ORIGINAL RESEARCH ARTICLE Open Access
Biomimetic Culture Reactor for Whole-Lung Engineering
Micha Sam Brickman Raredon,
1–3
Kevin A. Rocco,
1,2
Ciprian P. Gheorghe,
4
Amogh Sivarapatna,
1,2
Mahboobe Ghaedi,
2
Jenna L. Balestrini,
2,5
Thomas L. Raredon,
6
Elizabeth A. Calle,
1,2
and Laura E. Niklason
1,2,
*
Abstract
Decellularized organs are now established as promising scaffolds for whole-organ regeneration. For this work to
reach therapeutic practice, techniques and apparatus are necessary for doing human-scale clinically applicable
organ cultures. We have designed and constructed a bioreactor system capable of accommodating whole
human or porcine lungs, and we describe in this study relevant technical details, means of assembly and oper-
ation, and validation. The reactor has an artificial diaphragm that mimics the conditions found in the chest cavity
in vivo, driving hydraulically regulated negative pressure ventilation and custom-built pulsatile perfusion appa-
ratus capable of driving pressure-regulated or volume-regulated vascular flow. Both forms of mechanical actu-
ation can be tuned to match specific physio logic profiles. The organ is sealed in an elastic artificial pleura that
mounts to a support architecture. This pleura reduces the fluid volume required for organ culture, maintains the
organ’s position during mechanical conditioning, and creates a sterile barrier allowing disassembly and mainte-
nance outside of a biosafety cabinet. The combination of fluid suspension, negative-pressure ventilation, and
physiologic perfusion allows the described system to provide a biomimetic mechanical environment not
found in existing technologies and especially suited to whole-organ reg eneration. In this study, we explain
the design and operation of this apparatus and prese nt data validating intended functions.
Key words: tissue engineering; regeneration; extracellular matrix; bioprocessing
Introduction
The current standard for treating chronic terminal lung
disease is whole-organ transplantation. Although the pro-
cedure can give patients many more years of life, trans-
plantation is often delayed due to a shortage of viable
organs, and patients who do receive an organ must take
immunosuppressive drugs for the rest of their lives. An al-
ternative would be to grow new organs from a patient’s
own cells, and in 2010, two separate groups showed the po-
tential for ex vivo lung regeneration using decellularized
whole-organ scaffolds.
1,2
In these studies, lungs were
taken from a rat and perfused with detergent-based solu-
tions that removed cellular material. Such decellulariza-
tion techniques, when performed with care, leave the
extracellular matrix largely intact, preserving the micron-
scale histological architecture of the lungs and presenting
a fully vascularized whole-organ scaffold.
3–5
In principle,
these scaffolds can then be repopulated with cells derived
from the patient, yielding a transplantable artificial organ
with little or no immunogenicity.
6
Specialized bioreactors are essential for this type of
whole-organ engineering work, whether it be for lungs
or other organs such as livers.
7
For any organ type, re-
actors must provide adequate mass transport to all por-
tions of the tissue, control media characteristics such as
pH and oxygen, and should provide an organ-specific
mechanical environment that approximates the bio-
physical signals experienced in vivo.
8
In the lung,
shear stresses due to perfusion are important mediators
of endothelial phenotype,
9
and cyclic ventilation is
Departments of
1
Biomedical Engineering and
2
Anesthesia, Yale University, New Haven, Connecticut.
3
Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts.
Departments of
4
Obstetrics, Gynecology, and Reproductive Services, and
5
Pathology, Yale University, New Haven, Connecticut.
6
Raredon Resources, Inc., Northampton, Massachusetts.
*Address correspondence to: Laura E. Niklason, MD, PhD, Department Biomedical Engineering, Yale University, 10 Amistad Street, Room 301D, New Haven, CT 06511,
E-mail: laura.niklason@yale.edu
ª Micha Sam Brickman Raredon et al. 2016; Published by Mary Ann Liebert, Inc. This Open Access article is distributed under the terms of the Creative
Commons License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided
the original work is properly credited.
