Immunization with DAT fragments generated a transient auto-immune response, with a marked elevation in titers of circulating aAbs directed against DAT. Notably, this response resulted in long-lasting impairment of striatal DA systems’ function, with enduring rise of HVA and also DOPAC levels. Such neuro-chemical sequel was consistently associated with a profile of behavioural modifications, including a rearrangement in patterns of nocturnal activity and diurnal resting, spontaneously expressed within home-cages settings. Such subtle effects in the wake-sleep rhythm were expected based on earlier data [49, 55]: these effects are relevant, since they denote a longer and more intense beginning as well as an earlier end of the daily spontaneous activity in DAT-i mice. Interestingly, a considerable impairment of their adaptive capacity was found when considering choice behaviour in the delay-of-reward task contingencies.
Changes in behavioural flexibility and impairment in forebrain DA
In a classical study from our group , we assessed spontaneous home-cage activity in the SHR rat, a validated model for ADHD . Interestingly, these rats showed increased activity during specific time-points of the dark phase, as well as peaks of activity with interruption of rest during the light phase. Thus, the circadian rhythm of present DAT-i mice, consisting in a prolonged peak of spontaneous activity after wake-up, an anticipated start for the resting period, as well as the tendency for diurnal sleep to be interrupted by a wake-up peak, shares some interesting features with these rats modelling for ADHD.
As for operant-choice behaviour, control mice were able to perform the whole delay-of-reward task correctly. As expected, all of the subjects developed a clear LL preference during training phase and then modified their choice preference according to the introduction of delays. Indeed, they displayed either a slight and transient or a robust and persistent shift towards SS, denoting their subjective discounting of value attributed to the LL option. As classically reported with this kind of protocols [57, 71], discounting (or not) of LL value identifies these individuals as more (or less) delay-intolerant.
Notably, mice of the DAT-i group appeared greatly impaired in operant choice between two alternatives differing for reward size and delay contingencies. Noteworthy, they were perfectly capable to associate nose-poking in either hole with the corresponding magazine of food-delivery. Nose-poking was followed by ramp climbing and food eating, ruling out major biases due to gross motoric or cognitive deficit. Their impairment was apparently specific to the formation and/or the flexible modification of a choice strategy. In fact, half of the subjects (termed DAT-i flexible, with a “slow” profile) were hardly able 1) to discriminate the difference in reward size between LL and SS, and 2) to detect and/or react to delay intervals. These subjects were still flexible, but appeared to be slow both in developing preference for the larger-size reward and in reacting to its delay; this, despite extensive training and testing. In the other half of subjects (termed DAT-i inflexible, with a “stuck” profile), a more marked impairment emerged: a net preference for LL seemed to be expressed, which was likely the result of a rigid pattern of choice. Closer analyses on sequence of choices may be useful to ascertain if this was indeed a habit (informal observation of mice suggested that each SS option was followed by two LL options quite regularly).
Present DAT-i mice displayed two profiles, consisting of either slower acquisition of LL preference plus late reaction to delay, or rigid habit-based responding. Both profiles may be interpreted as reduced or entirely impaired ability to develop a new pattern of choice and/or to modify a previously acquired one. Thus, as an enduring consequence of DAT auto-immune reactions, functional impairment can be proposed for DAT-i mice within the striatum (possibly, its dorso-medial region). In other words, a difficulty in modifying choice strategy and/or a rigid preference profile, shown by DAT-i sub-populations of mice, reflect different degrees of inflexibility due to impaired striatal function. Neuro-chemical data are in agreement with this notion since both subgroups of DAT-i mice showed enhanced HVA (alone or together with DOPAC) in the striatum. Since DA levels were unchanged, it can be proposed that excessive DA metabolism occurred as an enduring adaptation to the auto-immune insult. None of the multiple possibilities can be dismissed (excessive DA synthesis, release or leakage; reduced re-uptake by diminished DAT function or levels); however, an enhanced DA catabolism (with a dynamic limit in HVA fate and its consequent accumulation) seems to be more likely. In some cases, a very big HVA accumulation was denoted by a rise of DOPAC, its intermediate precursor. DOPAC was increased in inflexible (with a “stuck” profile) but not in flexible (with a “slow” profile) DAT-i mice, suggesting that rate of DA metabolism was of an extreme extent in the former subgroup, causing the accumulation of catabolic products deriving from DA.
