In 1898, Camillo Golgi, an eminent Italian physician and pathologist, published a landmark paper on the structure of “nervous cells.” In addition to the organelle that still bears his name, the Golgi apparatus, he described “a delicate covering” surrounding neurons’ cell bodies and extending along their dendrites. That same year, another Italian researcher, Arturo Donaggio, observed that these coverings, now known as perineuronal nets (PNNs), had openings in them, through which, he correctly surmised, axon terminals from neighboring neurons make synapses.

Since then, however, PNNs have been largely neglected by the scientific community—especially after Santiago Ramón y Cajal, a fierce rival of Golgi (who would later share the Nobel Prize with him), dismissed them as a histological artifact. It wasn’t until the 1970s, thanks to the improvement of histological techniques and the development of immunohistochemistry, that researchers confirmed the existence of PNNs around some types of neurons...

Composed of extracellular matrix (ECM) molecules, PNNs form during postnatal development, marking the end of what’s known as the “critical period” of heightened brain plasticity. For a while after birth, the external environment has a profound effect on the wiring of neuronal circuits and, in turn, on the development of an organism’s skills and behaviors, such as language, sensory processing, and emotional traits. But during childhood and adolescence, neuronal networks become more fixed, allowing the individual to retain the acquired functions. Evidence gathered over the past 15 years suggests that PNNs contribute to this fixation in many brain areas, by stabilizing the existing contacts between neurons and repelling incoming axons.

Because limited neuronal plasticity underlies the irreversibility of many afflictions of the central nervous system (CNS), from stroke to spinal cord injury to neurodegenerative diseases, PNNs have been considered promising targets to enhance CNS repair. Moreover, they are increasingly recognized as important players in the regulation of memory processes.

Increasingly, researchers are turning to PNNs as potential targets to enhance plasticity for the treatment of various diseases, from amblyopia to neurodegenerative diseases to psychiatric disorders such as schizophrenia and addiction.

PNNs may also play a supportive role in the normal functioning of the CNS. These coatings have been repeatedly observed around highly active neurons, and researchers have proposed that the structures provide a buffered, negatively charged environment that controls the diffusion of ions such as sodium, potassium, and calcium, thus serving as a rapid cation exchanger to support neuronal activity.1 PNNs have also been shown to protect neurons from oxidative stress, as they limit the detrimental effect of excessive reactive oxygen species on neuronal function or survival. Indeed, enzymatic degradation of the PNNs renders neurons more susceptible to oxidative stress.2

A lot of progress has been made in the last two decades toward illuminating the structural and functional properties of PNNs, defining their roles in CNS plasticity, and developing methods to manipulate them to increase plasticity, memory, and CNS repair. Still, how exactly PNNs work in the brain, and which precise mechanisms underlie their remodeling in physiological or pathological conditions, are still open questions.

PNNs and the plastic brain

MISH MASH MESH: A cell from the spinal cord of a dog, stained by the Italian neurologist Carlo Besta in 1910 and republished in Trends in Neurosciences in 1998 (21:510-15). The PNN’s polygonal units are visible.COURTESY MARCO CELIOPNNs’ role in closing the critical period of brain plasticity is now well established. In 2010, for instance, my colleagues and I showed that knockout mice lacking a PNN component called link protein display a reduced formation of PNNs, and they maintain juvenile levels of plasticity throughout adulthood.3

Another example comes from rats with amblyopia, a neurodevelopmental disorder resulting from an imbalance between the neural signals coming from the two eyes during the critical period for visual development. Inputs from the right and left eyes compete when they first converge on neurons in the primary visual cortex, leading to a physiological and anatomical cortical representation of the relative inputs contributed by either eye. When one eye is deprived of visual input—for instance, due to a congenital cataract—individuals suffer a loss of cortical response to that eye and an overrepresentation of the input from the healthy eye, resulting in visual impairment. In adulthood, because the critical period is closed, vision will remain defective even if the cause of amblyopia is treated. The removal of PNNs in the visual cortex of adult rats, however, has proven effective in treating amblyopia.4

Under particular circumstances, such as enriched environmental stimulation, the adult brain regains certain levels of plasticity, and here too, PNNs appear to be important mediators. Adult amblyopic rats reared in a cage enriched with toys, ladders, and running wheels show a reduction of PNNs in the visual cortex and recover normal visual acuity after two to three weeks in this environment with no other treatment.5

