From Pituitary Expansion to Empty Sella: Disease Progression in a

GENERAL
ENDOCRINOLOGY
From Pituitary Expansion to Empty Sella: Disease
Progression in a Mouse Model of Autoimmune
Hypophysitis
Isabella Lupi, Jiangyang Zhang, Angelika Gutenberg, Melissa Landek-Salgado,
Shey-Cherng Tzou, Susumu Mori, and Patrizio Caturegli
Department of Endocrinology and Metabolism (I.L.), University of Pisa 56124 Pisa, Italy; Division of
Magnetic Resonance Research (J.Z., S.M.), Russell H. Morgan Department of Radiology and Radiological
Science and Department of Pathology (M.L.-S., S.-C.T., P.C.), The Johns Hopkins University School of
Medicine, and Feinstone Department of Molecular Microbiology and Immunology (P.C.), The Johns
Hopkins Bloomberg School of Public Health, Baltimore, Maryland 21205; and Department of
Neurosurgery (A.G.), Georg August University, 37073 Göttingen, Germany
Lymphocytic hypophysitis has a variable clinical course, where a swelling of the pituitary gland at
presentation is thought to be followed by pituitary atrophy and empty sella. Data in patients,
however, are scanty and contradictory. To better define the course of hypophysitis, we used an
experimental model based on the injection of pituitary proteins into SJL mice. A cohort of 33 mice
was divided into three groups: 18 cases were immunized with pituitary proteins emulsified in
complete Freund’s adjuvant; six controls were injected with adjuvant only; and nine controls were
left untreated. Mice were followed by cranial magnetic resonance imaging (MRI) for up to 300 d,
for a total of 106 MRI scans, and killed at different time points to correlate radiological and
pathological findings. Empty sella was defined as a reduction in pituitary volume greater than 2
SD below the mean volume. All immunized mice showed by MRI a significant expansion of pituitary
volume during the early phases of the disease. The volume then decreased gradually in the majority
of cases (14 of 18, 78%), reaching empty sella values by d 300 after immunization. In a minority of
cases (four of 18, 22%), the decrease was so rapid and marked to induce a central area of necrosis
accompanied by hemorrhages, mimicking the condition known in patients as pituitary apoplexy.
No radiological or pathological changes were observed in controls. Overall, these findings indicate
that the evolution of hypophysitis is complex but can lead, through different routes, to the development of empty sella. (Endocrinology 152: 4190 – 4198, 2011)
H
ypophysitis (inflammation of the pituitary gland) comprises a complex and expanding spectrum of pathological
lesions and causes. In most reported patients, the inflammation
is limited to the pituitary and has no identifiable causes (primary
hypophysitis). More rarely, the inflammation is secondary to
sellar processes (like Rathke cleft cysts or craniopharyngiomas),
systemic diseases (like tuberculosis, syphilis, Wegener granulomatosis, or sarcoidosis) (1), or pharmacological treatments like
blockade of cytotoxic T lymphocyte antigen 4 (2).
Primary hypophysitis typically presents as a sellar mass
with signs and symptoms originating from compression of
nearby structures (like headaches and visual-field defects)
or from hormonal deficiencies. Its diagnosis is often one of
exclusion and is based on either pathological examination
of the pituitary biopsy or clinico-radiological grounds
alone (3). When pathology is available, hypophysitis is
classified into five forms: lymphocytic, granulomatous,
xanthomatous, necrotizing, and plasma cell rich. Lymphocytic hypophysitis is the most prevalent form, with
approximately 400 biopsy-proven patients published
from 1962–2011. It is an autoimmune disease characterized by a marked infiltration of lymphocytes within
ISSN Print 0013-7227 ISSN Online 1945-7170
Printed in U.S.A.
Copyright © 2011 by The Endocrine Society
doi: 10.1210/en.2011-1004 Received March 15, 2011. Accepted August 1, 2011.
First Published Online August 23, 2011
Abbreviations: AH, Autoimmune hypophysitis; Gd, gadopentetate dimeglumine; MRI,
magnetic resonance imaging.
4190
endo.endojournals.org
Endocrinology, November 2011, 152(11):4190 – 4198
Endocrinology, November 2011, 152(11):4190 – 4198
the pituitary gland. It can be mimicked experimentally
in the mouse by immunization with pituitary proteins
(4). Granulomatous hypophysitis has been reported in
over 120 patients since 1908. It features multinucleated
giant cells that organize in granulomas with palisading
histiocytes surrounded by T cells and plasma cells (5).
Xanthomatous hypophysitis, described in 13 patients
since 1998, displays foamy histiocytes and macrophages, accompanied by plasma cells and lymphocytes
(6). Necrotizing hypophysitis, reported in three patients
since 1993, is characterized by mononuclear infiltration
within a pituitary tissue that shows significant nonhemorrhagic necrosis (7). Finally, IgG4-related hypophysitis,
the most recent addition to the hypophysitis spectrum,
described in 13 patients since 2004, is characterized by a
mononuclear infiltration of the pituitary gland containing
more than 10 IgG4-producing plasma cells per highpower field, usually accompanied by IgG4-positive lesions
in other organs (8).
