Roe, J. & Wunderlich, K. J. Dopamine D1-like receptors in
prelimbic, but not infralimbic, medial prefrontal cortex contribute to chronic stress-
induced increases in cue-induced relapse to palatable food seeking during forced
abstinence. Behav. Brain Res. 417, 113583 (2022).
165. Zubair, M. et al. Divergent whole brain projections from the ventral midbrain in
macaques. Cereb. Cortex 31, 2913 (2021).
166. Kramer, D. J. et al. Generation of a DAT-P2A-Flpo mouse line for intersectional genetic
targeting of dopamine neuron subpopulations. Cell Rep. 35, 109123 (2021).
167. Nakamura, S., Ito, Y., Shirasaki, R. & Murakami, F. Local directional cues control growth
polarity of dopaminergic axons along the rostrocaudal Axis. J. Neurosci. 20, 4112–4119
(2000).
168. Gates, M. A., Coupe, V. M., Torres, E. M., Fricker-Gates, R. A. & Dunnett, S. B. Spatially and
temporally restricted chemoattractive and chemorepulsive cues direct the formation of
the nigro-striatal circuit. Eur. J. Neurosci. 19, 831–844 (2004).
169. Prestoz, L., Jaber, M. & Gaillard, A. Dopaminergic axon guidance: which makes what?
Front. Cell. Neurosci. 6, 32 (2012).
170. van den Heuvel, D. M. A. & Pasterkamp, R. J. Getting connected in the dopamine system.
Prog. Neurobiol. 85, 75–93 (2008).
171. Marillat, V. et al. Spatiotemporal expression patterns of slit and robo genes in the rat
brain. J. Comp. Neurol. 442, 130–155 (2002).
172. Fenstermaker, A. G. et al. Wnt/planar cell polarity signaling controls the anterior–
posterior organization of monoaminergic axons in the brainstem. J. Neurosci. 30,
16053–16064 (2010).
173. Lin, L ., Rao, Y. & Isacson, O. Netrin-1 and slit-2 regulate and direct neurite growth
of ventral midbrain dopaminergic neurons. Mol. Cell. Neurosci. 28, 547–555
(2005).
174. Hernández-Montiel, H. L., Tamariz, E., Sandoval-Minero, M. T. & Varela-Echavarría, A.
Semaphorins 3A, 3C, and 3F in mesencephalic dopaminergic axon pathinding. J. Comp.
Neurol. 506, 387–397 (2008).
175. Yamauchi, K. et al. FGF8 signaling regulates growth of midbrain dopaminergic axons by
inducing semaphorin 3F. J. Neurosci. 29, 4044–4055 (2009).
176. Blakely, B. D. et al. Wnt5a regulates midbrain dopaminergic axon growth and guidance.
PLoS ONE 6, e18373 (2011).
177. Kolk, S. M. et al. Semaphorin 3F is a bifunctional guidance cue for dopaminergic axons
and controls their fasciculation, channeling, rostral growth, and intracortical targeting.
J. Neurosci. 29, 12542–12557 (2009).
178. Torre, E. R., Gutekunst, C. A. & Gross, R. E. Expression by midbrain dopamine neurons of
Sema3A and 3F receptors is associated with chemorepulsion in vitro but a mild in vivo
phenotype. Mol. Cell Neurosci. 44, 135–153 (2010).
179. Hammond, R., Blaess, S. & Abeliovich, A. Sonic hedgehog is a chemoattractant for
midbrain dopaminergic axons. PLoS ONE 4, e7007 (2009).
180. Soleilhavoup, C. et al. Nolz1 expression is required in dopaminergic axon guidance and
striatal innervation. Nat. Commun. 11, 3111 (2020).
181. Marín, O., Baker, J., Puelles, L. & Rubenstein, J. L. R. Patterning of the basal telencephalon
and hypothalamus is essential for guidance of cortical projections. Development 129,
761–773 (2002).
182. Dugan, J. P., Stratton, A., Riley, H. P., Farmer, W. T. & Mastick, G. S. Midbrain dopaminergic
axons are guided longitudinally through the diencephalon by Slit/Robo signals.
Mol. Cell. Neurosci. 46, 347–356 (2011).
183. Kawano, H. et al. Aberrant trajectory of ascending dopaminergic pathway in mice lacking
Nkx2.1. Exp. Neurol. 182, 103–112 (2003).
