Differentiation of precursors into
parathyroid-like cells for treatment
of hypoparathyroidism
Kathleen M. Woods Ignatoski, PhD, Eve L. Bingham, BS, Lauren K. Frome, and
Gerard M. Doherty, MD, Ann Arbor, MI
Background. Hypoparathyroidism is the most frequent permanent complication of thyroid surgery. Our
hypothesis is that human precursor cells in culture can be differentiated into parathyroid cells and used
to reconstitute function. Human embryonic stem cells (hESCs) are a stable model to study differentiation
into parathyroid-like cells. In prior work, the BG01-hESC line was stimulated to form parathyroid-like
cells. This cell line is no longer available, however, and additional studies were needed to confirm and
extend prior observations.
Methods. Increasing concentrations of fetal bovine serum and timed exposure to Activin A were used to
differentiate H1-hESC into parathyroid-like cells. The potential benefit of Sonic hedgehog exposure on
parathyroid-like cell development also was evaluated by serial alterations of culture conditions. Calciumsensing receptor (CaSR), GCM2, and PTH expression (RT-PCR) and PTH protein secretion (ELISA)
were used as markers of differentiated cells.
Results. We successfully modified our prior protocol to generate cells that express CaSR, GCM2, and
PTH RNA from undifferentiated H1-hESC. The cells also secreted PTH.
Conclusion. We replicated parathyroid differentiation using H1-hESC cells. Our data advance the
project toward in vitro differentiation of precursor cells isolated from individual patients for
autotransplantation. (Surgery 2010;148:1186-90.)
From the Department of Surgery, Division of Endocrine Surgery, University of Michigan Health System, Ann
Arbor, MI
BECAUSE THE PARATHYROID GLANDS AND THEIR BLOOD
SUPPLY can be difficult to identify and preserve in
situ, loss of parathyroid gland function (hypoparathyroidism) is the most frequent, permanent complication of thyroid and parathyroid surgery. Thus,
even though prevention of hypoparathyroidism by
autografting damaged parathyroids is widely recognized as the best approach, the rate of permanent
hypoparathyroidism after thyroid surgery is nearly
14% in population-based series.1,2 The chronic, detrimental effects of hypoparathyroidism on bone,
teeth, skin, and nails are well documented.3 Hypoparathyroidism results in chronic hypocalcaemia
and low-turnover bone disease that can be palliated
by multiple daily doses of vitamin D analogues and
calcium. The condition, however, is poorly managed by currently available replacement methods.
Accepted for publication September 16, 2010.
Reprint requests: Kathleen M. Woods Ignatoski, PhD, Department of Surgery, Division of Endocrine Surgery, University of
Michigan Health System, A556 MSRB2, 1500 W. Medical Center
Drive, Ann Arbor, MI 48109-5654. E-mail: [email protected].
0039-6060/$ - see front matter
Ó 2010 Mosby, Inc. All rights reserved.
doi:10.1016/j.surg.2010.09.021
1186 SURGERY
Replacement of parathyroid hormone (PTH) itself
is available with teriparatide (Forteo; Eli Lilly and
Company, Indianapolis, IN), a 1-34 N-terminal protein fragment of PTH; however, the serum half-life
of synthetic PTH is less than 5 minutes, providing
ineffective replacement therapy.
We began our parathyroid replacement studies
by using human embryonic stem cells (hESCs) as a
stable model system. We induced BG01-hESCs to
differentiate into cells that expressed the parathyroid markers PTH, calcium-sensing receptor
(CaSR), CXCR4, and GCM2. The differentiated
cells also secreted intact PTH as determined by
enzyme-linked immunosorbent assay (ELISA).4
Concurrent to our development of differentiating procedures, the BG01-hES cell line was determined to not have the appropriate informed
consent approval, and our university recommended ending research that relied on these cells.
Testing with BG01 cells was therefore terminated
after the publication of our article.4 Here we present data with another hESC line (H1 from WiCell)
that validates our original data with BG01 cells and
extends those observations to include in vitro effects of Sonic hedgehog exposure. Our ability to
differentiate 2 independent hESC cell lines into
Surgery
Volume 148, Number 6
parathyroid-like cells using our published protocol
strongly suggests our approach will be useful to
produce parathyroid cells from individual patient
progenitors and allows for continued use of a
stable hESC model for additional research.