The natural history of hypophysitis is variable, ranging
from complete resolution to death (1, 9). The majority of
patients (65%) require some form of long-term hormone
replacement; other patients (20%) improve after massreducing treatments (such as pituitary surgery or highdose glucocorticoids) without need of hormone replacement; some patients (10%) die because of hypophysitis
and are diagnosed at autopsy; in a minority of cases (5%),
hypophysitis is aggressive and recurs after the initial massreducing treatment, so that a second surgery is necessary
to alleviate the mass-effect symptoms. Part of this variability is explainable by differences in the modality and
length of the follow-up, which overall tends to be short
(less than 2 yr after diagnosis) in the majority of published
patients. Morphologically, the pituitary gland is typically
enlarged at presentation and likely shrinks during followup. Some authors have also reported by magnetic resonance imaging (MRI) a loss of intrasellar volume and secondary empty sella (10).
Empty sella is a herniation of the subarachnoid space
into the sella turcica, leading to stretching of the stalk and
flattening of the pituitary gland against the sellar walls
(11). The term empty sella was originally used at autopsy
to describe the appearance of the sellar region in patients
with postpartum pituitary necrosis (12) but is nowadays a
neuroradiological diagnosis. Morphologically, empty
sella is classified into total, when more than 50% of the
sella is filled with cerebrospinal fluid and the pituitary
gland thickness is less than 2 mm, or partial, when less than
50% is filled and the pituitary thickness is at least 3 mm
(13). Mechanistically, empty sella recognizes primary and
secondary forms. Primary empty sella associates with congenital anatomic variations of the sellar diaphragm that
endo.endojournals.org
4191
are typically discovered incidentally, or with suprasellar
causes that lead to increases in intracranial pressure (14).
Secondary empty sella, on the contrary, is caused by a loss
of pituitary volume. This loss can be seen after extensive
pituitary surgery, after prolonged pituitary radiotherapy,
in Sheehan syndrome (ischemic necrosis after peripartum
hemorrhages), or in pituitary adenomas that undergo
hemorrhagic necrosis (pituitary apoplexy). Hypophysitis
is a potential cause of secondary empty sella if it is true that
the pituitary gland targeted by autoimmunity becomes
atrophic with time.
Given the rarity of hypophysitis, the paucity of prolonged follow-up times, and the fact that most patients
undergo some forms of mass reduction (thus preventing
observation of a bona fide natural history), it remains uncertain whether empty sella is indeed a final outcome of
autoimmune hypophysitis. The goal of this study was to
characterize by MRI the evolution and late course of lymphocytic hypophysitis using a previously developed mouse
model of the disease (4).
Materials and Methods
Study design and immunization scheme
The study used a cohort of 33 SJL/J female mice (from The
Jackson Laboratory, Bar Harbor, ME) distributed into three
groups. Cases (n ⫽ 18) were immunized twice (on d 0 and 7,
corresponding to ages of 63 and 70 d) using 0.5 mg mouse
pituitary cytosolic proteins emulsified in complete Freund’s
adjuvant, according to the protocol previously described (4).
Controls included six age-matched mice that received on d 0
and 7 only the complete Freund’s adjuvant without pituitary
proteins (CFA controls), and nine age-matched mice that were
left untreated (nonimmunized controls). All experimental
procedures were approved by the Animal Care and Use Committee of the Johns Hopkins University School of Medicine.
TABLE 1. Distribution of the MRI studies (n ⫽ 106)
according to age (in days) and experimental group
Not immunized
Age (d)
50
60
80
100
140
180
200
240
300
340
Total
2
2
4
3
3
2
5
3
24
CFA only
3
6
3
12
Immunized
5
5
20
11
6
5
3
8
4
3
70
4192
Lupi et al.
MRI Studies in Murine Hypophysitis
Endocrinology, November 2011, 152(11):4190 – 4198
TABLE 2. Distribution of the MRI studies (n ⫽ 106)
according to the day after immunization and
experimental group
d after immunization
0
12
20
27
32
40
70
100
140
175
225
300
Total
Not
immunized
CFA
only
24
3
6
3
24
12
Immunized
5
10
5
6
4
7
10
4
4
4
8
3
70
CFA, Complete Freund’s adjuvant. Numbers in bold refer to the 32
MRI studies used to describe the normal mouse pituitary volume.