184. Bagri, A. et al. Slit proteins prevent midline crossing and determine the dorsoventral
position of major axonal pathways in the mammalian forebrain. Neuron 33, 233–248
(2002).
185. Deschamps, C. et al. EphrinA5 protein distribution in the developing mouse brain.
BMC Neurosci. 11, 105 (2010).
186. Deschamps, C., Faideau, M., Jaber, M., Gaillard, A. & Prestoz, L. Expression of ephrinA5
during development and potential involvement in the guidance of the mesostriatal
pathway. Exp. Neurol. 219, 466–480 (2009).
187. García-Peña, C. M. et al. Ascending midbrain dopaminergic axons require descending
GAD65 axon fascicles for normal pathinding. Front. Neuroanat. 8, 43 (2014).
188. Schmidt, E. R. E. et al. Subdomain-mediated axon-axon signaling and chemoattraction
cooperate to regulate aerent innervation of the lateral habenula. Neuron 83, 372–387
(2014).
189. Prensa, L. & Parent, A. The nigrostriatal pathway in the rat: a single-axon study of the
relationship between dorsal and ventral tier nigral neurons and the striosome/matrix
striatal compartments. J. Neurosci. 21, 7247–7260 (2001).
190. Matsuda, W. et al. Single nigrostriatal dopaminergic neurons form widely spread
and highly dense axonal arborizations in the neostriatum. J. Neurosci. 29, 444–453
(2009).
191. Aransay, A., Rodríguez-López, C., García-Amado, M., Clascá, F. & Prensa, L. Long-range
projection neurons of the mouse ventral tegmental area: a single-cell axon tracing
analysis. Front. Neuroanat. 9, 59 (2015).
192. Barker, D. J., Root, D. H., Zhang, S. & Morales, M. Multiplexed neurochemical signaling
by neurons of the ventral tegmental area. J. Chem. Neuroanat. 73, 33–42 (2016).
193. Gauthier, J., Parent, M., Lévesque, M. & Parent, A. The axonal arborization of single
nigrostriatal neurons in rats. Brain Res. 834, 228–232 (1999).
194. Zhang, S. et al. Dopaminergic and glutamatergic microdomains in a subset of rodent
mesoaccumbens axons. Nat. Neurosci. 18, 386–392 (2015).
195. Fortin, G. M. et al. Segregation of dopamine and glutamate release sites in dopamine
neuron axons: regulation by striatal target cells. FASEB J. 33, 400–417 (2019).
196. Banerjee, A. et al. Molecular and functional architecture of striatal dopamine release
sites. Neuron 110, 248–265.e9 (2022).
197. Pereira, D. B. et al. Fluorescent false neurotransmitter reveals functionally silent
dopamine vesicle clusters in the striatum. Nat. Neurosci. 19, 578–586 (2016).
198. Manier, M. et al. Striatal target‐induced axonal branching of dopaminergic mesencephalic
neurons in culture via diusible factors. J. Neurosci. Res. 48, 358–371 (1997).
199. Hu, Z., Cooper, M., Crockett, D. P. & Zhou, R. Dierentiation of the midbrain dopaminergic
pathways during mouse development. J. Comp. Neurol. 476, 301–311 (2004).
200. Flanagan, J. G. Neural map speciication by gradients. Curr. Opin. Neurobiol. 16, 59–66
(2006).
201. Jaumotte, J. D. & Zigmond, M. J. Dopaminergic innervation of forebrain by ventral
mesencephalon in organotypic slice co-cultures: eects of GDNF. Brain Res. Mol.
Brain Res. 134, 139–146 (2005).
202. Janis, L. S., Cassidy, R. M. & Kromer, L. F. Ephrin-A binding and EphA receptor
expression delineate the matrix compartment of the striatum. J. Neurosci. 19, 4962–4971
(1999).
203. Yamaguchi, T., Wang, H. L., Li, X., Ng, T. H. & Morales, M. Mesocorticolimbic
glutamatergic pathway. J. Neurosci. 31, 8476 (2011).
204. Islam, K. U. S., Meli, N. & Blaess, S. The development of the mesoprefrontal dopaminergic
system in health and disease. Front. Neural Circuits 15, 746582 (2021).
205. Reynolds, L. M. et al. DCC receptors drive prefrontal cortex maturation by determining
dopamine axon targeting in adolescence. Biol. Psychiatry 83, 181 (2018).