MATERIALS AND METHODS
Cell culture. Undifferentiated cells were cultured
as we previously described.4 Briefly, hESC lines were
maintained on feeder layers of mouse embryo fibroblasts (MEFs) (Global Stem, Inc., Rockville, MD) on
0.1% gelatin. Cells were cultured in hESC culture
medium (DMEM/F12; 20% Knockout Serum Replacer [KOSR]; Invitrogen, Carlsbad, CA), nonessential amino acids, 2-mmol/L L-glutamine, 0.1-mmol/L
b-mercaptoethanol, and 4-ng/mL basic fibroblast
growth factor (bFGF). Cultures were fed daily and
mechanically passed every 3--4 days.
Differentiation protocols. H1 cells were differentiated using the Bingham Protocol (Fig 1).4 Briefly,
cells were kept on MEFs and were fed Activin A and
increasing amounts of fetal bovine serum (FBS) for
5 days. Cells were cultured on or off the MEF feeder
layer and were fed 5% FBS and Activin A ± Sonic
hedgehog (Shh) for an additional 7 days. Cells
were then kept in culture for an additional 2 weeks
± Activin A ± Shh. Cells were tested for marker expression and PTH secretion at each major time point in
the protocol.
PCR. Total RNA was isolated using TRIzol reagent, and cDNA synthesis was performed using the
ReactionReady First Strand cDNA synthesis kit
(SuperArray, Frederick, MD). PCR reactions were
performed using primer sets we have previously
published4 that span exon splicing sites. GAPD internal standards were used in each reaction. The
PCR protocol was 1 cycle at 958C for 15 minutes,
then 30 cycles of 15 seconds at 958C, 30 seconds at
558C, and 30 seconds at 728C with a final extension
of 7 minutes at 728C.
ELISAs. At each major time point of the differentiation protocol, conditioned media were collected and stored at --808C until use. Media were
used in 2 different commercial PTH ELISA kits (Cat
#DSL-10-8000, Diagnostic Systems Laboratories,
Brea, CA; and Cat#PT019T, Calbiotech, Spring
Valley, CA) since the DSL kit became difficult to
obtain. TSH, Thyroxine (freeT4), and calcitonin
from conditioned media also were assayed by ELISA
(Cat# TS045T; Calbiotech; Cat# F4107T; Calbiotech, Cat# 40-056-205003; and GenWay Biotech,
respectively). The cell number was not obtained as
the conditioned media were taken from cells used
in the next step of the differentiation process. All
samples were tested in triplicate.
Woods Ignatoski et al 1187
RESULTS
Previously, we were able to differentiate BG01hESCs cells into parathyroid-like cells that expressed
parathyroid markers GCM2, CaSR, and PTH and that
also secreted PTH.4 As our ultimate goal is to use cells
from an individual, hypoparathyroid patient and
differentiate them into parathyroid cells for replacement, we wanted to show that we could differentiate
more than 1 cell line. Thus, we determined whether
the Bingham protocol could also differentiate the
H1 hESC into PTH-producing parathyroid-like cells.
H1 hES cells were subjected to the Bingham
protocol ± the differentiation factor Shh. Cellular
RNA was isolated at all of the major time points of the
differentiation protocol and characterized by RTPCR to determine the expression of parathyroid
markers. Differentiated H1 cells expressed all of the
parathyroid markers (GCM2, CaSR, CXCR4, and
PTH) at 1 week posttreatment with Activin A (Fig 2),
1 week earlier than the BG01 cells did.4 The addition
of Shh caused the cells to express more of the
markers at the same time point.
Media from the differentiated cells cultured
both on and off MEFs were tested in commercial
PTH ELISAs for i-PTH secretion. The most secreted PTH (Fig 3) was observed at the same time
point as PTH peaked in the PCR analysis (Fig 2),
2 weeks after Activin A treatment. The addition
of Shh caused PTH to be released earlier, at
1 week post--Activin A/Shh treatment. The earlier
release of PTH after treatment with Shh also was
observed previously with the BG01 cells (data not
shown). More PTH was secreted when the cells
were cultured throughout the differentiation protocol with MEFs (Fig 4). The cells had the same appearance as the in vitro differentiated BG01 cells
did.4
As we used general differentiation factors, it also
is conceivable that the cells are producing factors
specific to other endocrine organs produced from
pharyngeal endoderm, although we used assessment of parathyroid-specific transcription factors
to guide the development of the protocol. Thus,
we performed ELISAs for TSH, T4, and calcitonin.
