Cortical subpial aging-related tau astrogliopathy in older adult men who participated in youth amateur American-style football

Abstract

Astrocytes play essential roles in various functions including ionic homeostasis, energy metabolism, neurotransmission, and regulation of the blood-brain barrier. Researchers have reported that hyperphosphorylated tau in astrocytes can be identified in postmortem, particularly in individuals over the age of 60 years. Astrocytic p-tau has also been reported in association with chronic traumatic encephalopathy neuropathologic change (CTE-NC). This study investigates possible association between subpial aging-related tau astrogliopathy (ARTAG) and a history of playing high school American-style football. Postmortem brain tissue samples were obtained from 176 men (median age at death = 65 years; range = 50-96) from the Lieber Institute for Brain Development. There were 128 with no known history of participating in contact or collision sports and 48 (27.3%) who participated in football. Subpial ARTAG was identified in 23 cases (13.1%). Those who participated in football had a higher percentage with subpial ARTAG than those who did not participate in football (20.8% vs 10.2%), although this difference was not statistically significant (P =.061). The 2 groups did not differ in the percentages who had subpial ARTAG in a sulcal depth (10.4% of former football players and 7.8% of controls). More research is needed to understand possible associations between subpial ARTAG, aging, and CTE-NC.

INTRODUCTION

In the last century, chronic traumatic encephalopathy (CTE) was a heterogeneous neurological condition or disease that reflected frank and obvious chronic traumatic brain injury (TBI) in ultrahigh exposure boxers.^1–5 In this century, CTE has been conceptualized as microscopic neuropathology (chronic traumatic encephalopathy neuropathologic change [CTE-NC]^6–10) identified after death using hyperphosphorylated tau (p-tau) immunohistochemistry.^6,7,11 For the past 15 years, there has been interest and uncertainty regarding whether p-tau in astrocytes should be considered age-related, CTE-NC-related, or both.^6,7,11–21 The first consensus definition of CTE-NC included astrocytic p-tau as a required element: “The pathognomonic lesions consists of p-tau aggregates in neurons, astrocytes, and cell processes (italics added) around small vessels in an irregular pattern at the depths of the cortical sulci.”⁶ In the second consensus definition, astrocytic tau was permitted as part of the definition but was not required: “p-tau aggregates in neurons, with or without thorn-shaped astrocytes (italics added), at the depth of a cortical sulcus around a small blood vessel, deep in the parenchyma and not restricted to the subpial and superficial region of the sulcus.”⁷ Quotes reflecting this discussion and uncertainty are reprinted in Table 1.

Table 1.

Quotations from the literature reflecting interest and uncertainty regarding astrocytic tau as part of chronic traumatic encephalopathy neuropathology, aging, or both.

2009

“The neurofibrillary degeneration of CTE is distinguished from other tauopathies by preferential involvement of the superficial cortical layers, irregular patchy distribution in the frontal and temporal cortices, propensity for sulcal depths, prominent perivascular, periventricular, and subpial distribution, and marked accumulation of tau-immunoreactive astrocytes” (bolding added, from the abstract).12

2)   “Fibrillar astrocytic tangles are also common in the subpial and periventricular zones” (bolding added, page 726).12

2013

3)   …“the diagnosis of CTE was defined by the presence of the following criteria (Table 1 and Figs. 1 and 2): (i) perivascular foci of p-tau immunoreactive astrocytic tangles and neurofibrillary tangles; (ii) irregular cortical distribution of p-tau immunoreactive neurofibrillary tangles and astrocytic tangles with a predilection for the depth of cerebral sulci; (iii) clusters of subpial and periventricular astrocytic tangles in the cerebral cortex, diencephalon, basal ganglia and brainstem (bolding added); and (iv) neurofibrillary tangles in the cerebral cortex located preferentially in the superficial layers” (bolding added, page 45).11

2015

4)   “Mild forms of the disease affect primarily the neurons, although there may be focal subpial p-tau immunopositive astrocytes at the sulcal depths and occasional p-tau immunoreactive glia around the affected vessel (Figure 3). The pathological significance of isolated subpial p-tau astrocytes at the sulcal depths unaccompanied by perivascular NFT and NT in the underlying cortex is unknown, although subpial p-tau astrocytes at the depths of the cortical sulci is not a phenomenon found in normal aging and has not been found in the brains of subjects without a history of repetitive brain trauma” (bolding added, page 352).21

2016

5)   Supportive Neuropathological Feature of CTE: “p-Tau immunoreactive thorny astrocytes at the glial limitans most commonly found in the subpial and periventricular regions” (bolding added, page 81).6

6)   “ARTAG has features that overlap those of CTE, including the accentuation of tau pathology around small cerebral vessels and in subpial and periventricular areas. On the other hand, tau pathology, including neuronal and astroglial, in CTE is more abundant in the depths of the convexity cerebral sulci, especially in early stages, an aspect that has not been reported in tau astrogliopathy in the aging brain” (bolding added, page 94).6

2018

7)   “The overlapping histological features of ARTAG and CTE and the frequent concomitant findings of their features in the same individual raise the question of these two histological entities being on a spectrum of the same disorder” (bolding added, page 974).14

8)   “Subpial TSAs are frequent in CTE [44]. Blast injury has been also reported to be associated with tau positive astrocytes mainly in the frontal and parietal cortices [51]. It must be noted that in our series ARTAG was frequently not associated with the characteristic constellation of concomitant NFT pathology and ARTAG in the depth of the sulci as suggested for CTE [43]. However, our observations on two major patterns of subpial and WM ARTAG support the notion that the development of ARTAG and CTE type pathology shares common pathogenesis” (bolding added, page 14).22