206. Manitt, C. et al. The netrin receptor DCC is required in the pubertal organization
of mesocortical dopamine circuitry. J. Neurosci. 31, 8381 (2011).
207. Cuesta, S. et al. Dopamine axon targeting in the nucleus accumbens in adolescence
requires Netrin-1. Front. Cell Dev. Biol. 8, 487 (2020).
208. Pasterkamp, R. J., Kolk, S. M., Hellemons, A. J. & Kolodkin, A. L. Expression patterns
of semaphorin7A and plexinC1during rat neural development suggest roles in axon
guidance and neuronal migration. BMC Dev. Biol. 7, 98 (2007).
209. Chabrat, A. et al. Transcriptional repression of Plxnc1 by Lmx1a and Lmx1b directs
topographic dopaminergic circuit formation. Nat. Commun. 8, 933 (2017).
210. Chung, C. Y. et al. The transcription factor orthodenticle homeobox 2 inluences axonal
projections and vulnerability of midbrain dopaminergic neurons. Brain 133, 2022
(2010).
211. Shigeoka, T. et al. Dynamic axonal translation in developing and mature visual circuits.
Cell 166, 181–192 (2016).
212. Kegeles, L. S. et al. Increased synaptic dopamine function in associative regions of the
striatum in schizophrenia. Arch. Gen. Psychiatry 67, 231–239 (2010).
213. McCutcheon, R. A., Abi-Dargham, A. & Howes, O. D. Schizophrenia, dopamine and the
striatum: from biology to symptoms. Trends Neurosci. 42, 205–220 (2019).
214. Poisson, C. L., Engel, L. & Saunders, B. T. Dopamine circuit mechanisms of addiction-like
behaviors. Front. Neural Circuits 15, 752420 (2021).
215. Corre, J. et al. Dopamine neurons projecting to medial shell of the nucleus accumbens
drive heroin reinforcement. Elife 7, 1–22 (2018).
216. Cassidy, C. M. et al. Evidence for dopamine abnormalities in the substantia nigra in
cocaine addiction revealed by neuromelanin-sensitive MRI. Am. J. Psychiatry 177,
1038–1047 (2020).
217. Fearnley, J. M. & Lees, A. J. Ageing and Parkinson’s disease: substantia nigra regional
selectivity. Brain 114, 2283–2301 (1991).
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 151
Review article
218. Gibb, W. R. G. & Lees, A. J. Anatomy, pigmentation, ventral and dorsal subpopulations
of the substantia nigra, and dierential cell death in Parkinson’s disease. J. Neurol.
Neurosurg. Psychiatry 54, 388–396 (1991).
219. Liu, G. et al. Aldehyde dehydrogenase 1 deines and protects a nigrostriatal
dopaminergic neuron subpopulation. J. Clin. Invest. 124, 3032–3046 (2014).
220. Schwarz, S. T. et al. Parkinson’s disease related signal change in the nigrosomes 1–5 and the
substantia nigra using T2* weighted 7T MRI. Neuroimage Clin. 19, 683–689 (2018).
221. Huddleston, D. E. et al. In vivo detection of lateral–ventral tier nigral degeneration in
Parkinson’s disease. Hum. Brain Mapp. 38, 2627–2634 (2017).
222. Sulzer, D. et al. Neuromelanin biosynthesis is driven by excess cytosolic catecholamines
not accumulated by synaptic vesicles. Proc. Natl Acad. Sci. USA 97, 11869–11874 (2000).
223. Segura-Aguilar, J. et al. Protective and toxic roles of dopamine in Parkinson’s disease.
J. Neurochem. 129, 898–915 (2014).
224. Yamada, T., McGeer, P. L., Baimbridge, K. G. & McGeer, E. G. Relative sparing in
Parkinson’s disease of substantia nigra dopamine neurons containing calbindin-D28K.
Brain Res. 526, 303–307 (1990).
225. German, D. C., Manaye, K. F., Brooksd, B. A. & Sonsalla, P. K. Midbrain dopaminergic
cell loss in Parkinson’s disease and MPTP-induced parkinsonism: sparing of
calbindin-D28k–containing cells. Ann. N. Y. Acad. Sci. 648, 42–62 (1992).
226. Liang, C. L., Sinton, C. M., Sonsalla, P. K. & German, D. C. Midbrain dopaminergic neurons
in the mouse that contain calbindin-D28k exhibit reduced vulnerability to MPTP-induced
neurodegeneration. Neurodegeneration 5, 313–318 (1996).