In vivo MRI of the mouse brain
MRI was performed using a horizontal 9.4-Tesla nuclear
magnetic resonance spectrometer (Bruker Biospin, Billerica,
MA) and a 40-mm-diameter birdcage coil as radio frequency
transmitter and receiver. Mice were anesthetized using 1% isoflurane in a mixture of 75% air and 25% oxygen and placed in
a custom-designed head holder with ear pins and nose cone to
restrain head motion. Respiration was monitored using a smallanimal imaging system (SA instruments, Inc., Stony Brook, NY)
and kept at a rate of approximately 60 – 80 breaths per minute.
T1-weighted images were acquired using a two-dimensional
multiple-slice spin echo sequence, with an echo time of 11 msec,
a repetition time of 500 msec, in-plane resolution of 0.1 mm ⫻
0.1 mm, 24 coronal slices with 0.3-mm slice thickness, and 10
signal averages. Image reconstruction was performed on the
spectrometer console (Paravision, version 3.0.2; Bruker Biospin). Two of 33 mice (6%) died toward the end of the MRI
procedure because of the anesthesia.
Gadolinium [gadopentetate dimeglumine (Gd) Magnevist from
Berlex Imaging, Montville, NJ] was diluted 1:3 in sterile saline solution and injected intra-peritoneally in a volume of 100 ␮l for a
dose of 0.65 mmol/kg body weight. Pilot experiments performed in
normal SJL/J female mice not included in the 30 mice described in
the present study showed that Gd enhancement of the pituitary
gland began 10 min post-injection, disappeared by 100 min, and
required about 10 min to obtain an image.
A total of 106 MRI studies were performed at various ages
(Table 1) and days post-immunization (Table 2). The time points
after immunization were chosen based on previous knowledge of
the course of hypophysitis in this experimental mouse model (4).
In all studies, T1-weighted images after Gd injection (T1 postcontrast) were obtained sequentially at 10, 20, 30, 40, 50, 60,
and 70 min. In 16 studies, T1-weighted images were also acquired before (T1 precontrast) Gd, for a study duration of approximately 100 min per mouse.
Pituitary histopathology
Immunized mice were killed at sequential time points to correlate the pituitary MRI findings with pathological changes. Pituitary glands were fixed overnight in Beckstead’s solution, embedded in paraffin, and processed as described (4). Sections were
stained with hematoxylin and eosin for morphology and Masson’s trichrome for fibrosis. Sections were scored qualitatively by
three investigators (S.C.T., M.L.S., and P.C.).
Pituitary autoantibodies
Pituitary autoantibodies were measured by ELISA using the
previously described homemade assay (4). Results are expressed
as arbitrary units per microliter.
Statistical analysis
The study analyzed four outcomes: pituitary volume (Figs.
1–3), pituitary histopathology (Fig. 4), pituitary signal intensity
before and after Gd (Figs. 5 and 6), and pituitary antibodies (Fig. 7).
Pituitary volume was calculated by manually segmenting
the anterior and posterior pituitary using Amira software
(Mercury Computer Systems, Inc., Carlsbad, CA). Manual
segmentation was performed in T1-weighted
images acquired at 30 min after Gd injection
because anterior pituitary, posterior pituitary,
and other brain regions can be easily defined at
this time point. Pituitary volume followed a
normal distribution, and thus, parametric tests
were used to analyze differences among immunization groups and over time.
Pituitary histopathology was classified qualitatively as normal, infiltrated by lymphocytes, infiltrated by lymphocytes and atrophic, or infiltrated by lymphocytes and hemorrhagic.
Pituitary signal in the anterior and posterior
lobe before Gd injection was compared among
the immunization groups using the nonparametFIG. 1. Pituitary volume in normal (not immunized) mice. A, Total pituitary volume
ric Wilcoxon rank sum test. After Gd injection,
(anterior plus posterior lobe) throughout the lifetime. Note the significant increase from
the intensity at the various time points was aver50 – 60 d of age. B, Distribution of total pituitary volume in adult (⬎60 d) mice. The
aged and then normalized to the intensity of the
thick line shows the kernel density estimate of the volume based on the mean and SD of
cerebral cortex, which does not vary with Gd due
22 MRI scans. The thin line reports the normal density that would be obtained based on
to the presence of the blood-brain barrier. Averthe same mean and SD. The dotted lines in both panels represent the mean plus 2 SD
aged and normalized intensity of the anterior and
and the mean minus 2 SD values of the total pituitary volume distribution.
Endocrinology, November 2011, 152(11):4190 – 4198
endo.endojournals.org
4193
Statistical analyses were performed using Stata statistical
software, release 10 (StataCorp LP, College Station, TX).