Moreover, independently from immunization, enhanced products of DA metabolism were found in the PFC of some individuals: specifically, subgroups with a “flat” or a “slow” profile displayed more attraction to LL together with lower HVA in the PFC. Conversely, enhanced DA catabolism may suggest poorer dopaminergic function [72, 73], so that cortical control would be worsened, in the subjects with a “steep” or a “stuck” profile. These data point to the role of PFC as subserving inhibitory control over the striatum . The physiological role of PFC is promoting the development of inhibition over impulsive drives, at least in the less flexible sub-population within VEH controls; a reduced PFC function may thus explain the “impulsive” choice strategy shown by the most flexible subjects within VEH controls. Following immunization, and in the presence of a compromised striatal function, the PFC may allow residual learning abilities as observed within flexible DAT-i mice. A relatively reduced function within the PFC, on the other hand, would contribute to the choice-rigidity observed in inflexible subjects among DAT-i mice.
Since the PFC is known to be devoid of DAT, it cannot have been a direct target of the DAT-directed auto-immunity. Such PFC-related features were likely pre-existing across groups. It is noteworthy that we were presently able to identify two separate subgroups of mice, based on their behavioural strategy in the operant choice task: differing levels of flexibility could be associated with diverse DA dynamics in the PFC, and this in the absence of any difference concerning levels of 5-HT nor of its metabolite.
Implications: auto-immunity in neuro-psychiatry
To interpret the present results, we propose that a transient rise of DAT aAbs, generated by immunization with DAT fragments, induced sequelae onto the striatal complex (possibly in its specific sub-regions), and a long-lasting interference with the corresponding behaviours. Thus, auto-immune reactions, with an increase of circulating aAbs to specific neural targets, may contribute to the etiology of psychiatric symptoms, at least to a certain extent [43–46]. There are many open issues, however, about such a working hypothesis, namely: 1) how can anti-CNS aAbs develop; and 2) how can these aAbs reach their CNS targets and exert their consequences. To cover the first issue (generation of anti-CNS aAbs), two hypotheses can be proposed. First, antibodies generated by immune reactions against an infectious agent may turn out to recognize some epitopes on healthy cells (see ), a typical mechanism proposed to account for some neuro-behavioural alterations observed within the so-called pediatric autoimmune neuropsychiatric disorders associated with streptococcal infections (i.e., PANDAS). Second, neuro-receptor fragments may themselves cause generation of auto-immune antibodies, at least in some cases. Indeed, since receptors and proteins expressed by neurons are fragmented during degradation, these fragments may occasionally enter into the bloodstream and act as antigenic peptides [76, 77]. This second hypothesis obviously postulates a temporary leakage of blood–brain barrier (BBB).
Such hypothesis should also be involved under the second open issue (i.e., reaching of CNS targets by aAbs), and therefore deserves more investigation. The BBB prevents developing B-cells in the bloodstream from being exposed to unique brain antigens: therefore, there are no mechanisms to establish tolerance to brain antigens and/or to prevent the production of antibodies against them. BBB failure would facilitate drainage of CNS antigens to peripheral lymphoid organs: subsequently, B-cell recognition of CNS antigens would cause: 1) B-cell proliferation and aAbs production; 2) presentation of CNS antigens to T-cells, with interactions between T- and B- cells amplifying the auto-immune response; 3) suggestion that activated B-cells may release aAbs into the circulating bloodstream [28, 29]. A related question, about the present anti-DAT aAbs, is then how can they reach the DA pathway, and where the auto-immune insult actually takes place.