2020

9)   “Subpial astrocytic p-tau, which often involves thorn-shaped astrocytes or TSA, is a form of ARTAG (3–6). Subpial p-tau pathology may be found in CTE as a supportive feature, but in isolation, it is nondiagnostic for CTE (2)” (bolding added, page 921).19

10)   “Prior to wide recognition of ARTAG as an age-related p-tau pathology (3–5), there was some ambiguity in early publications as to whether astrocytic p-tau pathology in subpial regions, temporal lobe white matter, and brainstem was a feature of CTE (8–11)” (page 923).19

11)   “In previously published articles, ARTAG neuropathology, such as astrocytic p-tau in subpial, subependymal, and perivascular locations, was described unambiguously as characteristic of CTE (7,8). Moreover, the consensus criteria for CTE neuropathology describe ARTAG spectrum pathology as a ‘supportive feature’ of CTE neuropathology (ie, ‘p-Tau-immunoreactive thorny astrocytes at the glial limitans most commonly found in the subpial and periventricular regions’)” (page 925).20

12)   “Our data demonstrate that in chronic traumatic encephalopathy neuropathologic change the astroglial component alone shows preferential distribution to the depths of cortical sulci. In contrast, the neuronal pathology of chronic traumatic encephalopathy neuropathologic change is distributed more uniformly from gyral crest to sulcal depth and echoes that of Alzheimer’s disease. These observations provide new insight into the neuropathological features of chronic traumatic encephalopathy that distinguish it from other tau pathologies and suggest that current diagnostic criteria should perhaps be reviewed and refined” (bolding added, from abstract).15

2021

13)   “The group also endorsed the following features of the pathognomonic lesion as necessary: p-tau aggregates in neurons, with or without concomitant p tau-immunoreactive TSA, at the depth of the sulcus around small blood vessels, in deeper cortical layers not restricted to subpial and superficial region. The panel unanimously confirmed that, based on case material available, purely astrocytic perivascular p-tau lesions (including subpial ARTAG) did not meet criteria for CTE. Furthermore, clusters of p-tau-immunoreactive astrocytes in the white matter of the frontal and temporal cortex, basal ganglia, lateral or medial brainstem were considered consistent with ARTAG (40–43), and not specific features of CTE” (bolding added, page 214).7

14)   “The distribution of tau pathology in CTE–NC and ARTAG–NC suggests some shared etiology [41 to 43]. Further, the presence of tau pathology in astrocytes in cases with CTE–NC suggests that ARTAG–NC could be a response of astrocytes to chronic mechanical stress or repetitive acute traumatic brain injury [1, 2, 43]” (bolding added, page 6).23

2022

15)   “Astroglial p-tau is an inconsistent supporting feature and not in itself diagnostic of CTE” (bolding added, from Abstract).16

16)   “This study demonstrates, through the use of a large group of brain donors neuropathologically diagnosed with CTE, across a wide range of age at death and CTE stage, without comorbid neuropathology, that the primary p-tau containing cell type in CTE is neuronal, and that the presence of astrocytic p-tau is a factor of age, not CTE severity or years of exposure to RHI” (bolding added, page 780).16

17)   “Our observations demonstrate that the presence of both neuronal and astroglial tau pathologies facilitates detection of CTE-NC, with its detection less consistent when neuronal tau pathology alone is visible. The combination of both glial and neuronal pathologies, therefore, may be required for detection of CTE-NC” (bolding added, from abstract).17

2024

18)   “Although not part of the diagnostic criteria of CTE [13], there is a close association between ARTAG and CTE [53] and they may represent a spectrum of the same disorder with ARTAG variably reported in 25%–100% of CTE cases [14,18,29,32,33]” (bolding added, page 14).24

2025

19)   “Intriguingly, while subpial TSA at sulcal depths were occasionally observed in aged controls with (3.6%) and without (2.8%) NDD, this pathology was considerably more common following RHI/TBI (42.2%; p < 0.001). These findings support history of RHI and TBI as independent risk factors for the development of thorn-shaped tau astrogliopathy, over and above ARTAG observed in aging and wider neurodegenerative disease. Moreover, trauma may be associated with TSA within specific distributions, including the subpial region of the cortical sulcal depths. The clinical significance of these observations will be important to determine” (bolding added, from Abstract).18

The quotations are from published articles; bolding is added for emphasis. Some entries contain numbered citations, but those numbered citations are not included in our reference list. Only the article from which the quote is obtained is in the present reference list.

Abbreviations: ARTAG, aging-related tau astrogliopathy; ARTAG-NC, aging-related tau astrogliopathy neuropathologic change; CTE, chronic traumatic encephalopathy; CTE-NC, chronic traumatic encephalopathy neuropathologic change; NDD, neurodegenerative disease; NFT, neurofibrillary tangle; NT, neuropil threads; RHI, repetitive head impacts; TBI, traumatic brain injury; TSA, thorn-shaped astrocytes; WM, white matter.