Results
Normal pituitary volume
In normal (nonimmunized) mice (N ⫽ 32), the total
volume of the pituitary gland increased significantly from
50 – 60 d of age (P ⫽ 0.001, Fig. 1A, closed circles), a
period corresponding to completion of puberty and sexual
maturation. This increase was due exclusively to the anterior pituitary (Fig. 1A, open diamond) because the posterior pituitary remained largely unchanged throughout
life (Fig. 1A, closed diamond). In adult (⬎60 d of age) mice
(N ⫽ 25), the total pituitary volume had a mean of 2.47
mm3 and SD of 0.26 mm3 and followed a normal distribution (Fig. 1B). Volumes greater than 2.98 mm3 (mean ⫹
2 SD) were considered indicative of pituitary enlargement,
and volumes smaller than 1.69 mm3 (mean ⫺ 3 SD) indicative of pituitary atrophy and empty sella. The mean (SD)
of adult anterior and posterior pituitary volumes were
2.23 (0.26) and 0.23 (0.08) mm3, respectively.
FIG. 2. Pituitary volume after immunization. A, Trend of pituitary
volume over time in the entire cohort of 18 mice immunized with
pituitary proteins; B, volume in the 14 mice that developed
pathologically the lymphocytic form of hypophysitis; C, volume in the
four mice that developed the hemorrhagic (apoplectic) form of
hypophysitis. Each mouse is indicated by a different symbol. Note the
more rapid decline in pituitary volume. The dotted lines in all panels
represent the mean plus 2 SD and the mean minus 2 SD values of the
total pituitary volume distribution.
posterior lobe were then compared among groups by the Wilcoxon rank sum test.
Pituitary antibodies followed a nonnormal distribution and
were thus natural log-transformed to satisfy assumptions of normality and equal variances. Results were then analyzed by multiple linear regression to assess changes in antibody levels according to age and pituitary volume.
Pituitary volume after immunization
Immunization with pituitary proteins, begun after 60 d
of age, markedly modified the pituitary volume. Volume
became significantly greater than baseline on d 20 (3.37 ⫾
0.8 mm3, P ⫽ 0.0003 vs. d 0, Fig. 2A, white circle) and
peaked 1 month after immunization (3.75 ⫾ 1.9, P ⬍
0.001 vs. d 0, Fig. 2A, white diamond). Volume then remained significantly larger than baseline on d 40, 70, 100,
and 140 after immunization, started to decrease on d 175,
and approached baseline values on d 225 (1.9 ⫾ 0.68). On
d 300, volume became significantly smaller than baseline,
reaching atrophic values (1.08 ⫾ 0.38, P ⬍ 0.0001 vs. d 0,
Fig. 2A). Pituitary volume varied significantly during the
initial phase of hypophysitis (first month after immunization), a variation partly explainable by two distinct patterns of radiological and morphological lesions.
Lymphoid form
The majority of immunized mice (14 of 18, 78%) developed a clear initial enlargement of the pituitary gland
followed by a gradual and slow decline toward atrophy
(Fig. 2B). The initial enlargement was clearly detectable by
MRI (Fig. 3B) and sustained pathologically by a diffuse
lymphocytic infiltration of the adenohypophysis (Fig. 4B).
At later stages, the pituitary gland became markedly atrophic (Fig. 3C) and showed prominent fibrosis (Fig. 4C) in
addition to the lymphocytes that remained the predominant hematopoietic cell type throughout the course of the
4194
Lupi et al.
MRI Studies in Murine Hypophysitis
Endocrinology, November 2011, 152(11):4190 – 4198
pituitary cells, overall resembling a sort
of hypophyseal cirrhosis.
Apoplectic form
In the remaining minority of immunized mice (four of 18, 22%), the initial
expansion of the pituitary was so rapid
it induced a central area of necrosis accompanied by extravasation of blood
from the vessels (Figs. 3D and 4D). This
form of hemorrhagic necrosis caused a
rapid increase in pituitary volume followed by a brisk, rather than gradual,
decrease (Fig. 2C). It mimics closely a
condition known in patients as pituitary apoplexy that, although more often described in pituitary adenomas,
has also been reported in lymphocytic
hypophysitis (15–20).
The above-described changes in pituitary volume were exclusively seen in
FIG. 3. Axial T1-weighted, Gd-enhanced MR images of the mouse pituitary. A, Normal,
nonimmunized, mouse pituitary. Note the avidly enhancing posterior pituitary (arrow). For an
the adenohypophysis. The posterior piexplanation of the major anatomical structures in the mouse brain, see Supplemental Fig. 1.
tuitary volume remained unchanged afB, Lymphoid form of hypophysitis during the florid disease phase (d 32 after immunization).
ter immunization, in keeping with the
Note the enlargement of the anterior pituitary. C, Lymphoid form of hypophysitis during the
atrophic disease phase (d 300 after immunization). Note the marked reduction in the anterior
lack of lymphocytic infiltration in the
pituitary volume. D, Lymphoid form of hypophysitis during the florid disease phase (d 27 after
posterior pituitary (data not shown). Its
immunization) with hemorrhagic pituitary apoplexy. Note the area of hyperintensity
volume was 0.207 mm3 in all groups,
(corresponding to the hemorrhagic necrosis) within the anterior pituitary (arrow).
with a similar SD around 0.05.
disease. At these later stages, the anterior pituitary disPituitary volume assessed by MRI correlated strongly
played dense, interlacing bands of fibrosis (turquoise with postmortem pituitary weight [Pearson correlation
staining in Fig. 4C), demarking small areas of atrophic coefficient (r) of 0.895], so that it could be used to predict
weight in a univariate linear regression
analysis: for every unit (cubic millimeter) increase in pituitary volume, the pituitary weight increased 0.8576 mg
(95% confidence interval, 0.63– 0.87;
P ⬍ 0.0001).
