Introduction
The behavioral manipulation hypothesis predicts that parasites can change host behavior
in a way that benefits the parasites and not the host (extensively reviewed in [1–9]).
In other words, the hypothesis predicts that genes of a parasite can produce an "extended"
phenotype that manifests beyond a parasite's soma [10]. Protozoan parasite Toxoplasma
gondii (henceforth toxoplasma) is an often-cited example. Chronic toxoplasma infection
reduces aversion of rodents to cat odors, plausibly increasing predation by its definitive
felid host [11]. Here, I enumerate main narratives that have emerged in the past decade
about biological mechanisms of behavioral change in rodents after toxoplasma infection.
Cats are infected by toxoplasma when they eat infected prey. The parasite undergoes
gametogenesis in cat intestines, resulting in eventual shedding of fecal oocysts that
are ingested by intermediate hosts. Entry in the cat is important for the parasite
because it permits a) sexual recombination; b) infection of herbivore hosts who otherwise
cannot be infected through carnivory between intermediate hosts; and c) the discharge
of highly infectious and resilient oocysts into the environment. Yet, entry of the
parasite in the cat is constrained by predation rates. Preys of cats avoid cats and
cat odors [12]. Apropos, toxoplasma infection leads to reduced aversion of rodents
to cat odors [11]. A subset of animals also develops an atypical and “fatal” attraction
[11,13]. These behavioral observations suggest, but do not prove, that the parasite
creates an extended phenotype in the host behavior. The caution in the preceding sentence
is necessary because it is yet unknown if infected rodents are indeed predated more
frequently by cats.
Toxoplasma is also sexually transmitted through the male ejaculate in rats [14]. Apropos,
male rats infected with toxoplasma become more attractive to females [15]. Uninfected
females spend greater time near infected males and allow them greater reproductive
access [14]. These observations suggest a second parasitic manipulation of the host
behavior, whereby being infected creates greater avenues for sexual transmission of
the parasite itself [9].
Biological pathways underlying mate choice and innate aversion to predator odor are
relatively well-studied in rodents. This has allowed researchers to study proximate
mechanisms of parasitic behavioral manipulation in greater detail in this association
compared to other host–parasite relationships. This mechanistic research has focused
on three main narratives.
Narrative #1: Tropism to Specific Regions of the Brain
This narrative posits that toxoplasma preferentially concentrates in certain brain
regions; and this tropism can explain host behavioral changes through local manipulation
of neuronal signaling and/or damage. Toxoplasma exhibits a decided tropism to brain,
testes, and eyes. These organs are immune-privileged, in the sense that immune cells
have limited access to these sites. Some experimental evidence suggests that toxoplasma
gains entry into these sites through brain endothelial cells [16] or by using dendritic
immune cells as Trojan horses [17]. The parasite then forms bradyzoite-containing
tissue cysts that undergo periodic cycles of rupture and encystment (i.e., recrudescence).
Several studies have mapped sites of encystment within the brain, with the hope that
location of these cysts has some bearing on the host behavior. Two earliest studies
in this regard reported a rather widespread occurrence of tissue cysts in a variety
of brain regions [13,18]. Both of these reports suggested a mild tropism to nucleus
accumbens, ventromedial hypothalamus, or amygdala. These brain structures are involved
in decision making and generation of fear [19,20]. Thus, the suggestion was that presence
of cysts somehow compromised normal functioning of these brain regions and led to
deficits in processing of fear or decision-making capacities. Yet, subsequent detailed
analysis of cyst distribution failed to reveal substantial tropism in any of these
three brain structures in mice and rats, instead reporting a rather “probabilistic”
spread of the parasites [21]. Even more surprisingly, further experiments demonstrated
that change in host behavior could be observed even after extensive clearance of parasite
cysts within brain [22]. This suggests that toxoplasma tropism or lack of tropism
does not have a causal relationship with the behavioral change. A nontropic model
must thus be sought. Opinions still remain divided on this issue. For example, it
has been suggested that within infected animals, those with presence of parasitic
cysts within certain brain regions experience greater magnitude of behavioral change
[23]. Another possibility in this regard is that toxoplasma could “coopt” brain cells
without invading them [24]. The parasite is known to inject effector proteins inside
host cells during early phases of invasion [24]. In many of the brain cells, the parasite
injects these effector proteins but then does not, or fails to, gain residence. Such
covert tropism will be difficult to detect by mere enumeration of parasite presence
in various brain regions. Excitingly, new transgenic methods now allow visualization
of “toxoplasma-kissed” neurons [25], although selective tropism of such events is
presently unstudied.
Narrative #2: Disruption of Dopamine Signaling in the Brain
An important candidate in the nontropic model is disruption of brain dopamine signaling.
The genome of toxoplasma contains two genes (AAH1 and AAH2) that bear striking sequence
similarity to a mammalian enzyme called tyrosine hydroxylase [26]. This enzyme in
mammals catalyzes conversion of L-tyrosine to L-3,4-dihydroxyphenylalanine. This reaction
is a rate-limiting step in synthesis of dopamine. While dopamine is often presented
as the neurotransmitter that signals reward or pleasure, its more parsimonious role
in the brain is in motivation and goal-directed behaviors [27]. The narrative then
is that toxoplasma increases dopamine signaling in the host brain by virtue of supplying
a rate-limiting enzyme for its synthesis [28,29]. This increase in dopamine signaling
then interferes with host behavior, creating atypical motivation to explore predator
odors. In support of this, drugs that interfere with the binding of dopamine to its
receptors ameliorate effects of the infection on aversion to predator odors [30].