The differentiated cells did not produce TSH, T4,
or calcitonin (data not shown).
DISCUSSION
Mechanical damage to the parathyroids is most
often a result of operations on the thyroid. The
nature of hypoparathyroidism along with the simplicity of the parathyroid glands makes hypoparathyroidism an ideal candidate for treatment by
cellular replacement.
1188 Woods Ignatoski et al
Surgery
December 2010
Fig 1. Bingham protocol. Schematic representation of the protocol used to differentiate hESC in vitro. We used a modified D’Amour protocol to drive the early endoderm differentiation. The modifications include increasing FBS instead
of decreasing it, while adding Activin A (Act A). After 5 days, we replated the cells without mouse embryo fibroblasts
(MEFs) in 5% FBS for 1 week. We then removed the ActA and kept the cells in culture. For some cultures, we also added
Sonic hedgehog (Shh). At all major time points, cells and conditioned medium were collected for RT-PCR analysis and
ELISAs, respectively.
Fig 2. Expression of parathyroid markers from H1 hES
cells differentiated with Activin A or with Activin A/Shh.
RT-PCR was performed at all major time points of the
differentiation scheme on cells from day 5 until 2 weeks
post-treatment. By 1 week post treatment with Activin A
and Shh, the cells expressed the definitive endoderm
marker Sox 17, the epithelial, marker CXCR4, and the
parathyroid markers CaSR, PTH and GCM2.
To begin to understand how to differentiate cells
into parathyroid cells in vitro, we treated BG01
hESC with a combination of increasing FBS and
timed Activin A exposure.4 The differentiated cells
expressed the parathyroid markers CaSR, GCM2,
and PTH, and they secreted PTH. Our overall
objective, however, is to isolate precursor cells
from patients, differentiate them in vitro, and return the differentiated cells as an autograft. Thus,
it is important to know that we could differentiate
more than 1 cell line. To this end, we repeated our
experiments with the H1 hESC. The results presented here show that our differentiation protocol
has the potential to be used to differentiate cells
originating from individual patients into PTHsecreting cells.
Fig 3. PTH secretion from differentiated H1 cells. Conditioned medium was collected from H1 hESC undergoing differentiation at each of the same major time points
as were used for Fig 1 and assessed by i-PTH ELISA. PTH
was secreted at the same time points at which the cells
expressed parathyroid markers (Fig 1).
In current clinical practice, if parathyroid tissue
is damaged during thyroid surgery, it is devascularized and mechanically disrupted into small bits of
tissue that can survive on diffused nutrients at the
autograft site until neovascularization occurs. Resumption of measurable, normal parathyroid function occurs in 6--10 weeks postgrafting for patients
who otherwise have no endogenous parathyroid
function.6 The rate of complete normalization of
parathyroid function after fresh autograft of parathyroid tissue is >90% and is approximately 50%
using cryo-preserved tissue,5,6 and the grafted parathyroid tissue responds normally to changes in
serum calcium concentration. Unfortunately, unless damage to the parathyroids is recognized in
the operating room, autografting is not possible.
Many challenges of organ replacement therapy
are avoided during parathyroid replacement because of the nature of the parathyroid glands.
Parathyroid glands are optimal for cellular replacement therapy because: (1) each parathyroid cell
Woods Ignatoski et al 1189
Surgery
Volume 148, Number 6
Fig 4. PTH secretion from differentiated H1 cells. Conditioned medium was collected from H1 hESC undergoing differentiation at each of the same major time points
as were used for Fig 1 and assessed by i-PTH ELISA. For
this assay, the cells were left on MEFs for the entire protocol. This ELISA was from a different manufacturer
than the one used for Fig 3.
contains the complete function of the organ; (2) no
architectural arrangement of parathyroid cells is
needed to support or enhance the function of the
organ; (3) transplantation of a small number of
parathyroid cells reconstitutes normal parathyroid
function; and (4) patients who have lost this function because of operative complication have no
autoimmune reactivity to their parathyroid tissue.2,5-8 Thus, the parathyroid gland is straightforward to graft and restore function, if the cells are
available to do so. The simplicity of the parathyroid
glands focuses the treatment challenges on the
development of the replacement cells themselves
and not on the development of glandular architecture, making the parathyroids ideal for in vitro organogenesis. To this end, we propose to replace
damaged parathyroids with autologous grafts of
cells that are engineered in vitro without genetic
manipulation to replace parathyroid function.
