INTRODUCTION

Abnormal bone growth is a common type of pathology that varies greatlyin its appearance (^{Mann & Hunt, 2005}), largely because the development and physiology of bone and related mesodermal structures is complex (^{Hall, 2005}). The development of dermal bones and how they relate to the dura mater is particularly important in this way, as this relationship can limit the available space for central nervous system growth (^{Opperman et al., 1993}; ^{Presley, 1993}; ^{Jeffery, 2002}). One such abnormal growth that is extremely common, yet whose etiology is still unknown, is hyperostosis frontalis interna (HFI). HFI is characterized by lumpy growths of bone on the inner surface of the frontal bone, some of which can be almost a centimeter in additional thickness. MicroCT studies differentiate it from other skull thickening diseases such as Paget’s disease quite clearly, yet also illustrate that simple lateral x-rays do not sufficiently allow their identification (^{Rühli et al., 2007}). The best means to recognize it remains observation from the inside of the skull, either visually or by reconstructed CT. For this reason, HFI has been studied more by archaeologists and anthropologists than by physicians (^{Hershkovitz et al., 1999}).

HFI is best known from archaeological studies in which the skulls are studied without available medical records or soft tissues (Hershkovitz et al.). Because of its prevalence in 80 % of post-menopausal women, many archaeologists use it as a means to identify the sex of partial skulls (^{Devriendt et al., 2004}; ^{Belcastro et al., 2006}; ^{Mulhern et al., 2006}). Incidence of HFI in men is rare, and the magnitude of manifestation is less severe when reported (Hershkovitz et al.). Thus, there may be some hormonal correlations with HFI, such as estrogen or testosterone levels, which affect the distinguishable presence and severity of HFI in men versus women.

Investigations of abnormal growth at the interface between dura and bone are uncommonly done in living human patients because of the risks to vital structures, resulting in most research to be limited to animal models, cadavers, and radiological studies of living human subjects. Due to regulations on access to medical records for cadaveric studies, this presents a conundrum of sorts. To best identify the etiology of these abnormal growths, it would be best to link them to medical history. However, to study their morphology, some of the most effective methods are invasive (including histopathology) and are difficult to access in living people.

The prevalence of HFI in women and the challenge of recognizing it in living people have caused this to be relatively less studied. No negative symptoms have been identified yet for HFI, but this may be due to its prevalence causing its effects to be considered normal. But, just as hyperostoses of the dura and ligaments around the spinal cord can cause problems (^{Dunlop & Stirling, 1996}), it isn’t unreasonable to consider that the prevalence of HFI in postmenopausal women may be linked to other prevalent symptoms in this population, such as migraines (^{Stewart et al., 1992}). Hormonal relationships during the course of a woman’s life may be an effect of HFI. Findings which show that greater magnitudes of HFI increase with age (Hershkovitz et al.) further support this theory. During the course of normal menstrual cycles in young women, follicle stimulating hormone (FSH) and luteinizing hormone (LH) are released from the pituitary to regulate the cycle, and estradiol is produced by the ovaries. When a woman becomes menopausal, the levels of FSH and LH increase, while estradiol and progesterone levels decrease. Ovarian produced estradiol directly affects the pituitary and thus plays an essential role in the hypothalamic-pituitary-gonadal axis which regulates gonadotropin production (Morishima et al., 1995; Meethal et al., 2009). One may suggest that a decrease in estradiol increases the presence of HFI. According to this assumption, post-menopausal women and men with decreased testosterone levels should present with HFI.

In women, estradiol is produced by the ovaries. In men, interstitial cells (Leydig cells) in the testes produce testosterone, which is then converted to its active metabolites, one of which is estradiol (Carani et al., 1997). In both sexes, estradiol is also produced in the arterial walls and the brain (Meethal et al., 2009). Since estradiol is thus found in excess in the cranial region, it makes sense that a change in estradiol would affect that area. Additionally, enterohepatic circulation contributes to maintaining estradiol levels since it is conjugated in the liver by sulfate and glucuronide formation. Estradiol is also known to be linked with cholestasis, which histologically presents as bile stasis (Kobayashi et al., 2013).