Astrocytes are glial cells in the brain that play essential roles in various functions, including ionic homeostasis, energy metabolism, neurotransmission, and the regulation of the blood-brain barrier.^25–30 P-tau is identified in tissues by immunohistochemistry and has been shown to accumulate within astrocytes in association with aging and with neurodegenerative diseases.^31–36 The identification of p-tau in astrocytes in Alzheimer disease dates to the late 1980s.³⁵ In 1995, Ikeda et al. coined the term “thorn-shaped astrocytes” (TSA) as a “secondary” phenomenon in the aged brain, and also speculated about a role in traumatic brain injury, stating: “As noted in boxer’s brain and other diseases, TSA tend to form in the cleft of the gyrus, where damage due to cerebral edema would be more prominent than that at the crest of the gyrus. It can be deduced that the repeated mechanical damage probably provokes frequent cerebral edema and would result in gliosis in this region in dementia pugilistica.”³² A more detailed analysis of astrocytic p-tau was published in 2004, in a study of 100 brains, from people between the ages of 42 and 97 years. In this article, TSA were described, preferentially located in periventricular, subependymal, and subpial areas of the medial temporal lobe.³¹ These authors reported that TSA were absent in brains from donors under the age of 60 years but they were present in approximately 50% of those between the ages of 75 and 98 years.³¹ Botez described “bush-like” astrocytic p-tau in argyrophilic grain disease in 1999, which would later be termed “granular/fuzzy astrocytes” (GFA) in 2016, as part of the umbrella term “aging-related tau astrogliopathy” (ARTAG),¹³ encompassing both TSA and GFA. The ARTAG acronym was introduced in this article as part of a harmonized strategy for coding aging-related astrocytic tau.

Aging-related tau astrogliopathy is conceptualized as a spectrum of astroglial tau morphologies found in locations that are subpial, subependymal, and perivascular in gray and white matter.¹³ Subpial, subependymal, perivascular, and white matter ARTAG typically has thorn-shaped morphology (Figure 1), whereas gray matter ARTAG normally has the granular/fuzzy morphology. Less common morphologies include tufted astrocytes, astrocytic plaques, ramified astrocytes, and globular glial inclusions, which are generally associated with primary 3-repeat or 4-repeat tauopathies^32,33 (Figure 2).

Figure 1.

Examples of thorn-shaped astrocytes (TSA) and granular/fuzzy astrocytes (GFA). Thorn-shaped astrocytes (TSA) (left image) and granular/fuzzy astrocytes (GFA) (right image) are descriptive terms for appearances of immunolabeling by p-tau immunohistochemistry. Both are included in the broader umbrella term aging-related tau astrogliopathy (ARTAG). As described by Kovacs et al.,¹³ TSA demonstrate p-tau labeling in astrocytic perikarya with extension into the proximal parts of the astrocytic processes, with p-tau also in astrocytic endfeet at the glia limitans around blood vessels and at the pial surface. The processes are thick and short, reminiscent of thorns. They may also occur in the white matter and less often in the gray matter. Granular/fuzzy astrocytes appear typically in gray matter as finely granular immunoreactivity of branching processes with dense labeling of the perinuclear soma. White scale bars in lower left corners = 200 μm.

Figure 2.

Primary 3-repeat (eg, Pick disease) and 4-repeat tauopathies (eg, corticobasal degeneration, progressive supranuclear palsy, globular glial tauopathy) show p-tau immunolabeling that is relatively specific for those pathological subtypes. (A) Tufted astrocyte in progressive supranuclear palsy, with dense p-tau-positive tufts in proximal astrocytic processes. (B) Astrocytic plaque consisting of p-tau-positive stubby dilatations of distal astrocytic processes in a plaque-like appearance in a case of corticobasal degeneration. (C) Ramified astrocytes in a case of Pick disease, which shows p-tau labeling mostly in perikarya and cell processes in a somewhat asymmetric fashion. (D) Globular astroglial inclusions in a case of globular glial tauopathy. Images were obtained from cases that were not part of this study. Black scale bars in lower right corners = 100 μm.

Aging-related tau astrogliopathy in different forms and locations often coexist.¹³ ARTAG may also co-occur with neurodegenerative diseases, including Alzheimer disease and primary tauopathies such as progressive supranuclear palsy.²² The first consensus criteria for defining CTE-NC were published in 2016,⁶ the same year that the harmonized strategy for coding ARTAG was published.¹³ The authors agreed that astrocytic tau was part of the supportive diagnostic criteria for CTE-NC. Specifically, they considered astrocytic tau a supportive neuropathological feature of CTE-NC that was described as follows: “p-Tau immunoreactive thorny astrocytes at the glial limitans most commonly found in the subpial and periventricular regions” (page 81).⁶ Therefore, in 2016, subpial astrocytic tau was defined as both ARTAG and a supportive diagnostic feature of CTE-NC. See Figures 3 and 4 for other examples of images of subpial ARTAG.

Figure 3.

Subpial aging-related tau astrogliopathy in the parietal lobe. Immunohistochemical stain for p-tau (AT8) at medium magnification shows subpial thorn-shaped astrocytes at a sulcal depth of the inferior parietal cortex. White scale bar in lower left corner = 200 μm.

Figure 4.

Subpial aging-related tau astrogliopathy. Immunohistochemical stain for p-tau (AT8) at low magnification shows subpial thorn-shaped astrocytes at a sulcal depth of the frontal cortex. White scale bar in lower left corner = 200 μm.