BioResearch Open Access
Volume 5.1, 2016
DOI: 10.1089/biores.2016.0006
BioResearch
OPEN ACCESS
72
known to affect organ development in utero and exerts
effects on epithelium during regeneration in vitro.
10,11
Any system for the culture of whole lung should there-
fore allow fine control of both breathing and perfusion,
in addition to maintaining sterility and allowing for
suitable nutrient and gas exchange.
The predominant commercially available apparatus
designed to work with human lungs are ex vivo lung per-
fusion (EVLP) systems, used in the pioneering clinical
work of the Lund grou p
12–16
and the Toronto group.
17 –19
In EVL P, lung s are taken fro m a donor, can nulated with
flow lines, placed under mechanical perfusion and ventila-
tion in the operating room, and monit ored before trans-
plant f or relevant outputs such as DP
O2
and v ascular
resistanc e. While this process mirrors some aspects of
the whole-o rgan regene ration process, EVLP systems are
not well suited for long-term tissue culture. Because
EVLP systems are designed for surgical use in an operating
room, they do not create a sterile barrier around the organ.
During procedures, the lungs are placed on a hard surface
and support their own weight, creating a mechanical envi-
ronment very different from that found in the body; when
in vivo, the lungs are suspended in fluid and held in place
by thin pleural membranes. In addition, these systems use
positive-pressure ventilators to breathe the lungs, and
there is substantial evidence that positive-pressure ventila-
tion can cause mechanical tissue damage and incite bio-
chemical cascades that negatively alter pulmonary cell
behavior and morphology.
20,21
Although the various de-
signs now used by laboratories for EVLP have some tech-
nical variation (see these references
22–27
), no currently
published EVLP apparatus is well suited for long-term
(e.g., more than 1 day) ex vivo whole-lung culture.
Based on the needs described above, we determined a
set of functional parameters for a regenerative whole-
lung bioreactor. Sterility was of primary importance,
and care was taken to build components from materials
that could be frequently autoclaved. We designed our sys-
tem to allow for sterile benchtop maintenance. Since ven-
tilator injury can be somewhat prevented by preserving
end-expiratory lung volume through a maintenance of
negative pleural pressure,
28
an overarching goal of this
project was to develop an apparatus capable of reliable
negative pressure ventilation. Suspending the cultured
lung in fluid encased by a ‘‘pleural membrane’ ’ was
deemed preferable, as it more closely mimics the condi-
tions found in the body. Previous studies with small an-
imal models suspended the entire regenerating organ in
the culture media. However, at human scale, this tech-
nique would have proven prohibitively expensive. We,
therefore, aimed to minimize media volume as much as
possible and to isolate media to contact with the tissue
itself. We also designed the system to be sufficiently com-
pact to be transportable between workstations. Most im-
portantly, the bioreactor had to be capable of mirroring
breathing and pulmonary vascular flow profiles corre-
sponding to physiologic profiles, as the utmost goal is
to provide a mechanical and chemical environment
that delivers appropriate cues to cultured lung cells.
The most important functional parameters are delineated
in Table 1.
Materials and Methods
Bioreactor design
The bioreactor consists of a ventilation apparatus, per-
fusion apparatus, and an organ culture chamber. The
breathing and perfusion units are integrated into a
wheeled cart that acts as a mobile platform for the
organ chamber. Technical drawings of the apparatus
are shown in Figure 1.