Tissue cysts within the infected brain contain high amounts of dopamine, and cultured
dopaminergic cells secrete greater dopamine when infected with toxoplasma, suggesting
the possibility of hyperdopaminergic drive in the infected brain [28]. This hypothesis
does not necessarily require tropism of the parasite within brain regions that endogenously
produce this neurotransmitter. A stochastic distribution of the parasite will provide
tyrosine hydroxylase to a wide variety of brain regions. Yet, it is unlikely that
nondopaminergic neurons will contain complementary enzymes required for dopamine synthesis.
Moreover, any residual dopamine production in nonendogenous circuits is unlikely to
have any effect because of lack of dopamine receptors in downstream efferent. In short,
an intersection of generalized increase in tyrosine hydroxylase in brain with specific
distribution of complementary proteins in endogenous dopaminergic circuit could create
a rather specific behavior alteration. This hypothesis can reconcile lack of strong
tropism of the parasite. Yet it obligatorily requires persistent presence of the parasite
within brain. The relationship of dopamine alteration and increase in sexual attractiveness
of infected males also remains unclear at present. A definitive proof of this hypothesis
will require a future experiment with possible demonstration that disruption of parasitic
tyrosine hydroxylase genes within brain necessarily results in loss of host behavioral
change without affecting parasite survival itself.
Recently, a mutant parasite with ablation of one of the AAH genes (AAH2) has been
described [29]. This ablation does not affect parasite viability, invasion, and transmission.
It will be interesting to ask if infection with this mutant still causes host behavioral
change. Contrary to predictions from dopaminergic mediation, infection with wild-type
or ΔAAH2 mutant toxoplasma did not result in greater dopamine in mice brain. Similarly,
overexpression of AAH2 genes in cell culture failed to augment dopamine content [29].
A double mutant parasite lacking both AAH genes has still not been successfully created.
Narrative #3: Hormonal Upheavals and Concomitant Epigenetic Changes
An alternative nontropic model invokes parasite-induced tweaking of communication
lines between brain and gonadal hormones. Toxoplasma invades rat testes upon infection,
leading to a heavy cyst burden in epididymis and the ejaculates [14]. This is not
surprising because the testes, like the brain, is an immune-privileged site. The infection
results in up-regulation of testosterone synthesis within Leydig cells of the testes
[31]. The narrative here posits that greater testosterone synthesis results in two
simultaneous effects. One, it increases synthesis of male sexual pheromone by virtue
of androgen dependence of these molecules [15]. Two, excess testosterone shifts the
host towards sexual behaviors and away from defensive behaviors [32]. In support of
this narrative, castrating male rats before infection prevents host behavioral change
[31]. An important caveat in this experiment is the possibility that removal of testosterone
by castration can potentially increase tonicity of the immune response during the
acute phase of the infection, thereby changing the course of the infection. An unequivocal
demonstration of the direct role of testosterone would require selective ablation
of testosterone “increase” postinfection rather than removal of all testicular steroidogenesis
preinfection.
This narrative nonetheless provides a plausible chain of events. In male rats, testosterone
sustains synthesis of major urinary proteins in the liver. These proteins are necessary
and sufficient to signal sexual attractiveness in a dose-dependent manner to females
when eventually excreted in the urine [33]. Testosterone and/or its metabolic derivatives
bind to their receptor found in the medial amygdala, a brain region involved in signaling
presence of sexual opportunities [34,35]. This brain region contains population of
neurons expressing arginine vasopressin, a neurotransmitter that mediates reproductive
behaviors in many species (e.g., [36,37]). Male rats infected with toxoplasma exhibit
reduced DNA methylation in promoter sites upstream of the arginine vasopressin gene
in the medial amygdala, resulting in its greater production [32]. This is akin to
observations noted during testosterone supplementation in uninfected rats [38]. These
arginine vasopressin neurons are typically recruited during copulation or sensory
stimulation by female presence in uninfected rats [37]. Atypically, the same population
of neurons becomes activated during exposure to cat odors in infected male rats. Furthermore,
pharmacological mimicry of this molecular change institutes decreased predator aversion
akin to the effects of toxoplasma infection [32]. Thus, the proposal here is that
the increase in testosterone synthesis mediates both aspects of behavioral manipulation
of sexual attractiveness and reduction in aversion to cats. The presence of the parasite
within the brain or its tropism to certain brain regions becomes merely incidental
and non-necessary in this narrative.
The narrative detailed above encompasses substrates that are highly dimorphic between
genders. Congruent to males, toxoplasma also reduces aversion to cat odors in female
mice and rats [13,39]. Thus, proximate mechanisms involving gender-dimorphic biology
needs to be yet reconciled with gender-nondimorphic behavioral effects of the parasitism.
Mechanisms of the behavioral change in females are currently understudied, though
preliminary evidence suggests changes in progesterone levels postinfection [39].
Conclusions
Rats and mice infected with toxoplasma exhibit behavioral change in their aversion
to cat odor and their sexual attractiveness to females. Previous work has resulted
in three main classes of hypotheses pertaining to proximate mechanism of this phenomenon.
Current work continues to test these hypotheses. More clarity about the mechanisms
will plausibly inform ultimate causation of host behavioral change.