The second consensus criteria for defining CTE-NC were published in 2021.⁷ The consensus panel attempted to clearly differentiate ARTAG from CTE-NC. They wrote: “The panel unanimously confirmed that, based on case material available, purely astrocytic perivascular p-tau lesions (including subpial ARTAG) did not meet criteria for CTE. Furthermore, clusters of p-tau-immunoreactive astrocytes in the white matter of the frontal and temporal cortex, basal ganglia, lateral or medial brainstem were considered consistent with ARTAG (40-43), and not specific features of CTE (page 214).”⁷

A recent study was designed to determine whether exposure to TBI among people from the general population or exposure to repetitive head impacts in former contact and collision sport athletes is associated with p-tau astrogliopathy.¹⁸ They examined postmortem tissue samples from 556 people selected to include cases who (1) had a history of moderate or severe TBI (n = 77), (2) had a history of combat, collision, or contact sports participation (n = 45), or (3) were controls with (n = 397) or without (n = 37) neuropathologically confirmed neurodegenerative disease. They reported that subpial TSA at sulcal depths were present in 19 of 45 former athletes (42.2%), compared to only 3.6% of controls with neurodegenerative diseases, only 2.7% of controls who did not have a neurodegenerative disease, and 9.2% of those with a personal history of moderate or severe TBI.¹⁸ The authors discussed the importance of subpial TSA as being associated with a personal history of participation in combat, collision, and contact sports.¹⁸ In contrast, other authors have emphasized that this subpial pathology is age related and not associated with repetitive head impacts from sports.¹⁶

In summary, researchers have reported that p-tau in astrocytes is primarily age related but that it may also co-occur more often in individuals with a history of participation in contact, collision, and combat sports. Noteworthy is that subpial ARTAG and particularly subpial ARTAG at a sulcal depth is often depicted in illustrations of CTE-NC.^{6,11,21,37,38} One study suggested that subpial astrocytic p-tau preferentially concentrates at sulcal depths in association with CTE-NC.¹⁵ It has also been suggested by computational modeling that mechanical strain might be concentrated at sulcal depths.³⁹ Assuming that subpial p-tau in astrocytes relates to aging and may be associated with participation in contact and collision sports, it is reasonable to hypothesize that former amateur football players might be more likely to have cortical subpial ARTAG than men who did not play football. The purpose of this study was to screen for cortical subpial ARTAG in the brains of men and determine if there is an association between subpial ARTAG and a personal history of playing amateur football.

METHODS

Brain donation program

The slides of postmortem brain tissue for this study were obtained through a collaboration between the Lieber Institute for Brain Development (LIBD) and the Department of Pathology at Western Michigan University Homer Stryker MD School of Medicine (WMed) (WCG protocol #1126332). During the data collection period (2016-2020), the WMed Department of Pathology conducted death investigations and forensic autopsies in 13 counties across Western Michigan. Some cases were referred during organ donation through the Gift of Life Michigan. The donors were men aged 50 years and older; immunohistochemistry results were obtained for these cases. Only men were studied because the aim was to explore the potential association between a personal history of participating in American football during youth and p-tau astrogliopathy. To be included in the study, information from next of kin relating to earlier in life sports participation must have been available. Audiotaped informed consent from the legal next of kin was obtained for each case at the time of autopsy.

History of participation in combat and collision sports, such as boxing, football, and rugby, was obtained from the next of kin; this variable was coded as “yes” or “no.” The question asked during the interview was about a, “History of engagement in sports with a high risk of repetitive brain injury, including boxing, football, mixed martial arts, lacrosse, rugby, or auto racing?” Forty-eight men participated in high school football; 4 of them also played at the college level. There were no other sports with a sufficient number of subjects to be included as subgroups. The focus of this study was earlier-in-life participation in amateur football, mostly high school (with 4 subjects playing at the college level). Nonetheless, for completeness, supplementary analyses were conducted that included the 10 subjects who participated in other sports.

The next of kin were asked an open-ended question about: “History of head trauma?” The interviewer then took notes describing characteristics of the injury. Some information was also available in medical records. A broad range of injury mechanisms (eg, sports, falls from a height, and car accidents), types, and severities were reported (eg, concussion, skull fractures, subdural hematomas, and prolonged posttraumatic unconsciousness or coma). For most of the decedents, there was insufficient information to classify TBI severity. Therefore, the variable was coded as binary. There was missing data on TBI for 8 decedents (4.5%). This variable was used for supplementary analyses.

Brain sampling and immunohistochemistry

Per the LIBD protocol, sampling of 7 brain regions was obtained for histopathology as previously described.⁴⁰ These included the right middle frontal gyrus (Brodmann area 8), the right superior/middle temporal cortex (Brodmann area 22), the right inferior parietal lobule (Brodmann area 40), the right occipital calcarine cortex (Brodmann area 17), hippocampus at the level of the lateral geniculate nucleus (including the contiguous parahippocampal gyrus and inferior temporal cortex in most cases), midbrain, and cerebellar hemisphere with dentate nucleus. For the cortical samples, research technicians were instructed to include sulcal depths in each sample. All samples were stained with hematoxylin and eosin (H&E) for microscopic examination. P-tau immunohistochemistry (using monoclonal antibody AT8) (hippocampus and cortical samples) was used. Slides were scanned into virtual microscopy (Aperio ImageScope Pathology View software), as described.⁴¹ We recently published a study illustrating our methodology for examining postmortem brain tissue from the LIBD brain donation repository.⁴² P-tau and Aβ are semiquantitated using modifications of Braak⁴³ and CERAD (Consortium to Establish a Registry for Alzheimer’s Disease).⁴⁴ Modified Braak and CERAD ratings are provided for descriptive purposes.