227. Rcom-H’cheo-Gauthier, A., Goodwin, J. & Pountney, D. L. Interactions between calcium
and alpha-synuclein in neurodegeneration. Biomolecules 4, 795–811 (2014).
228. Post, M. R., Lieberman, O. J. & Mosharov, E. V. Can interactions between α-synuclein,
dopamine and calcium explain selective neurodegeneration in Parkinson’s disease?
Front. Neurosci. 12, 161 (2018).
229. Uittenbogaard, M., Baxter, K. K. & Chiaramello, A. The neurogenic basic helix-loop-helix
transcription factor NeuroD6 confers tolerance to oxidative stress by triggering an
antioxidant response and sustaining the mitochondrial biomass. ASN Neuro 2, 115–133
(2010).
230. Buck, S. A. et al. VGLUT2 is a determinant of dopamine neuron resilience in a rotenone
model of dopamine neurodegeneration. J. Neurosci. 41, 4937–4947 (2021).
231. Buck, S. A. et al. Roles of VGLUT2 and dopamine/glutamate co-transmission in selective
vulnerability to dopamine neurodegeneration. ACS Chem. Neurosci. 13, 187–193 (2022).
232. Steinkellner, T. et al. Dopamine neurons exhibit emergent glutamatergic identity in
Parkinson’s disease. Brain 143, 879–886 (2021).
233. Björklund, A. & Stenevi, U. Reconstruction of the nigrostriatal dopamine pathway by
intracerebral nigral transplants. Brain Res. 177, 555–560 (1979).
234. Lindvall, O. et al. Human fetal dopamine neurons grafted into the striatum in two patients
with severe Parkinson’s disease: a detailed account of methodology and a 6-month
follow-up. Arch. Neurol. 46, 615–631 (1989).
235. Lindvall, O. et al. Grafts of fetal dopamine neurons survive and improve motor function
in Parkinson’s disease. Science 247, 574–577 (1990).
236. Parmar, M., Torper, O. & Drouin-Ouellet, J. Cell-based therapy for Parkinson’s disease: a
journey through decades toward the light side of the Force. Eur. J. Neurosci. 49, 463–471
(2019).
237. Henchclie, C. & Sarva, H. Restoring function to dopaminergic neurons: progress in the
development of cell-based therapies for Parkinson’s disease. CNS Drugs 34, 559–577
(2020).
238. Björklund, A. & Parmar, M. Dopamine cell therapy: from cell replacement to circuitry
repair. J. Parkinsons Dis. 11, S159–S165 (2021).
239. Guo, X., Tang, L. & Tang, X. Current developments in cell replacement therapy for
Parkinson’s disease. Neuroscience 463, 370–382 (2021).
240. Li, J. Y. & Li, W. Postmortem studies of fetal grafts in Parkinson’s disease: what lessons
have we learned? Front. Cell Dev. Biol. 9, 666675 (2021).
241. Rodríguez-Pallares, J., García-Garrote, M., Parga, J. & Labandeira-García, J. Combined
cell-based therapy strategies for the treatment of Parkinson’s disease: focus on
mesenchymal stromal cells. Neural Regen. Res. 18, 478 (2023).
242. Gaillard, A. et al. Anatomical and functional reconstruction of the nigrostriatal pathway
by intranigral transplants. Neurobiol. Dis. 35, 477–488 (2009).
243. Kirkeby, A. et al. Predictive markers guide dierentiation to improve graft outcome in
clinical translation of hESC-based therapy for Parkinson’s disease. Cell Stem Cell 20,
135–148 (2017).
244. Grealish, S. et al. Human ESC-derived dopamine neurons show similar preclinical
eicacy and potency to fetal neurons when grafted in a rat model of Parkinson’s disease.
Cell Stem Cell 15, 653–665 (2014).
245. Aldrin-Kirk, P. et al. A novel two-factor monosynaptic TRIO tracing method for
assessment of circuit integration of hESC-derived dopamine transplants. Stem Cell Rep.
17, 159–172 (2022).
246. Morizane, A. et al. MHC matching improves engraftment of iPSC-derived neurons in non-
human primates. Nat. Commun. 8, 385 (2017).
247. Morizane, A. et al. Direct comparison of autologous and allogeneic transplantation of
IPSC-derived neural cells in the brain of a nonhuman primate. Stem Cell Rep. 1, 283–292
(2013).