In 2016, my colleagues and I documented similar links between PNNs and plasticity in the vestibular system, which detects head position and acceleration, stabilizes gaze and body posture, and contributes to self-motion perception. Mice that have suffered permanent damage to the inner ear vestibular receptors generally show severe deficits in their posture and balance, but will improve over time. This improvement comes along with an initial decrease of PNNs in the areas of the brain stem that regulate vestibular functions, followed by a complete restoration of the PNNs after posture and balance strengthen.6

Beyond these examples of recovering from sensory deficiencies, the adult brain exhibits plastic tendencies during normal learning, and recent evidence points to the role of PNNs in memory formation and retention. In rodents, explicit memory—information that can be consciously recollected—can be assessed by a novel object recognition test: when animals are exposed to familiar and novel objects, they spend more time exploring the novel one. This type of memory requires synaptic plasticity in a specific area of the cerebral cortex. Mutant mice lacking link proteins, and thus having reduced PNNs, exhibit a prolonged memory for familiar objects, and degradation of the PNN via the application of the bacterial enzyme chondroitinase ABC gives similar results. In both cases, the removal of PNNs facilitates the induction of synaptic plasticity.7

In addition to resisting memory formation, PNNs may also be to blame for blocking memory destruction. Whereas young individuals can permanently erase a fear memory by extinction training—a form of learning involving associating the fear-inducing stimulus with neutral scenarios—adults exhibit fear behaviors that are resistant to erasure. These behaviors depend on the amygdala, where PNNs are present in adult, but not young, animals. Interestingly, in adult mice, PNN degradation in the amygdala by chondroitinase ABC reopens a critical period during which fear memories can be fully erased by extinction training.8 In addition, PNNs in various cortical areas have recently been shown to be important for storage of fear memories, as their removal disrupts such memories.9,10

Currently, chondroitinase ABC is widely applied for removing PNNs in experimental animals, but it lacks specificity, causing a degradation of ECM molecules not only in PNNs but throughout CNS tissue. Researchers are looking for more subtle ways to manipulate the PNN in animal models in order to further understand their functions as well as to fine-tune neuronal plasticity. Additionally, while behavioral studies have clearly demonstrated the role of PNNs in mediating the plasticity of the brain, researchers still don’t have a good grasp on the molecular details of these processes. Recently, revelations about the composition of PNNs have begun to yield clues.

PNN structure and the control of plasticity

A number of ECM molecules are present at higher concentrations within the PNN than in the rest of the extracellular space. The sugar hyaluronan serves as the backbone of PNN structure. Bound to hyaluronan are chondroitin sulfate proteoglycans (CSPGs). This binding is stabilized by the link proteins. CSPGs are composed of a core protein and attached sugar chains. Different sulfation patterns in the CSPG sugar chains create specific binding sites for a wide variety of molecules and receptors, affecting CSPG function. (See illustration below.)

One mechanism by which PNNs control neuronal plasticity is the interaction between CSPGs and the homeoprotein Otx2. Homeoproteins are transcription factors that play major roles during embryonic development, controlling the organization of the vertebrate brain into distinct regions. Many homeoproteins also serve as paracrine signaling factors that shuttle between cells. In mice, experimentally reducing the capture of Otx2 by visual cortex neurons, which happens through binding to CSPGs, reduced PNN assembly, increased plasticity, and prompted the recovery of visual acuity in adult animals with amblyopia.11 And research last year demonstrated Otx2’s role in regulating PNNs’ influence on the experience-dependent formation of tonotopic maps, the spatial arrangement of neurons according to their sound frequency responses, in the primary auditory cortex and the acquisition of acoustic preferences (which is mediated by the medial prefrontal cortex).12

THE STRUCTURE OF THE PNN: The PNN is composed of chondroitin sulfate proteoglycans (CSPGs), which are made of a core protein (blue) flanked by a number of sugar chains (dark purple). CSPGs bind to hyaluronic acid (pink balls), which is secreted by membrane-bound enzymes. Link proteins (orange) stabilize the interaction between hyaluronic acid and CSPGs. Sema3A and Otx2 (pink pyramid and red ball, respectively) bind to the sugar chains of the CSPGs. Tenascin-R (green) acts as a cross-linking protein among several CSPGs, contributing to the macromolecular assembly of the PNN.
See full infographic: WEB | PDF
© 2018, LISA A. CLARK