Thus, abnormal liver states and diseases, such as an increase in estrogenic levels leading to cirrhosis, hepatitis, or even resulting bile stasis may affect the presence of HFI.

According to Hershkovitz et al. there are 4 categories of HFI, varying in the size and location of the hyperostosis. Type A is classified as elevated bony overgrowths that may be unilateral, bilateral, single, or multiple. They are generally less than 10mm in size with indented margins. Type B overgrowths are more nodular and do not have discrete margins. Type C is identified by more extensive nodular overgrowth than B and greater, more irregular thickening of the bone that can cover up to 50 % of the frontal endocranial surface and can be greater than 10mm. Type D is classified as a continuous overgrowth covering more than 50 % of the frontal endocranial surface. The size will be greater than 10mm, with sharp, clearly demarcated margins.

This makes HFI easy to recognize in cadaver populations. Human cadavers in anatomy labs are a resource for sampling histopathology that one cannot attain from anthropology osteological collections. Though HIPPAA rules prevent access to medical records that would make these cadavers more useful in the interpretation of associations of the share etiology of HFI and other pathologies, the ability to extensively sample the tissues of cadavers is an asset in comparison to the limits of biopsy in living people. Also, HFI in females is extremely common in anatomy lab populations because the average ages of body donors tend to be among the elderly, including males that mostly lack HFI and postmenopausal women that have it. Together, this makes the sampling of cadavers from anatomy lab an opportunity to thoroughly sample a sample of appropriate age. Here, we have histologically sampled the livers, gonads, and pituitaries of cadavers over a period of 3 different years, accumulating data for 36 individuals. Associating histopathological findings with the appearance of HFI in cadavers here will allow us assess the assumption that HFI is linked to changes in hormone levels.

Does HFI significantly differ in its distribution among males and females from the Anatomy Lab cadaver sample? If not, then we should expect the distribution of HFI to be equal among males and females when controlled for age. If so, the distribution of HFI should be greater among females than among males in older age groups that include postmenopausal women. Among males with HFI, is there significant evidence of a greater prevalence of endocrine disrupting pathology? If not, then the distribution of HFI among men should not significantly relate to the distribution of endocrine disrupting pathologies. If so, the distribution of HFI among men should significantly relate to the distribution of endocrine disrupting pathologies, such as testicular atrophy and decreased numbers of interstitial cells. Lastly, is there a relation between liver abnormalities and HFI? If not, the distribution of HFI should show no significant correlation with liver pathologies. If so, the distribution of HFI should significantly relate to liver pathologies found.

MATERIAL AND METHOD

Cadavers from the Anatomy Lab at the NYIT College of Osteopathic Medicine were used in this study. Cadavers sampled included 14 males and 22 females (Table I). Because these specimens do not meet the definition of a human subject according to the IRB, they do not require IRB exempt review. The study was HIPAA compliant and adhered to the tenets of the Declaration of Helsinki. Each individual was given a unique ID number by the NYITCOM Anatomy Lab to protect their identity from the authors.

For each individual, the deep surface of the cranial fossa was inspected for evidence of HFI, including distinct regions and the severity. The regions included were the different portions of the frontal, including the frontal sinus, orbital roof and anterior and posterior portions of the frontal’s contribution to the calvarium. In addition, HFI extending beyond the frontal was noted if present on the temporal, sphenoid, parietal, and occipital. For each individual, the grade of HFI severity followed the standardized way done for archaeological specimens to ensure data equivalency (Hershkovitz et al.). Photographs of the deep surface of the calvaria were taken for reference.

In addition, tissue samples of livers, pituitaries, and gonads were collected, if available. All tissues were placed in cassettes and were processed in the NYIT College of Osteopathic Medicine Histology Lab by Alice O’Connor. Once sectioned and stained with hematoxylin and eosin, slides were studied using a bright field light microscope and with consultation by Dr. Maria Plummer (clinical pathologist) for confirmations of interpretation.