248. Schweitzer, J. S. et al. Personalized iPSC-derived dopamine progenitor cells for
Parkinson’s disease. N. Engl. J. Med. 382, 1926–1932 (2020).
249. Tao, Y. et al. Autologous transplant therapy alleviates motor and depressive behaviors in
parkinsonian monkeys. Nat. Med. 27, 632–639 (2021).
250. van de Haar, L. L. et al. Molecular signatures and cellular diversity during mouse
habenula development. Cell Rep. 40, 111029 (2022).
251. Melani, R. & Tritsch, N. X. Inhibitory co-transmission from midbrain dopamine neurons
relies on presynaptic GABA uptake. Cell Rep. 39, 110716 (2022).
252. Parkinson, J. An essay on the shaking palsy. J. Neuropsychiatry Clin. Neurosci. 14,
223–236 (2002).
253. Lees, A. J., Hardy, J. & Revesz, T. Parkinson’s disease. Lancet 373, 2055–2066 (2009).
254. Dickson, D. W. Parkinson’s disease and parkinsonism: neuropathology. Cold Spring Harb.
Perspect. Med. 2, a009258 (2012).
255. Schapira, A. H. V., Chaudhuri, K. R. & Jenner, P. Non-motor features of Parkinson disease.
Nat. Rev. Neurosci. 18, 435–450 (2017).
256. Blauwendraat, C., Nalls, M. A. & Singleton, A. B. The genetic architecture of Parkinson’s
disease. Lancet Neurol. 19, 170–178 (2020).
257. Polymeropoulos, M. H. et al. Mutation in the alpha-synuclein gene identiied in families
with Parkinson’s disease. Science 276, 2045–2047 (1997).
258. Healy, D. G. et al. Phenotype, genotype, and worldwide genetic penetrance of
LRRK2-associated Parkinson’s disease: a case-control study. Lancet Neurol. 7, 583–590
(2008).
259. Goker-Alpan, O. et al. Parkinsonism among Gaucher disease carriers. J. Med. Genet. 41,
937–940 (2004).
260. Kakkar, A. K. & Dahiya, N. Management of Parkinson’s disease: current and future
pharmacotherapy. Eur. J. Pharmacol. 750, 74–81 (2015).
261. Oxtoby, N. P. et al. Sequence of clinical and neurodegeneration events in Parkinson’s
disease progression. Brain 144, 975–988 (2021).
262. Elkouzi, A., Vedam-Mai, V., Eisinger, R. S. & Okun, M. S. Emerging therapies in Parkinson
disease — repurposed drugs and new approaches. Nat. Rev. Neurol. 15, 204–223 (2019).
263. Bolam, J. P. & Pissadaki, E. K. Living on the edge with too many mouths to feed: why
dopamine neurons die. Mov. Disord. 27, 1478–1483 (2012).
264. Pacelli, C. et al. Elevated mitochondrial bioenergetics and axonal arborization size are
key contributors to the vulnerability of dopamine neurons. Curr. Biol. 25, 2349–2360
(2015).
265. Giguère, N. et al. Increased vulnerability of nigral dopamine neurons after expansion
of their axonal arborization size through D2 dopamine receptor conditional knockout.
PLoS Genet. 15, 1–26 (2019).
266. Ricke, K. M. et al. Mitochondrial dysfunction combined with high calcium load leads
to impaired antioxidant defense underlying the selective loss of nigral dopaminergic
neurons. J. Neurosci. 40, 1975–1986 (2020).
267. Kanaan, N. M., Kordower, J. H. & Collier, T. J. Age-related changes in dopamine
transporters and accumulation of 3-nitrotyrosine in rhesus monkey midbrain dopamine
neurons: Relevance in selective neuronal vulnerability to degeneration. Eur. J. Neurosci.
27, 3205–3215 (2008).
268. Nakajima, S. et al. Age-related vulnerability to nigral dopaminergic degeneration in rats
via Zn2+-permeable GluR2-lacking AMPA receptor activation. Neurotoxicology 83, 69–76
(2021).
269. Shi, H. et al. Sirt3 protects dopaminergic neurons from mitochondrial oxidative stress.
Hum. Mol. Genet. 26, 1915–1926 (2017).
270. Guillot, T. S. & Miller, G. W. Protective actions of the vesicular monoamine transporter 2
(VMAT2) in monoaminergic neurons. Mol. Neurobiol. 39, 149–170 (2009).