As a matter of fact, DAT is not only present in striatal synapses: DA neurons in ventral mesencephalon, i.e. in the Substantia Nigra (SN) pars compacta (SNc; cell group A9) and in the Ventral Tegmental Area (VTA; cell group A10), are enriched with moderate to high DAT immuno-reactive intensity [78, 79], where it is concentrated in perikarya, dendrites, and axons. The presence of DAT in the somato-dendritic and axonal compartments of the VTA and the SNc is particularly intriguing, since the nigro-striatal dopaminergic system is specially sensitive to changes in BBB integrity , a feature recently associated to Parkinson’s disease (PD). The specific vulnerability of nigral dopaminergic neurons has already been demonstrated in many degeneration models: the Vascular Endothelial Growth Factor (VEGF), Histamine, and Lipo-poly-saccharide (LPS), are all able to disrupt the BBB in the entire ventral mesencephalon, and to reduce cell bodies and fibres of DA neurons in the SNc [81–83]. Previous studies have suggested that microglial density is high in the SNc , mediating excessive inflammatory reactions and causing dopaminergic neurons in the SNc to be vulnerable: any transient BBB disruption in the SNc likely results in the vigorous infiltration of neutrophils, which in turn could trigger a vicious cycle of astrocyte and endothelial cell damage, BBB permeability, and further neutrophil infiltration . Therefore, since BBB integrity is crucial to preserve dopaminergic neurons, we propose that a transiently disrupted BBB, in association with immunization procedures, opened the ventral mesencephalon to the action of anti-DAT aAbs, with the permanent consequence of a reduced number of DA cell bodies in the SNc and of subsequent alterations within mesencephalic fibres projecting to the dorsal striatum.
Limitations and future perspectives
The present results do raise a number of questions which remain unfortunately unanswered. For instance, details about production of DAT aAbs as well as in vivo effects of these aAbs on DAT activity are unknown yet (though, check with ref. ). Future investigation is warranted to gather full details of antibody development, to verify BBB permeability of such aAbs, and to assess selectivity of effects on DAT activity. We underline that several experimental approaches could be used.
In order to address the first point (i.e., how can DAT aAbs develop in pathogenesis), two main questions should be answered: Are neuro-receptor fragments produced in vivo as a side consequence of protein degradation? Are neuro-receptor fragments able to efflux across the BBB? In addition, once postulated that anti-DAT aAbs could be somehow produced, it would be interesting to investigate the BBB influx in the mouse, using isolated/purified and I-125/123/131-radiolabelled anti-DAT aAbs. An alternative way to explore the BBB permeability issue could be to detect these aAbs directly within the brain tissues or at least in the CSF.
Finally, it would be helpful to differentiate between a direct effect (linked to putative DAT protein / anti-DAT aAb interaction) and indirect effects brought by circulating aAbs, as it cannot be excluded that central consequences may arise without the need for these aAbs to cross the BBB. In this line, a direct intra-cerebral injection of these anti-DAT aAbs would allow to study the putative direct effects on functional DAT activity and behaviour. In vivo micro-dialysis and fast-scan cyclic voltammetry (FSCV) could be used to assess the actual DA dynamics in brain areas like e.g. the dorsal and ventral striata.
Present findings indicate a profile of slight spontaneous hyperactivity in DAT-i mice, in the form of a 2 h-longer wake-up peak and a wake-up attempt during rest (spontaneous rhythm in their home cages) together with reduced or overtly impaired patterns of flexibility (operant choice task in Skinner-box cages). It is well known that the ventral portion of the striatum subserves a motivational drive to flexibly change choice strategies [58, 60, 64]. After ventral striatum has provided motivation, such behavioural flexibility requires the recruitment of specific circuits within dorsal striatum [59, 62]; if these are compromised, the expression of habits is observed, whereby semi-automatic behaviour does not respond to devaluation of its outcomes [61, 63].
Our data are consistent with a functional impairment of the striatal complex, as a long-lasting consequence tapping onto the dopaminergic pathways and owing to the DAT immunization. Namely, the transient (but detectable) rise in circulating aAbs, generated by specific neuro-receptor fragments used as antigen, possibly together with a concomitantly permeable BBB (occasionally possible in stressful conditions; see ), might have generated persistent neuro-behavioural sequelae. Similarly, in a recent study , mice specifically immunized with fragments of glutamate receptors (GluR1) also demonstrated altered spontaneous activity and a stereotyped behavioural syndrome in response to a novelty-induced stress. A possible limitation in our study may be the lack of a further control group, immunized with a non-specific peptide. It is not easy however to figure out what this control peptide could be, since the exact sequence of the DAT antigens is currently protected information.
In conclusion, brain-directed auto-immune responses are likely to interfere with (or produce enduring impairments to) function of brain areas, with all the consequences thereof. More research effort is needed to verify potential psycho-immune alterations, via changes in the brain and behaviour of immunized mice. An altered psycho-immune modulation of neuro-receptors, like GluR1 or DAT, may play a role in vulnerability to neuro-psychiatric disorders like ADHD and OCD, as well as other impulse-control disorders like e.g. pathological gambling and sensation seeking.