Another potential mediator of PNN-controlled plasticity is the axon guidance molecule Semaphorin 3A, a CSPG-binding molecule that is highly concentrated in the PNNs of distinct neuronal populations in the mature brain. Recent experiments have shown that neurons cultured on PNN sugars grow shorter neurites and that this inhibition is enhanced by the presence of Semaphorin 3A.13 But the effect of Semaphorin 3A binding to the PNNs has yet to be determined. PNNs may also act through direct interactions with receptors for CSPGs on neurons, as occurs after a CNS injury, in which CSPGs upregulated in the injury site inhibit axon regrowth by binding to specific receptors. However, so far, no clear evidence exists about the presence of CSPG receptors on PNN-bearing neurons or synapses.Although the mechanisms that allow PNNs to influence neuronal plasticity remain unclear, the effects of that influence are well documented. Increasingly, researchers are turning to PNNs as potential targets to enhance plasticity for the treatment of various diseases, from amblyopia to neurodegenerative diseases to psychiatric disorders such as schizophrenia and addiction.

Targeting the PNN to treat disease

In accordance with findings showing a role for PNNs in memory, increasing evidence points to the involvement of PNNs in drug-associated memories. Environmental cues formerly associated with drug use (such as people or situations) can become strong triggers for drug-seeking behavior, contributing to the development of addiction. Therefore, disrupting these associations could aid in the treatment of addiction.

Genetic studies have identified several ECM- and PNN-regulating molecules as potential contributors to the etiology of autism.

Studies on rodents show that PNNs in the dorsal cerebellar cortex might play an important role in the formation and maintenance of cocaine-associated memories,14 and in the prefrontal cortex, PNNs are decreased after heroin self-administration but rapidly increased after reexposure to heroin-associated cues.15 Conversely, degrading PNNs in the prefrontal cortex reduces the acquisition and maintenance of cocaine memory in rats,16 while PNN degradation in the amygdala following an exposure to morphine, cocaine, or heroin inhibits relapse in the animals.17 On the whole, PNNs appear to be necessary for creating and/or maintaining drug-related memories, and thus may serve as targets for weakening memories that drive relapse.

Experimentally degrading PNNs has also shown promise in treating various forms of brain damage. Research in rodents has demonstrated that the degradation of PNNs by chondroitinase induces the formation of new axon branches and synapses, for example, and improves specific functions after stroke, trauma, and spinal cord injury.18 And in mouse models of Alzheimer’s disease, in which memory formation is compromised, a chondroitinase injection into the brain can successfully restore the ability to form new memories, even in the presence of diffuse neuronal dysfunction and cell death.19

Sometimes it’s not the presence of PNNs that is the problem, but rather aberrations in their structure. For instance, researchers have observed decreased densities of PNNs, or PNNs with degraded morphology, in brain areas responsible for complex cognitive functions, such as the frontal cortex and entorhinal cortex, in subjects with Alzheimer’s disease, suggesting that neurons with PNN alterations might be vulnerable to cell death.

BUNDLED UP: PNNs, visualized by immunostaining, wrap around cells in a mouse brain.DANIELA CARULLIAbnormal PNNs have been observed in the postmortem brains of schizophrenia patients—specifically, in regions involved in emotion-related learning and associative sensory information processing such as the amygdala, entorhinal cortex, and prefrontal cortex. Researchers have linked mutations in the genes encoding CSPGs, Semaphorin 3A, and other components of the normal ECM such as integrins and remodeling enzymes to schizophrenia risk. Loss of PNNs may render neurons more susceptible to the excitotoxic effects of oxidative stress believed to occur in schizophrenia.20

Several other psychiatric disorders have also been linked to PNN abnormalities. For example, genetic studies have identified several ECM- and PNN-regulating molecules, including Semaphorin 3A, the hyaluronan surface receptor CD44, and Otx2, as potential contributors to the etiology of autism. And postmortem studies of bipolar patients have shown a marked decrease in specific sugars or proteins associated with PNNs in the amygdala. Scientists have also suggested that a variant of the gene encoding the CSPG neurocan could be a risk factor for the disorder. Consistently, neurocan knockout mice show manic-like behaviors. Furthermore, increased levels of matrix-degrading enzymes appear in blood samples from subjects with major depression, bipolar disorder in a depressed state, schizophrenia, and autism.20