Cortical subpial ARTAG was assessed according to the criteria described by Kovacs et al.¹³ Each immunostained slide for p-tau (AT8) was examined for the presence or absence of subpial TSA. Thorn-shaped astrocytes are typically observed in subpial, subependymal, perivascular, and white matter locations, as well as occasionally in gray matter.¹³ We also noted whether TSA were observed at a sulcal depth. We limited the study to cortical subpial ARTAG because a recent study has emphasized its importance in research with former athletes.¹⁸ There has been an ongoing debate as to whether subpial ARTAG is etiologically and mechanistically associated with low-level repetitive neurotrauma experienced during participation in amateur and professional collision and contact sports.^{14–16,18,45} Cases were coded by region: hippocampus with adjacent parahippocampal and fusiform gyrus, superior/middle temporal gyrus, middle frontal gyrus, inferior parietal lobule, and occipital calcarine cortex and presence or absence of involvement of a sulcal depth.

Statistical analyses

Cohen’s Kappas and percentages in agreement were computed as measures of interrater agreement. A conventional interpretation for Kappa is as follows: values ≤ 0 as no agreement and 0.01-0.20 as none to slight, 0.21-0.40 as fair, 0.41-0.60 as moderate, 0.61-0.80 as substantial, and 0.81-1.00 as almost perfect agreement. χ² tests were used to compare proportions and independent t-tests were used to compare groups on age. An odds ratio (OR) above 1.00 and a 95% CI higher or lower than 1.00 indicated the predictor was significantly associated with greater odds of having preferential subpial ARTAG. In contrast, an OR below 1.00 with a 95% CI not including 1.00 indicated the predictor was associated with reduced odds of having preferential subpial ARTAG. All statistical analyses were performed using IBM SPSS Statistics 26.

RESULTS

Complete clinical information, tissue samples, and immunohistochemical stains were available for 186 men. Their median age at the time of death was 65 years (interquartile range [IQR]=57-75, range = 50-96). Their modified Braak ratings were: B1 = 105 (56.5%), B2 = 58 (31.2%), B3 = 15 (8.1%), and missing = 8 (4.3%). Their modified CERAD ratings were: C0 = 99 (53.2%), C1 = 35 (18.8%), C2 = 36 (19.4%), C3 = 15 (8.1%), and missing = 1 (0.5%). Of those, 10 were excluded from all but the supplementary analyses because they participated in a contact or collision sport other than football. The final sample comparing those who participated in football to the control subjects included 176 donors. Their median age at the time of death was 65 years (IQR = 57-75, range = 50-96). Their race and ethnicity were as follows: 98.3% White, 98.3% Black or African American, and 1.1% Hispanic. Their manner of death was: 29.5% natural, 31.8% accident, 35.8% suicide, and 2.8% undetermined. Their modified Braak ratings were: B1 = 99 (56.3%), B2 = 54 (30.7%), B3 = 15 (8.5%), and missing = 8 (4.5%). Their modified CERAD ratings were: C0 = 91 (51.7%), C1 = 34 (19.3%), C2 = 35 (19.9%), C3 = 15 (8.5%), and missing = 1 (0.6%). Among them, 128 men (72.7%) had no known history of participation in contact or collision sports; 48 men (27.3%) had a history of playing amateur football. Of the total sample, there were 46 who had a history of TBI (26.1%). The 2 groups did not differ significantly in age (No Sports mean [M]=66.78, SD=11.17; Football M = 65.63, SD = 11.82; t = 0.602, P = .548, Cohen’s d = 0.10).

Interrater agreement

Two raters independently coded the cases, and the percentage agreements and Cohen’s Kappa values were calculated. The Kappas ranged from fair to almost perfect agreement (Table 2). Kappas were sometimes low because most of the ratings agreed that the finding was not present, with a small number of disagreements regarding the presence of the finding. Initial percentage agreement ratings ranged from 96.1% to 100% (Table 2). After their initial blinded ratings, the 2 raters collaborated to reach consensus on the individual ratings that differed and these final ratings were used for the analyses.

Table 2.

Initial blinded neuropathology ratings from 2 raters.

Variable	n	Percent agreement	Kappa
Subpial ARTAG	181	96.1	.807
Subpial ARTAG frontal	181	98.9	.745
Subpial ARTAG temporal	181	98.9	.828
Subpial ARTAG parietal	181	97.2	.430
Subpial ARTAG occipital	180	98.9	NA
Subpial ARTAG inferior temporal	181	97.8	.738

Inferior temporal refers to cortical tissue in the parahippocampal and fusiform gyrus that is included with the sample of hippocampus. A small number of tissue samples from the 186 cases were not available to the second rater, and so those samples were coded by a single rater. Not applicable (NA) because the measure of agreement (Kappa) was not calculated because one rater recorded ‘no’ for the 180 cases that were rated by both authors.

Abbreviations: ARTAG, aging-related tau astrogliopathy; CERAD, Consortium to Establish a Registry for Alzheimer’s Disease.

Presence of cortical subpial ARTAG

Cortical subpial ARTAG was identified in 23 cases (13.1%). The median age of those with subpial ARTAG was 67 years, the IQR for their age was 58-71 years, and the full range of their age was 51-95 years. Those with cortical subpial ARTAG (M age = 68.26 years, SD = 13.28) were not significantly older than those who did not have subpial ARTAG (M age = 66.20 years, SD = 11.03; t(1,174)=−0.814, P=.417, Cohen’s d = −0.18, small effect size). The presence of cortical subpial ARTAG by age group was: 50-59 years (n = 54)=13.0%, 60-69 years (n = 60)=11.7%, 70-79 years (n = 37)=8.1%, and 80+ (n = 25)=24.0%. Former football players with cortical subpial ARTAG (n = 10; M age = 65.30 years, SD = 12.85) were similar in age to former football players who did not have subpial ARTAG (n = 38, M age = 65.71 years, SD = 11.72; P =.923, d=0.034). Control subjects with cortical subpial ARTAG (n = 13) were a little older (M age = 70.54 years, SD = 13.67), on average, than those without subpial ARTAG (n = 115, M age = 66.36 years, SD = 10.85, P =.202, d = −0.37); this difference was not statistically significant. Those former football players with cortical subpial ARTAG (n = 10, M age = 65.30 years, SD = 12.85) were younger, on average, than controls who had cortical subpial ARTAG (n = 13, M age = 70.54 years, SD = 13.67; P =0.361, d = −0.44), although this difference was not statistically significant and the sample sizes for this comparison were very small.