271. Fahn, S. Does levodopa slow or hasten the rate of progression of Parkinson’s disease?
J. Neurol. 252, 37–42 (2005).
272. Mosharov, E. V. et al. Interplay between cytosolic dopamine, calcium, and α-synuclein
causes selective death of substantia nigra neurons. Neuron 62, 218–229 (2009).
273. Surmeier, D. J., Guzman, J. N., Sanchez-Padilla, J. & Schumacker, P. T. The role of calcium and
mitochondrial oxidant stress in the loss of substantia nigra pars compacta dopaminergic
neurons in Parkinson’s disease. Neuroscience 198, 221–231 (2011).
274. Zucca, F. A. et al. Interactions of iron, dopamine and neuromelanin pathways in brain
aging and Parkinson’s disease. Prog. Neurobiol. 155, 96–119 (2017).
275. Jansen van Rensburg, Z., Abrahams, S., Bardien, S. & Kenyon, C. Toxic feedback loop
involving iron, reactive oxygen species, α-synuclein and neuromelanin in Parkinson’s
disease and intervention with turmeric. Mol. Neurobiol. 58, 5920–5936 (2021).
276. Nedergaard, S., Flatman, J. A. & Engberg, I. Nifedipine- and omega-conotoxin-sensitive
Ca2+ conductances in guinea-pig substantia nigra pars compacta neurones. J. Physiol.
466, 727–747 (1993).
277. Philippart, F. et al. Dierential somatic Ca2+ channel proile in midbrain dopaminergic
neurons. J. Neurosci. 36, 7234–7245 (2016).
278. Conway, K. A., Rochet, J. C., Bieganski, R. M. & Lansbury, J. Kinetic stabilization of the
α-synuclein protoibril by a dopamine-α-synuclein adduct. Science 294, 1346–1349
(2001).
279. Ren, Y., Liu, W., Jiang, H., Jiang, Q. & Feng, J. Selective vulnerability of dopaminergic
neurons to microtubule depolymerization. J. Biol. Chem. 280, 34105–34112
(2005).
280. Ulusoy, A., Björklund, T., Buck, K. & Kirik, D. Dysregulated dopamine storage
increases the vulnerability to α-synuclein in nigral neurons. Neurobiol. Dis. 47,
367–377 (2012).
281. Biondetti, E. et al. The spatiotemporal changes in dopamine, neuromelanin and iron
characterizing Parkinson’s disease. Brain 144, 3114–3125 (2021).
282. Thomsen, M. B. et al. PET imaging reveals early and progressive dopaminergic deicits
after intra-striatal injection of preformed alpha-synuclein ibrils in rats. Neurobiol. Dis.
149, 105229 (2021).
Nature Reviews Neuroscience | Voume 24 | March 2023 | 134–152 152
Review article
283. Uchihara, T. An order in Lewy body disorders: retrograde degeneration in hyperbranching
axons as a fundamental structural template accounting for focal/multifocal Lewy body
disease. Neuropathology 37, 129–149 (2017).
284. Bellucci, A., Antonini, A., Pizzi, M. & Spano, P. F. The end is the beginning: Parkinson’s
disease in the light of brain imaging. Front. Aging Neurosci. 9, 330 (2017).
Acknowledgements
The authors thank P. Lingor for input on the manuscript. Work on the dopamine system in
the laboratory of the authors is supported by Stichting Parkinson Fonds, the Dutch Research
Council (NWO; ALW-VICI 865.14.004) and the NWO Gravitation programme BRAINSCAPES:
A Roadmap from Neurogenetics to Neurobiology (NWO: 024.004.012) to R.J.P. The authors
apologize to all investigators whose research could not be appropriately cited owing
to space limitations.
Author contributions
All authors contributed to all aspects of the article.
Competing interests
The authors declare no competing interests.
Additional information
Correspondence should be addressed to R. Jeroen Pasterkamp.
Peer review information Nature Reviews Neuroscience thanks S. Blaess, L. Zweifel
and the other, anonymous, reviewer(s) for their contribution to the peer review of
this work.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional ailiations.
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this
article under a publishing agreement with the author(s) or other rightsholder(s); author self-
archiving of the accepted manuscript version of this article is solely governed by the terms of
such publishing agreement and applicable law.
© Springer Nature Limited 2023