Interestingly, general differentiation factors,
Activin A and Shh, were useful to provide a specific
differentiation phenotype, the secretion of PTH
and expression of CaSR. Activin A has been used to
differentiate other pleuripotent cells into bone;
however, the concentration used to produce bone
is greater than the concentration we used to
produce parathyroid-like cells. Taken together,
these data suggest that a gradient of concentration
and timing of Activin A may be used to dictate
specific differentiation outcomes.
Parathyroid cells release PTH in response to
changing serum calcium concentrations to balance
appropriate serum levels of ionized calcium, with
solid form calcium salts used for bone structure. A
phenotype of parathyroid cells in culture is their
ability to release PTH in response to varying
calcium levels. Our differentiated cells released
only modest amounts of PTH and were grown in a
medium that had moderate levels of calcium. If
our cells are truly responsive to calcium concentrations, then we would expect to observe only
modest levels of PTH secreted at those levels of
calcium, which is what we observed. Future studies
will incorporate calcium concentration-dependent
PTH analysis.
Because general differentiation factors were used,
another possibility is that the cells producing PTH
also may produce other factors specific to the endocrine organs derived from the pharyngeal endoderm, such as thyroid hormone or thymus markers.
Production of these factors, particularly of thyroid
hormone, could prove therapeutic for other diseases. The thyroid markers, TSH, T4, and calcitonin
were not produced by our cells, however, indicating
that the cells were not differentiated into thyroid.
This specificity of parathyroid-like cell development
also validates our approach of optimizing the
appearance of parathyroid-specific transcription factors in the evolution of the Bingham differentiation
protocol.
The data presented here bring us closer to
autologous cellular replacement therapy for hypoparathyroidism. Work is ongoing to assess whether
the cells can compensate for PTH deficiency in an
animal model.
REFERENCES
1. Scott-Coombs D, Kinsman R. Second National Audit Report.
In: British Association of Endocrine Surgeons, editors. Parathyroid: outcomes, Vol. 1. Oxfordshire, UK: Henley-onThames; 2007. p. 110-2.
2. Doherty GM. Complications of thyroid and parathyroid surgery. In: Mulholland MW, Doherty GM, editors. Complications
in surgery. Philadelphia: Lippincott, Williams & Wilkins; 2006.
3. Rubin M, Dempster DW, Zhou H, Shane E, Nickolas T, Sliney
J Jr, et al. Dynamic and structural properties of the skeleton
in hypoparathyroidism. J Bone Miner Res 2008;23:2018-24.
4. Bingham EL, Cheng S-P, Ignatoski KMW, Doherty GM. Differentiation of hES cells to a parathyroid-like phenotype.
Stem Cell Develop 2009;18:1071-80.
5. Doherty GM, Skogseid BM, editors. Surgical endocrinology.
Philadelphia: Lippincott, Williams & Wilkins; 2001.
6. Olson JA Jr, DeBenedetti MK, Baumann DS, Wells SA Jr. Parathyroid autotransplantation during thyroidectomy. Results
of long-term follow-up [See comment]. Ann Surg 1996;223:
472-8; discussion 478-480.
7. Lo CY. Parathyroid autotransplantation during thyroidectomy [See comment]. ANZ J Surg 2002;72:902-7.
8. Wells SA Jr, Gunnells JC, Shelburne JD, Schneider AB, Sherwood LM. Transplantation of the parathyroid glands in man:
clinical indications and results. Surgery 1975;78:34-44.
DISCUSSION
Dr Bradford Mitchell (Morgantown, WV): I assume the
assay that you are using is intact PTH?