The percentages of the sample that had subpial ARTAG in specific locations are presented in Table 3. Subpial sulcal depth ARTAG was present in 15/23 cases with subpial ARTAG (ie, 65.2%). Those who participated in football had a higher percentage with cortical subpial ARTAG than those who did not participate in football (20.8% vs 10.2%), although this difference was not significant (P =0.061). The 2 groups did not differ in the percentages who had cortical subpial ARTAG in a sulcal depth (Table 3). There were 114 men whose age of death was under the age of 70 years (32 who participated in youth football and 82 who did not). Of those under 70, men who participated in football had a higher percentage with subpial ARTAG than those who did not participate in football (21.9% vs 8.5%), although this difference was not significant (P =.051; χ²=3.80, OR = 3.00, 95% CI, 0.96-9.39).

Table 3.

Subpial aging-related tau astrogliopathy in the total sample and stratified by participation in football.

	Total		No Sports		Football				95% CI
Variable	f	%	f	%	f	%	P	OR	Lower	Upper
Subpial ARTAG	23	13.1	13	10.2	10	20.8	.061	2.32	0.94	5.74
Subpial ARTAG at sulcal depth	15	8.5	10	7.8	5	10.4	.582	1.37	0.44	4.24
Subpial ARTAG frontal	4	2.3	3	2.3	1	2.1	.918	0.89	0.09	8.74
Subpial ARTAG temporal	7	4.0	5	3.9	2	4.2	.937	1.07	0.20	5.71
Subpial ARTAG parietal	6	3.4	4	3.1	2	4.2	.734	1.35	0.24	7.61
Subpial ARTAG occipital	0	0	0	0	0	0	—	—	—	—
Subpial ARTAG inferior temporal	12	6.8	6	4.7	6	12.5	.067	2.91	0.89	9.50

Inferior temporal refers to cortical tissue in the parahippocampal and fusiform gyrus that is included with the sample of hippocampus.

Abbreviations: CERAD, Consortium to Establish a Registry for Alzheimer’s Disease; OR, odds ratio; f, frequency; %, percentage.

Supplementary analyses

Supplementary analyses were conducted on the total sample, with the 10 men who participated in other sports during their youth, including 3 who participated in hockey, 3 who participated in wrestling, 1 who participated in kickboxing, 1 who participated in boxing, and 2 who participated in auto racing (with multiple suspected concussions). There was no significant difference in the percentages with cortical subpial ARTAG in the combined group who participated in collision, contact, or combat sports (17.2%) compared to the control group (10.2%) who did not participate in these sports (P =.174; χ²=1.85, OR = 1.84, 95% CI, 0.76-4.49). There also was no significant difference in the percentages with cortical subpial sulcal depth ARTAG in the combined group who participated in collision, contact, or combat sports (8.6%) compared to the control group (7.8%) who did not participate in these sports (P =.851; χ²=0.04, OR = 1.11, 95% CI, 0.36-3.42).

Additional supplementary analyses were conducted on the total sample stratified by TBI history. There was no significant difference in the percentages with cortical subpial ARTAG between the group with a history of TBI (12.0%) compared to the group without a history of TBI (13.3%; P =.819; χ²=0.052, OR = 0.89, 95% CI, 0.33-2.41). There was also no significant difference in the percentages with cortical subpial sulcal depth ARTAG in the group with a history of TBI (10.0%) compared to the group without a history of TBI (7.8%; P =.637; χ²=0.223, OR = 1.31, 95% CI, 0.43-4.05). Chronic sequelae of TBI in the form of remote contusions were noted in 3 decedents. None had subpial ARTAG.

DISCUSSION

By definition, ARTAG is associated with age and subpial TSA are one type of ARTAG.^13,31 The path to understanding the possible association between participation in American-style football early in life and the presence of subpial TSA in the depths of cortical sulci after death has been tortuous. This is illustrated in quotes from articles published between 2009 and 2025 (Table 1). In the present study, subpial ARTAG was present in 13.1% of the total sample, including 20.8% of those who participated in high school football and 10.2% of those with no known participation in combat, collision, or contact sports. Although twice as common in men who participated in football, this finding did not reach conventional statistical significance (P =.061). Nonetheless, this finding suggests a possible basic association between participation in football and later-in-life subpial ARTAG. However, there was no clear association between subpial ARTAG in a sulcal depth and participation in football, with 10.4% of former football players having this pathology and 7.8% of the control subjects having this finding.