FIG. 4. Histopathological appearance of hypophysitis according to the histopathological type
and time after immunization. A, Normal mouse pituitary gland; B, acute lymphoid form; C,
chronic (atrophic) lymphoid form; D, acute hemorrhagic form.
Pregadolinium T1signal intensity
Before gadolinium, the mean ⫾ SD
T1 signal intensity of the anterior pituitary gland was similar in immunized
(1.11 ⫾ 0.07) and control mice (1.15 ⫾
0.05, P ⫽ 0.145 by Wilcoxon rank sum
test, Fig. 5A, black boxes) and did not
correlate with anterior pituitary volume (data not shown), overall suggesting that precontrast signal intensity
does not discriminate normal pituitary
from hypophysitis.
Before gadolinium, the T1 signal intensity of the posterior pituitary in un-
Endocrinology, November 2011, 152(11):4190 – 4198
endo.endojournals.org
4195
after 120 d after immunization. Overall, these data suggest that gadoliniumenhanced images distinguish hypophysitis from normal pituitary if acquired
early during disease development.
After gadolinium, the peak (mean ⫾
SD) T1 signal intensity of the posterior
pituitary was similar in immunized
(2.39 ⫾ 0.15) and control (2.36 ⫾ 0.20)
mice.
Dynamic postgadolinium T1signal
intensity
T1 MRI images were acquired 10,
FIG. 5. MRI signal intensity of the mouse pituitary. A, Intensity before gadolinium (Gd)
administration. Note that the posterior pituitary (gray boxes) is significantly brighter than the
20, 30, 40, 50, 60, and 70 min after
anterior pituitary (black boxes). The baseline intensity of both the anterior and posterior
gadolinium administration to characpituitary does not change after immunization. B, Intensity after Gd administration. The
terize the dynamic of contrast uptake.
posterior pituitary retains its inherent stronger brightness but is unaffected by immunization.
The anterior pituitary becomes brighter after Gd, and more significantly so in immunized
The T1 intensity tended to peak at 40
mice. NS, Not significant.
min, but overall, it was not significantly
different from the other time points,
immunized mice was significantly higher than that of the
both
in
controls
(Fig.
6A) and immunized mice (Fig. 6B).
anterior pituitary (1.53 ⫾ 0.06, P ⫽ 0.0001, Fig. 5A),
mimicking the bright spot observed in human pituitary
glands. After immunization, the precontrast posterior pituitary bright spot remained (1.57 ⫾ 0.18, P ⫽ 0.0001 vs.
anterior pituitary, Fig. 5A) and was not different from the
one observed in unimmunized controls.
Peak postgadolinium T1signal intensity
After gadolinium, the peak ⫾ SD T1 signal intensity of
the anterior pituitary gland was significantly higher in the
immunized group (1.97 ⫾ 0.23) than in controls (1.78 ⫾
0.13, P ⫽ 0.0016 by Wilcoxon rank sum test, Fig. 5B,
black boxes). The intensity was directly proportional to
the anterior pituitary volume during the initial disease
phase (regression coefficient 2.73, P ⫽ 0.015), but not
Pituitary antibodies and pituitary volume
Pituitary antibodies, not surprisingly, higher in immunized mice (Fig. 7A, circles) than in unimmunized controls
(Fig. 7A, diamonds), declined gradually with age but remained higher than controls for up to 1 yr after immunization. Antibodies correlated directly with pituitary volume, being higher in mice with larger pituitary volumes
and lower in smaller volumes (Fig. 7B), in keeping with the
notion that autoantibodies decline with decreasing antigenic loads (21). In particular, for every unit increase in the
log of pituitary antibodies, the volume increased 1.10 mm3
(95% confidence interval, 0.48 –1.73; P ⫽ 0.002).
Discussion
FIG. 6. Dynamic T1 pituitary intensity after Gd administration in controls (A) and immunized
(B) mice. Note the small variation around the mean T1 signal intensity after Gd injection in
immunized mice.
We report in a mouse model of lymphocytic hypophysitis that progression toward pituitary atrophy and empty sella
is the final outcome of this disease.
Empty sella was reached gradually over
the course of several months in the majority of mice, or abruptly in those that
developed pituitary apoplexy. The
study thus contributes to the understanding of an aspect of hypophysitis
that in humans remains difficult to ascertain. First of all, because a diagnosis
of certainty of hypophysitis depends
upon surgical intervention, the natural
history of the disease cannot be ob-
4196
Lupi et al.