Table I Data for each individual in this study, including sex, HFI status, and histopathology of liver and gonads. For HFI status, other bones include temporal (T), sphenoid (S), parietal (P), and occipital (O). 

RESULTS

Out of the 36 cases examined, some level of HFI was present in50 % of males and 95 % of females. All males with HFI had testicular atrophy or had the testes surgically removed pre-mortem. Of the males that did have HFI, category C was most commonly present, appearing in 25 % of all males. All male cases with HFI category C had moderate testicular atrophy and 100 % of male cases with HFI category D displayed a severe amount of testicular atrophy (Figs. 1 and 2). Decreased numbers of interstitial cells were present in 83.3 % of males with HFI and across each category of HFI, indicating a correlational trend between decreased numbers of interstitial cells and the presence of HFI. Some form of testicular atrophy was common for most of the males in this study, though the only healthy individual and all of the males that showed only slight evidence of atrophy all lacked evidence of HFI.

Fig. 1 Males with Class D HFI. 

Fig. 2 Microscopic images of testicular atrophy and decreased interstitial cells (Leydig cells) presence. Pictures taken at 2x zoom (right) and 4x zoom (left). 

All but besides one female (21/22) from this study exhibited evidence of HFI. Ovarian tissues from these provided evidence of fibromas in two, complete atrophy in one, and being healthy and post-menopausal in nine. Four were too poorly preserved to read and four were unavailable because of pre-mortem hysterectomy.

Liver pathologies were chosen by the most frequent observations made on the histology of the samples taken (Fig. 3). Males had largely negative associations of HFI with liver pathologies. A negative pattern was found between bile stasis and HFI in males, such that 5 of the of cases without HFI had bile stasis and only one of the males with HFI had bile stasis. Similarly, of the four cases with signs of passive congestion, one male with hepatitis, and one male with signs of fibrosis, none of them showed evidence of HFI. The only exceptions were that the one case of liver neoplasia was found in a male with HFI and steotosis was found in two cases with HFI and one case without HFI.

Female cases generally had the opposite pattern, where all cases of liver pathologies were associated with HFI. There were four cases of steatosis, one of hepatitis, six of passive congestion, six of fibrosis, and two of bile stasis.

Fig. 3 Microscopic image of a liver at 2x zoom with extreme bile stasis, nodular cirrhosis, and possible cirrhotic implications (left). Microscopic image of focal hepatic vessels at 10x zoom that show evidence of a metastatic neoplasm of neuroendrocrine origin (right). 

DISCUSSION

HFI is found more frequently in women than men (Hershkovitz et al.), leading us to we hypothesize that there may be some endocrine related origin or correlation, specifically in relation to estradiol. Estradiol is produced in women during their menstrual cycle and is the essential messenger hormone in the hypothalamus-pituitary-gonadal axis that regulates hormonal secretion and the menstrual cycle (Morishima et al., 1995; Meethal et al., 2009). In post- menopausal women, estradiol concentration decreases drastically, roughly to the same level found in men. Men produce estradiol as an essential metabolite of testosterone, produced in the interstitial cells. Decreased numbers of interstitial cells result in less testosterone, which thus results in a decreased concentration of estradiol (Carani et al., 1997). Estradiol is also found in the brain and in the arterial walls (Meethal et al., 2009), possibly contributing to the development of HFI.

In this study, there was a correlation found between atrophic testes and the presence of HFI. Furthermore, 83.3% of males that had HFI also showed decreased numbers of interstitial cells in the histology samples taken. In accordance with our hypothesis, these findings suggest that most men with HFI had decreased testosterone production, and thus a decreased amount of estradiol. Since there is no current access to cadaveric medical records, it is unclear as to what age the atrophic states in these males began and if there was any physiologic manifestation recorded. Additionally, the severity of the atrophy was seen to increase as the level of HFI increased, with all males that had HFI category C showing moderate atrophy and all males with HFI category D showing severe atrophy. These findings suggest that the level of HFI increases as the drop of estradiol from the baseline increases. Since most of the female cadavers were of post-menopausal age, their ovarian samples appeared hyalinized, which is normal in post-menopausal women (O’Neill & Eden, 2017). This is consistent with the post- menopausal drop in estradiol production and the greater frequency of HFI found in post-menopausal women.