The etiology of subpial ARTAG is unknown.²² Astrocytes form borders (eg, glia limitans) that separate neural from nonneural tissue along perivascular spaces and meninges⁴⁶; these borders can be subpial in the depths of sulci and vulnerable to mechanical stress. Astrocytes have also been shown to induce blood-brain barrier properties in endothelial cells.⁴⁷ An association between subpial ARTAG and dysfunction of the blood-brain barrier is supported by its association with greater amounts of astroglial expression of connexin-43 (a protein expressed by astrocytes at the blood-brain barrier) and aquaporin-4 (a water-channel protein expressed in the foot processes of glial cells surrounding capillaries).⁴⁸ The mechanisms and temporal kinetics by which connexin membrane channel proteins and the formation of ARTAG are neuroprotective, neurotoxic, and/or neurodegenerative are not well understood.^48,49

Traumatic brain injury can induce blood-brain barrier breakdown that might contribute to astrogliosis.^23,50 In a swine model study of TBI, researchers reported blood-brain barrier disruption and acute perivascular uptake of blood-borne proteins in both neurons and astrocytes.⁵¹ However, in a postmortem study that examined the frequency of ARTAG in community-dwelling older adults with and without a lifetime history of moderate-to-severe TBI, there was no association between ARTAG and TBI.⁴⁵ Regarding subpial ARTAG specifically, 14.2% of those with a history of moderate-to-severe TBI had neocortical subpial ARTAG compared to 15.2% of those who did not have the brain injury history.⁴⁵ In the present study, we also found no association between TBI and cortical subpial ARTAG or subpial ARTAG in a sulcal depth.

Shively et al. discussed 2 mechanistic hypotheses for how p-tau might accumulate in astrocytes and neurons in the depths of a sulcus in a pattern resembling CTE-NC.⁵² First, tau phosphorylation causes microtubule disassembly in neurons, which might lead to aggregation of p-tau over time. Second, the depths of sulci are particularly vulnerable to the effects of mechanical stress rendering them susceptible to p-tau deposition.⁵² Following axonal injury, Shively et al. hypothesized that the sulcal depth serves as an exodus for tau clearance possibly via the “glymphatic” pathway of interstitial solute clearance through local diffusion and fluid bulk flow through channels of astrocytic endfeet into perivenous spaces.⁵²

In their review of the literature, Bachstetter et al. suggested that these 2 hypothetical models are not mutually exclusive and that both could underlie the formation of ARTAG.⁵³ First, neuronal injury results in the release of tau protein ‘seeds’ that accumulate in astrocytes. This accumulation occurs at least in part because the protein seeds are not being cleared from the brain via glymphatic pathways and/or the seeds accumulate during the process of being cleared from the brain according to the hypothetical model.⁵³ This assumes that the origins of ARTAG primarily arise from neuronal injury. Their second hypothetical model proposes that brain injury in combination with aging leads to a reactive transformation of astrocytes promoting ARTAG. The second model assumes that astrocytes have the capacity to generate their own tau.⁵³ Kovacs et al. alternatively hypothesized that astroglia might phagocytose p-tau, derived from the endings of projecting neurons.²²

Prior studies of former football players from the LIBD brain repository

Three prior studies have utilized this sample of brain donors from the LIBD to investigate the clinical and neuropathological aspects of later-life brain health in former amateur football players.^40,54 The first study examined whether there was an association between participation in amateur football and suicide as a manner of death.⁵⁴ There was no association between participation in amateur football and completed suicide. The men who participated in amateur football were significantly less likely to have a lifetime history of a suicide attempt. Those men who experienced mood disorders or substance use disorders during life and those who had a family history of suicide were more likely to have completed suicide.⁵⁴ The second study examined all cases for CTE-NC; the authors did not identify a single definitive case of CTE-NC in former amateur football players or control subjects.⁴⁰ In that study, 10 cases (5.4% of the sample) had some features or characteristics of CTE-NC, including 8.3% of those who participated in amateur football and 3.9% of those who did not participate in contact or collision sports.⁴⁰ The third study examined an association between preferential p-tau in the hippocampal cornu ammonis 2 subfield (CA2) region and a history of playing amateur football.⁵⁵ That study was done because past researchers reported that p-tau aggregates in the CA2 region of the hippocampus in association with aging (ie, Braak stage II),^56,57 preferentially as part of primary age-related tauopathy,⁵⁸ in association with early Alzheimer disease,⁵⁹ and preferentially as part of CTE-NC.⁶ Preferential CA2 p-tau was present in 29.9% of the total sample, and it was associated with older age at the time of death and the postmortem presence of amyloid-β (Aβ). There was no significant association between participation in amateur football and preferential CA2 p-tau identified after death.⁵⁵

Limitations

This study has important limitations. The exposure variable, participation in amateur football, was derived from an interview with the next of kin and it was binary (ie, yes or no). Participation in amateur football was not verified with school yearbooks or other sources and we could not quantify the amount of exposure in years. This study has sampling bias because the LIBD research program specializes in brain donations from people who suffered from psychiatric and neuropsychiatric disorders. That said, the LIBD research program does not specialize in collecting tissue from former athletes, so brain donations are pursued irrespective of the donors’ history of sports participation. This sample is from a single medical examiner’s office serving western Michigan and northern Indiana and it is not representative of men from the general population. The sample is also not representative of men who played high school football because it included virtually no men who were Black or African American. The LIBD research program protocol involves neuropathological screening of donors aged 50 years and older; therefore, the results should not be considered generalizable to women or men under the age of 50 years. Finally, the sampling protocol is limited to 5 tissue samples from the frontal, parietal, occipital, temporal, and hippocampus with adjacent inferior temporal cortex (ie, neocortical tissue in the parahippocampal and fusiform gyrus) regions. It likely underestimates the frequency of subpial ARTAG in the donors. The harmonized strategy for the study and coding of ARTAG¹³ involves more samples and additional brain regions (eg, amygdala, subcortical structures, brainstem). Using this approach would likely have resulted in higher percentages being identified as having ARTAG.