MRI Studies in Murine Hypophysitis
FIG. 7. Pituitary antibodies after immunization. A, After immunization
on d 0 and 7 (corresponding to 60 and 67 d of age), antibodies
decrease over time, approaching the levels seen in unimmunized
controls at approximately 1 yr; B, antibodies directly correlated with
pituitary volume.
served. More importantly, the follow-up is often incompletely reported or short. When coupled with the rarity of
hypophysitis, these aspects make it difficult to reach firm
conclusions.
Review of the hypophysitis literature for the presence of
an empty sella identified 27 patients, 10 biopsy proven (all
lymphocytic) and 17 clinically suspected (Supplemental
Table 1, published on The Endocrine Society’s Journals
Online web site at http://endo.endojournals.org). The first
case was reported in 1986 by Okada et al. (22) in a 28yr-old woman with clinically suspected hypophysitis and
pituitary antibodies who developed empty sella 2 yr after
diagnosis. Empty sella was present at diagnosis in two
patients (23, 24) and reported at some time during follow-up in the remaining 25 cases. Follow-up, which
ranged from 0.25 (25) to 24 (26) years, was short in cases
where empty sella developed after pituitary surgery and
more prolonged in those where no mass-reducing treatment (either surgical or with high-dose steroids) was used
Endocrinology, November 2011, 152(11):4190 – 4198
(Supplemental Table 1). Pituitary antibodies were measured in 11 of the 27 patients and found positive in three
(11%). Although variable and limited in number, these
human data suggest that hypophysitis can be a cause of
secondary empty sella and are supported by the present
experimental study.
For primary empty sella, the issue of a possible association with hypophysitis and pituitary autoimmunity is
more controversial. The clinical setting here is that of a
patient diagnosed with empty sella without demonstrable
causes (including hypophysitis) who is found to be positive
for pituitary antibodies. Considering that hypophysitis
can be asymptomatic and undetectable for years, some
authors have wondered whether the presence of pituitary
antibodies in these patients actually reflect a subclinical
form of hypophysitis that has then evolved toward pituitary atrophy and empty sella. The published studies have
been difficult to interpret and compare, largely due to the
enormous variability in the methods used to detect pituitary antibodies. Komatsu et al. (27) first analyzed 32 patients with primary empty sella for the presence of pituitary antibodies measured by immunofluorescence using
mouse corticotroph cells as the antigenic substrate and
found them in 24 (75%), significantly more than what was
found in normal subjects (zero of 25, 0%). Mau et al. (28)
studied six patients with primary empty sella and measured pituitary antibodies by immunoblotting using human GH and ACTH as antigens, reporting two positive
(33%), whereas none of the six healthy controls were.
Keda et al. (29) studied 48 patients with primary empty
sella and measured pituitary antibodies by ELISA using as
the antigenic substrate surgically removed human prolactinomas and somatotropinomas, then trypsinized to
obtain a suspension of single cells that were then immobilized to the ELISA plates and fixed with glutaraldehyde.
Antibodies were found in 25 patients (52%), a significantly higher prevalence than that found in normal adults
(three of 65, 5%). Bensing et al. (30) analyzed 30 patients
with primary empty sella measuring pituitary antibodies
by immunoblotting using 49-kDa ␣-enolase as the antigen. Antibodies were found in six of the 30 patients (20%),
a prevalence not different from that found in healthy controls (11 of 50, 22%). Finally, we reported recently 85
patients with primary empty sella where pituitary antibodies were measured by immunofluorescence using monkey pituitary as substrate. Antibodies were found in five of
the 85 empty sella patients (6%), a prevalence significantly
higher than that seen in healthy controls (one of 214,
0.5%, P ⫽ 0.008 by ␹2 test) (31). Interestingly, the antibody-positive patients were also hypopituitaric, suggesting that hypopituitarism in empty sella may be the consequence of tissue damage due to hypophysitis.
Endocrinology, November 2011, 152(11):4190 – 4198
The variability and lack of standardization in the methods used to detect pituitary antibodies make these five
studies difficult to compare. In addition, it is important to
consider the timing of antibody measurement. We have
shown in this study that pituitary antibodies decrease in
titer with decreasing pituitary volume, approaching the
upper limit of normality in mice that have an atrophic
pituitary. In addition, it is known that antibodies to thyroid antigens progressively disappear after complete ablation of the thyroid (either by surgery or radioiodine),
indicating that continued antibody production requires
persistence of the targeted antigens (21). Therefore, if primary empty sella is indeed caused by subclinical autoimmune hypophysitis, by the time the empty sella is discovered, the pituitary gland is by definition atrophic and thus
predicted to have low antigenic load and consequently low
pituitary antibody titers.