Dr Kathleen M. Woods Ignatoski (Ann Arbor, MI): Yes.
Dr Bradford Mitchell (Morgantown, WV): You are
showing RT-PCR and mRNA and what looks to me like
1190 Woods Ignatoski et al
constitutive secretion of PTH. Have you looked at it in
varying concentrations of calcium because otherwise
this would not be a great model. If it is not responding
to calcium, it would be a great concern.
Dr Kathleen M. Woods Ignatoski (Ann Arbor, MI): We
started looking at calcium with the BGO-1 cells. And before we could optimize the conditions, we were told we
could no longer use the BGO-1 cells, and at the same
time, we were trying to switch and use another cell
line, the H-1s. We got the H-1s up and going, and then
we decided we should just go for it and try the thymus
because we received similar results.
So, right now we are just using thymus. And the next
step is to see whether we have calcium regulation, but we
have not done that yet.
Dr Bradford Mitchell (Morgantown, WV): And the
next suggestion would be to look at a calcium-sensing receptor stain to see that not only is it in the cell but that
also it is expressed appropriately on the surface.
Dr Kathleen M. Woods Ignatoski (Ann Arbor, MI): We
actually just got a thymus the other day before I left. And
we had that slide prepared, but we just have not been
able to do that yet.
Dr Mark Cohen (Kansas City, KS): I have a question
about 1 of the statements you made about parathyroid
and immunogenicity lacking of autografts. And so the
question I have is, are you planning to take thymic tissue
from the same patient and then deliver it to that patient
later after the development? Or are you taking thymic
tissue from an unknown source and then delivering it?
Dr Kathleen M. Woods Ignatoski (Ann Arbor, MI):
No, our ultimate goal is to take it from the same patient.
Dr Jennifer Rosen (Boston, MA): A few questions:
The first is, have you thought about using calciumsensing receptor to do live-cell sorting in your thymus?
And the second question is, have you tried taking
parathyroid tissue from patients and growing it in
different concentrations of calcium, using that as, say,
a growth factor for the cells?
Dr Kathleen M. Woods Ignatoski (Ann Arbor, MI): For
your first question, we are thinking about it. We don’t want
to flow sort. We are worried about if the flow sorting would
damage the cells in some way. So, we are trying to figure
Surgery
December 2010
out a way to affinity-purifying the cells using calciumsensing receptor and magnetic beads? And we have only
tried that once. We did get a population of cells out, but
then we have to figure out how to grow them from there.
Regarding your second question, using conditioned
media as a growth factor, that was very intriguing. We have
not done that. We would have to get parathyroid tissue to
do it, but we have not done it. It would be interesting.
Dr James Lee (New York, NY): The 2 questions I have,
number 1, have you thought about looking at fat-derived
stem cells? It seems like it would be an easier source of
cells to harvest, but also many people have had a lot of
good results de-differentiating them and then taking
them down the neuroendocrine line.
The second question, as always with stem cells, what is
the durability of these stem cells? How many passages
can you get from these cells?
Dr Kathleen M. Woods Ignatoski (Ann Arbor, MI):
Let me go back to the second question first. The durability is, what we saw there was at a week, at the end, when
we pulled out the thymus, that was only a week we received expression of PTH. A month later, we did not
have any expression of PTH, so we need to figure out
the cell biology to keep the PTH expression.
Regarding your second question on fat-derived stem
cells, that is our alternative step in our grant to use. We
decided to go with the thymus because it was from the
same primordium. We thought it would be easier to pull
them out from there, but we could go back and use
fat-derived stem cells.
Dr Herb Chen (Madison, WI): I will take the privilege
of the last question, and it has to do with sonic hedgehog
signaling. Did you look at downstream markers like Lee
1 when you introduced the Sonic hedgehog? And is the
whole process of differentiation dependent on, like, did
you treat with an inhibitor like cyclopamine?
Dr Kathleen M. Woods Ignatoski (Ann Arbor, MI): We
have not looked at any of the downstream markers. Our
whole goal was to try and drive the cells to produce PTH
and then also produce PTH in a calcium-sensing manner. And we have not looked to see how any of those
drugs were acting. We just were trying to manipulate
the cells. So we have not done any of the signaling.