Estrogenic effects are also found on the liver because estradiol undergoes hepatic metabolism (Kobayashi et al., 2013). The most common findings and estrogen related pathologies in this study were recorded when the histology was viewed. A significant inverse relationship was found between bile stasis and HFI. Estrogenic effects have been noted in relation to cholestasis (Kobayashi et al., 2013), which histologically manifests as bile stasis. A significant decrease in estradiol, hypothesized to relate to HFI, would then show no bile stasis, as observed in our research findings. There was one histological liver sample in which a metastasis of an endocrinological origin was observed. In this specific case, atrophic testes and decreased numbers of interstitial cells were observed as well as HFI category C. Since this study was small and only one such case was found, this particular finding does not hold much value. However, it may be beneficial to further investigate such correlations and the effect of such metastasis on estradiol production.

The results of this study strongly indicate a hormonal relationship to HFI. Overall, the findings indicate that a decrease from a baseline level of estradiol has an effect on HFI. Further studies may be performed using a larger population of cadavers to gain more accurate statistics and results, and further sampling of other endocrine tissues, such as the pituitary and pancreas. Although estradiol has been shown to maintain bone mass density (Grundberg et al., 2007), additional research may be useful for examining the relationship between estradiol and bone mass formation, specifically in relation to the flat bones of the skull (Niada et al 2016). In males there is a well-known relationship of estradiol to maintenance of bone during aging (Russell & Grossman, 2019), though it isn’t clear whether this manifests as a mobilization of calcium for deposition elsewhere in soft tissues as can be found in cases of cholecalciferol toxicity (^{Kocher et al., 2010}). Furthermore, since estradiol is necessary for epiphyseal fusion (MacGillivray et al., 1998), studies should look into the effects of estradiol on the sutures of the skull, particularly the coronal suture. Though most research seems to indicate a relationship of estradiol to bone growth early in development (^{Flöter et al., 2016}), studies of animal models and/or humans of advanced age may be better suited for understanding this etiology. If access is granted to medical records of cadavers, the physiologic manifestation of HFI causes and effects may be more accurately observed and recorded.

In addition, an important consideration is the possible relation of HFI to dementia, as this relationship can be defined with two distinct mechanisms. One such mechanism could involve the direct effect of the cranial bone thickening on the brain tissue itself. The severity of HFI has been correlated with the decrease in intracranial volume. There may be a possible association between HFI and the beginning of cognitive decline (^{May et al., 2012}). Further studies can investigate the symptomatic effects of HFI on postmenopausal women, especially since it has been more prevalent nowadays compared to older generations (May et al.).

Another mechanism for HFI connection to dementia could be the estrogen exposure itself. Estrogenic receptors have been noted to be present in important regions of the brain for learning and memory, such as the prefrontal cortex and hippocampus (^{McEwen et al., 2012}). In fact, estrogen also affects glucose metabolism in the brain. Towards menopause, the decrease in estrogen results in a decrease in glucose metabolism in the brain. The compensatory pathways have been observed to contribute to the deposition of amyloid beta proteins, a key mechanism is the development of Alzheimer’s disease (^{Yao et al., 2012}; ^{Ding, et al., 2013}; ^{Yin et al., 2015}; ^{Rettberg et al, 2016}). Further studies could examine whether HFI can indirectly affect brain function through the effects of estrogen exposure. Although studying HFI is limited by physicians today, understanding the effects of HFI on brain function could lead to a screening test for post-menopausal women in the future. HFI has been associated with decreased estrogen, which in turn has already been shown to be connected to dementia on many accounts. Therefore, it would stand to reason that HFI may indirectly serve as a good screening tool for possible development of dementia.

Such an investigation has the power to unlock the mysteries of HFI and potentially prevent patients from any suffering that may be caused from this unknown etiology. If such a large percentage of post-menopausal women have this, it is essentially “normal”. Could this be a treatable pathology hiding in its prevalence?