Another interesting aspect of this study was the comparison of the MRI findings between mouse and human
hypophysitis. In patients, hypophysitis at presentation is
radiologically characterized by symmetric enlargement of
the pituitary, intense and homogeneous enhancement after gadolinium administration, loss of posterior pituitary
bright spot, and thickened pituitary stalk (3). Enlargement
and enhancement were mimicked by our mouse model,
which also showed that gadolinium enhancement is useful
only during the early phase of hypophysitis. This enhancement reflects a hypervascularity of the pituitary that is
present early in the disease course but is then replaced by
fibrosis in more advanced stages, contributing to both the
delayed and reduced enhancement seen in our study. Delayed enhancement has been described in adults with hypophysitis (32) and in children with hypopituitarism (33).
Loss of posterior bright spot and thickened stalk, on the
contrary, were not seen in murine hypophysitis, in keeping
with the observation that this experimental model induces
only adenohypophysitis and not infundibulo-neurohypophysitis (4).
In summary, we report the first MRI study of hypophysitis in mice and show that pituitary atrophy and secondary empty sella is the final outcome of autoimmune
hypophysitis.
Acknowledgments
Address all correspondence and requests for reprints to: Patrizio
Caturegli, Johns Hopkins University, Department of Pathology,
Ross Building Room 632, 720 Rutland Avenue, Baltimore,
Maryland 21205. E-mail: pcat@jhmi.edu.
This work was supported by Grant DK080351 to P.C. A.G.
was supported by the Research Program from the Faculty of
endo.endojournals.org
4197
Medicine, Georg-August-University Göttingen, Germany, and
the Tönnis-Stipendium by the German Society for Neurosurgery
(DGNC).
Disclosure Summary: The authors have nothing to disclose.
References
1. Caturegli P, Newschaffer C, Olivi A, Pomper MG, Burger PC, Rose
NR 2005 Autoimmune Hypophysitis. Endocr Rev 26:599 – 614
2. Gutenberg A, Landek-Salgado M, Tzou SC, Lupi I, Geis A, Kimura
H, Caturegli P 2009 Autoimmune hypophysitis: expanding the differential diagnosis to CTLA-4 blockade. Expert Rev Endocrinol
Metab 4:681– 698
3. Gutenberg A, Larsen J, Lupi I, Rohde V, Caturegli P 2009 A radiologic score to distinguish autoimmune hypophysitis from nonsecreting pituitary adenoma preoperatively. AJNR Am J Neuroradiol 30:
1766 –1772
4. Tzou SC, Lupi I, Landek M, Gutenberg A, Tzou YM, Kimura H,
Pinna G, Rose NR, Caturegli P 2008 Autoimmune hypophysitis of
SJL mice: clinical insights from a new animal model. Endocrinology
149:3461–3469
5. Shi J, Zhang JM, Wu Q, Chen G, Zhang H, Bo WL 2009 Granulomatous hypophysitis: two case reports and literature review.