Conclusions and clinical implications

There is an ongoing debate about whether subpial ARTAG is etiologically and mechanistically associated with: (1) repetitive head impacts from combat, collision, and contact sports, (2) TBI, and/or (3) CTE-NC.^{14–16,18,45} One study on CTE-NC asserts “that the primary p-tau containing cell type in CTE is neuronal and that the presence of astrocytic p-tau is a factor of age, not CTE severity or years of exposure to RHI” (page 780).¹⁶ In contrast, another recent study reported that TSAs at sulcal depths are strongly associated with a personal history of participation in contact, collision, and combat sports.¹⁸ Amplifying the debate on etiology, the authors adopted different terminology to discuss subpial ARTAG in former contact and collision sport athletes, that is, dropping the words “aging-related” and the acronym ARTAG and referring to it as “tau astrogliopathy.”¹⁸

It is known that: (1) subpial ARTAG and CTE-NC co-occur,^6,14,15,17 (2) subpial ARTAG was considered part of CTE-NC in the past,^6,11,12,21 (3) both subpial ARTAG and CTE-NC can be present in postmortem brain tissue from individuals who experienced frontal leucotomy,⁵² space-occupying arachnoid cysts,⁵³ a resection of a cerebral arteriovenous malformation,⁶⁰ glioblastoma,⁶⁰ and cerebral infarcts⁶⁰ during life, and (4) subpial ARTAG is associated with aging in the general population.²⁴ Our findings suggest the possibility of greater subpial ARTAG in former high school football players compared to men who did not play football. However, the difference did not reach statistical significance. A larger study is needed to determine whether there is a clear and basic association between amateur football participation and greater subpial ARTAG. The mechanisms and temporal kinetics driving the aggregation of subpial ARTAG and the extent to which it is reactive, neuroprotective, neurotoxic, and/or neurodegenerative are not well understood and require future study.

Little is known about the clinical correlates or implications of ARTAG in general or of subpial ARTAG specifically. In the paper introducing the harmonized strategy for the coding ARTAG the authors wrote: “Despite its high prevalence, there is a lack of consensus on whether these astroglial tau pathologies in the elderly are clinically relevant, even as a concomitant pathology that might lower an individual’s threshold for the development of clinical symptoms” (page 90).¹³ Studies have reported that ARTAG is not associated with dementia.^61,62 However, in one study of people over the age of 90 years at the time of death, cortical ARTAG was associated with dementia.²⁴ The authors emphasized that multiple pathologies occurring together that are separate from Alzheimer disease pathology, such as cerebrovascular disease, ARTAG, hippocampal sclerosis, TDP-43, and Lewy body pathology can act as multiple insults on the brain and that in combination they are sufficient for correlating with dementia.²⁴ In their recent study of subpial ARTAG, Arena et al. emphasized the importance of determining the clinical significance, if any, of their findings.¹⁸ To our knowledge, no studies have shown a clear association between subpial ARTAG and any neurological, neuropsychological, or psychiatric problem during life.

Conflicts of interest

G.L.I. serves or has served as a scientific advisor for NanoDX, Sway Operations, LLC, and Highmark, Inc. He has a clinical and consulting practice in forensic neuropsychology, including expert testimony, involving individuals who have sustained mild TBIs (including former athletes), and on the topic of suicide. He has received research funding from several test publishing companies, including ImPACT Applications, Inc., CNS Vital Signs, and Psychological Assessment Resources (PAR, Inc.). He has received research funding as a principal investigator from the National Football League, and subcontract grant funding as a collaborator from the Harvard Integrated Program to Protect and Improve the Health of National Football League Players Association Members. He acknowledges unrestricted philanthropic support from ImPACT Applications, Inc., the Mooney-Reed Charitable Foundation, and the National Rugby League. These entities were not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the decision to submit it for publication. P.J. serves as a consultant neuropathologist for Novocure GmbH. A.D.-S. has no disclosures or potential conflicts of interest to declare. T.M.H. has no disclosures or potential conflicts of interest to declare. J.E.K. has no disclosures or potential conflicts of interest to declare. C.E.S. is a member of the steering committee for the Australian Football League Brain Health Initiative. L.-N.H. has no disclosures or potential conflicts of interest to declare. R.J.C. is subcontracted to the Lieber Institute for Brain Development to assist with brain examinations. He is a collaborator on a grant funded by the National Football League to study the spectrum of concussion, including possible long-term effects. He has a consulting practice in forensic neuropathology, including expert testimony, which has involved former athletes at amateur and professional levels as well as sport organizations. He was reimbursed for hotel costs at the World Rugby Medical Commission Conference in 2022. He is an Editorial Board Member of this journal but was not involved in the peer-review process nor had access to any information regarding its peer review.

Funding

G.L.I. acknowledges unrestricted philanthropic support from the Schoen Adams Research Institute at Spaulding Rehabilitation. He has received research funding from the Wounded Warrior Project to conduct research relating to CTE and traumatic encephalopathy syndrome, but not for this study. The above entities were not involved in the study design, data collection, analysis, interpretation, writing of this article, or the decision to submit it for publication.

Ethical considerations

The postmortem brain tissue, demographic, and clinical information used in this study were obtained from the Lieber Institute for Brain Development (LIBD). All tissue samples examined were obtained from the collaboration between LIBD and Western Michigan University Homer Stryker MD School of Medicine, Department of Pathology (WCG protocol #1126332).

Consent to participate

At the time of autopsy, audiotaped witnessed informed consent from legal next of kin was obtained for each case.