Yet another brain disorder that might be affected by the state of PNNs is epilepsy, which is characterized by abnormal patterns of neuronal activity that cause convulsions, unusual emotions and sensations, and a loss of consciousness. PNNs are decreased in animal models of epilepsy, putatively allowing for synaptic reorganization, such as occurs following seizures. Conversely, PNN abnormalities may contribute to a susceptibility to seizures. For example, increased epileptic activity has been found in mice lacking an enzyme that helps synthesize hyaluronan.20

Although many questions remain unanswered, research has clearly demonstrated that targeting PNNs holds great promise for the treatment of several brain diseases. So far, no compounds that target PNNs are drug candidates, as they are too general in their disruption, but scientists are working to identify specific PNN components to zero in on. While finding noninvasive ways to precisely target PNNs in specific areas of the human brain still represents a challenge, researchers are hopeful that this could be a promising avenue for a variety of therapies in the coming decades.

Daniela Carulli is an assistant professor in the Department of Neuroscience at the University of Turin in Italy and a member of the Neuroscience Institute Cavalieri Ottolenghi.


  1. W. Härtig et al., “Cortical neurons immunoreactive for the potassium channel Kv3.1b subunit are predominantly surrounded by perineuronal nets presumed as a buffering system for cations,” Brain Res, 842:15-29, 1999.
  2. J.-H. Cabungcal et al., “Perineuronal nets protect fast-spiking interneurons against oxidative stress,” PNAS, 110:9130-35, 2013.
  3. D. Carulli et al., “Animals lacking link protein have attenuated perineuronal nets and persistent plasticity,” Brain, 133:2331-47, 2010.
  4. T. Pizzorusso et al., “Structural and functional recovery from early monocular deprivation in adult rats,” PNAS, 103:8517-22, 2006.
  5. A. Sale et al., “Environmental enrichment in adulthood promotes amblyopia recovery through a reduction of intracortical inhibition,” Nat Neurosci, 10:679-81, 2007.
  6. A. Faralli et al., “Modifications of perineuronal nets and remodelling of excitatory and inhibitory afferents during vestibular compensation in the adult mouse,” Brain Struct Funct, 221:3193-209, 2016.
  7. C. Romberg et al., “Depletion of perineuronal nets enhances recognition memory and long-term depression in the perirhinal cortex,” J Neurosci, 33:7057-65, 2013.
  8. N. Gogolla et al., “Perineuronal nets protect fear memories from erasure,” Science, 325:1258-61, 2009.
  9. S.B. Banerjee et al., “Perineuronal nets in the adult sensory cortex are necessary for fear learning,” Neuron, 95:169-79, 2017.
  10. E.H. Thompson et al., “Removal of perineuronal nets disrupts recall of a remote fear memory,” PNAS, 115:607-12, 2018.
  11. M. Beurdeley et al., “Otx2 binding to perineuronal nets persistently regulates plasticity in the mature visual cortex,” J Neurosci, 32:9429-37, 2012.
  12. H.H.C. Lee et al., “Genetic Otx2 mis-localization delays critical period plasticity across brain regions,” Mol Psychiatry, 22:680-88, 2017.
  13. F. de Winter et al., “The chemorepulsive protein Semaphorin 3a and perineuronal net-mediated plasticity,” Neural Plast, 2016:3679545, 2016.
  14. M. Miquel et al., “Have we been ignoring the elephant in the room? Seven arguments for considering the cerebellum as part of addiction circuitry,” Neurosci Biobehav Rev, 60:1-11, 2016.
  15. M.C. Van den Oever et al., “Extracellular matrix plasticity and GABAergic inhibition of prefrontal cortex pyramidal cells facilitates relapse to heroin seeking,” Neuropsychopharmacology, 35:2120-33, 2010.
  16. M. Slaker et al., “Removal of perineuronal nets in the medial prefrontal cortex impairs the acquisition and reconsolidation of a cocaine-induced conditioned place preference memory,” J Neurosci, 35:4190-202, 2015.
  17. Y.-X. Xue et al., “Depletion of perineuronal nets in the amygdala to enhance the erasure of drug memories,” J Neurosci, 34:6647-58, 2014.
  18. D. Wang, J. Fawcett, “The perineuronal net and the control of CNS plasticity,” Cell Tissue Res, 349:147-60, 2012.
  19. S. Yang et al., “Perineuronal net digestion with chondroitinase restores memory in mice with tau pathology,” Exp Neurol, 265:48-58, 2015.
  20. H. Pantazopoulos, S. Berretta, “In sickness and in health: Perineuronal nets and synaptic plasticity in psychiatric disorders,” Neural Plast, 2016:9847696, 2016.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?