J Zhejiang Univ Sci B 10:552–558
6. Aste L, Bellinzona M, Meleddu V, Farci G, Manieli C, Godano U
2010 Xanthomatous hypophysitis mimicking a pituitary adenoma:
case report and review of the literature. J Oncol 2010:195323
7. Gutenberg A, Caturegli P, Metz I, Martinez R, Mohr A, Bruck W,
Rohde V 10 April 2011 Necrotizing infundibulo-hypophysitis: an
entity too rare to be true? Pituitary 10.1007/s11102-011-0307-2
8. Leporati P, Landek-Salgado MA, Lupi I, Chiovato L, Caturegli P
2011 IgG4-related hypophysitis: a new addition to the hypophysitis
spectrum. J Clin Endocrinol Metab 96:1971–1980
9. Caturegli P, Lupi I, Landek-Salgado M, Kimura H, Rose NR 2008
Pituitary autoimmunity: 30 years later. Autoimmun Rev 7:631– 637
10. Karaca Z, Tanriverdi F, Unluhizarci K, Kelestimur F, Donmez H
2009 Empty sella may be the final outcome in lymphocytic hypophysitis. Endocr Res 34:10 –17
11. Naing S, Frohman LA 2007 The empty sella. Pediatr Endocrinol Rev
4:335–342
12. Sheehan HL, Summers VK 1949 The syndrome of hypopituitarism.
Q J Med 18:319 –378
13. De Marinis L, Bonadonna S, Bianchi A, Maira G, Giustina A 2005
Primary empty sella. J Clin Endocrinol Metab 90:5471–5477
14. Sage MR, Blumbergs PC 2000 Primary empty sella turcica: a radiological-anatomical correlation. Australas Radiol 44:341–348
15. Dan NG, Feiner RI, Houang MT, Turner JJ 2002 Pituitary apoplexy
in association with lymphocytic hypophysitis. J Clin Neurosci
9:577–580
16. Fujimaki T, Hotta S, Mochizuki T, Ayabe T, Matsuno A, Takagi K,
Nakagomi T, Tamura A 2005 Pituitary apoplexy as a consequence
of lymphocytic adenohypophysitis in a pregnant woman: a case
report. Neurol Res 27:399 – 402
17. Lee MS, Pless M 2003 Apoplectic lymphocytic hypophysitis. Case
report. J Neurosurg 98:183–185
18. Minakshi B, Alok S, Hillol KP 2005 Lymphocytic hypophysitis presenting as pituitary apoplexy in a male. Neurol India 53:363–364
19. Noordzij MJ, de Heide LJ, van den Berg G, Hoving EW 2005 [Pituitary apoplexy: an endocrinologic emergency]. Ned Tijdschr Geneeskd 149:2748 –2751 (Dutch)
20. Tanaka Y, Hirao T, Tsutsumi K, Miyashita T, Izumi Y, Mihara Y,
Ito M, Baba H, Migita K 1 June 2010 A case of apoplectic lymphocytic hypophysitis complicated by polymyalgia rheumatica. Rheumatol Int 10.1007/s00296-010-1535-x
4198
Lupi et al.
MRI Studies in Murine Hypophysitis
21. Chiovato L, Latrofa F, Braverman LE, Pacini F, Capezzone M, Masserini L, Grasso L, Pinchera A 2003 Disappearance of humoral thyroid autoimmunity after complete removal of thyroid antigens. Ann
Intern Med 139:346 –351
22. Okada K, Ishikawa S, Saito T, Kumakura S, Sakamoto Y, Kuzuya
T 1986 A case of partial hypopituitarism with empty sella following
normal course of pregnancy and delivery. Endocrinol Jpn 33:117–
123
23. Komada H, Yamamoto M, Okubo S, Nagai K, Iida K, Nakamura T,
Hirota Y, Sakaguchi K, Kasuga M, Takahashi Y 2009 A case of
hypothalamic panhypopituitarism with empty sella syndrome: case
report and review of the literature. Endocr J 56:585–589
24. Unlühizarci K, Bayram F, Colak R, Oztürk F, Selçuklu A, Durak AC,
Kele°timur F 2001 Distinct radiological and clinical appearance of
lymphocytic hypophysitis. J Clin Endocrinol Metab 86:1861–1864
25. Chada C, Seaquist ER 2009 An unusual case of lymphocytic hypophysitis in a young man presenting with elevated serum IGF-1.
Acta Endocrinol (Buc) 5:391–397
26. Nishiyama S, Takano T, Hidaka Y, Takada K, Iwatani Y, Amino N
1993 A case of postpartum hypopituitarism associated with empty
sella: possible relation to postpartum autoimmune hypophysitis.
Endocr J 40:431– 438
27. Komatsu M, Kondo T, Yamauchi K, Yokokawa N, Ichikawa K,
Ishihara M, Aizawa T, Yamada T, Imai Y, Tanaka K, Taniguchi K,
Endocrinology, November 2011, 152(11):4190 – 4198
28.
29.
30.
31.
32.
33.
Watanabe T, Takahashi Y 1988 Antipituitary antibodies in patients
with the primary empty sella syndrome. J Clin Endocrinol Metab
67:633– 638
Mau M, Phillips TM, Ratner RE 1993 Presence of anti-pituitary
hormone antibodies in patients with empty sella syndrome and pituitary tumours. Clin Endocrinol (Oxf) 38:495–500
Keda YM, Krjukova IV, Ilovaiskaia IA, Morozova MS, Fofanova
OV, Babarina MB, Marova EI, Pankov YA, Kandror VI 2002 Antibodies to pituitary surface antigens during various pituitary disease
states. J Endocrinol 175:417– 423
Bensing S, Rorsman F, Crock P, Sanjeevi C, Ericson K, Kämpe O,
Brismar K, Hulting AL 2004 No evidence for autoimmunity as a
major cause of the empty sella syndrome. Exp Clin Endocrinol Diabetes 112:231–235
Lupi I, Manetti L, Raffaelli V, Grasso L, Sardella C, Cosottini M,
Iannelli A, Gasperi M, Bogazzi F, Caturegli P, Martino E 27 May
2011 Pituitary autoimmunity is associated with hypopituitarism in patients with primary empty sella. J Endocrinol Invest 10.3275/7758
Sato N, Sze G, Endo K 1998 Hypophysitis: endocrinologic and dynamic MR findings. AJNR Am J Neuroradiol 19:439 – 444
Maghnie M, Genovese E, Aricò M, Villa A, Beluffi G, Campani R,
Severi F 1994 Evolving pituitary hormone deficiency is associated
with pituitary vasculopathy: dynamic MR study in children with
hypopituitarism, diabetes insipidus, and Langerhans cell histiocytosis. Radiology 193:493– 499
Sign up for eTOC alerts today
to get the latest articles as soon as they are online.
http://jcem.endojournals.org/subscriptions/etoc.shtml