Table 1. Functional Parameters for Regenerative Ex Vivo Lung Culture
Parameter Required specifications Commercial EVLP systems
Sterility Fully sealed, sterile for long culture periods
(1–21 days)
Designed for operating room use, open to room
air, nonsterile
Ventilation method Negative pressure Positive pressure
Ventilation rate 0–15 breaths per minute Dependent on model and tissue used. Generally,
7–15 breaths per minute, 100 mL$kg
1
$min
1
0–750 mL per breath
Ventilation media Fluid or air Air
Perfusion method Pulsatile Continuous
Perfusion rate 0–90 beats per minute Dependent on model and tissue used. Maximum
flow generally considered 4 L/min0–50 mL per beat (up to 4.5 L/min)
Medium volume outside of tissue Minimal (<1 L), mechanism to maintain this volume
when culturing decellularized (permeable) tissue
0 L (used with intact native tissue)
Disassembly and storage Allow for benchtop-disassembly and sterile storage
of decellularized lung tissue
Not designed for manipulation or storage of lungs
outside of operating room
Local tissue environment Fluid suspension In contact with air, resting on platform
EVLP, ex vivo lung perfusion.
Raredon, et al.; BioResearch Open Access 2016, 5.1
http://online.liebertpub.com/doi/10.1089/biores.2016.0006
73
FIG. 1. Technical drawings of the bioreactor. The apparatus is composed of three mechanical modules: the
organ chamber (A), ventilation apparatus (B), and perfusion apparatus (C). These three modules are linked
through ½† internal diameter tubing (not shown), with flow paths dependent on experimental conditions. The
lung is sealed within a silicone pleura (A.7) and any cannulae linked to the cannulation port (A.8, drawn
schematically). The construct is then suspended from the removable stainless steel internal architecture (A.9)
through loops in the silicone pleura, and flow lines run from the cannulation port to the sealing ring (A.5).
Quick-connect fittings (A.6) allow easy management of flow lines to and from the reactor. Oscillation of the
hydraulic cylinder (B.5) drives fluid flow in/out of the top plate ventilation dome (A.1), causing flexing of the
artificial diaphragm (A.3). Expansion and contraction of the polyethylene bellows pump in the perfusio n
apparatus (C.4) drives cyclical pulsatile vascular flow. Ventilation can be varied from 0 to 15 bpm at 10–750 mL
per cyc le. Perfusion can be varied from 0 to 94 bpm at 5–55 mL per cycle. bpm, beats per minute.
74
The organ chamber consists of a 26.5 L glass jar
(Fig. 1, A.10), sealed under compression between a
top and a bottom plate (A.1 and A.11) clamped with
thumbscrews and drop bars (A.2). Handles allow for
maneuvering and transport of the filled chamber. The
top plate has eight ports for fluid flow, with sterile-
disconnect quick-connect fittings (A.4).
Directly beneath the top plate is a silicone membrane
(A.3) that generates the seal against the jar/organ cham-
ber. This membrane acts as an artificial diaphragm in
the reactor, forming a compliant wall for a variable-
volume fluid reservoir in the top plate. The membrane
is sealed to the top plate by a custom-machined polysul-
fone ring (A.5) that isolates the hydraulic fluid behind
the diaphragm from the interior of the jar, while allow-
ing fluid flow through the assembly.
Descending from the top plate assembly is a detach-
able stainless steel architecture (A.9) that helps to posi-
tion and orient the organ within the organ chamber,
providing anchor points for the perfused organ (A.6).
This structure doubles as a stand for the upper portion
of the organ chamber and the attached organ when dis-
assembled and worked with on a benchtop, allowing
manual access to the interior of the organ chamber
without physical disruption of the growing construct.
The organ itself is placed within a custom-fabricated
silicone pleura (A.7) that is sealed by a 4-inch diameter
cannulationport (A.8). The purpose of the silicone pleura
is severalfold. First, it creates an isolated media reservoir
immediately surrounding the tissue, which minimizes
culture medium requirements because it is not necessary
to fill the entire glass bioreactor (25 L) with the medium.
Second, it holds the organ in a desired position and ori-
entation within the reactor throughout decellularization,
seeding, and mechanical conditioning, even when sub-
merged and ventilated with air. And third, it generates
a sterile barrier around the organ, allowing the discon-
nection, storage, and transport of the lung outside of
the bioreactor pre- and post-culture. The artificial pleura
is highly elastic, made from Shore 10A hardness material
with 1,000% elongation at break, which allows it to ex-
pand and contract with the organ during perfusion and
ventilation. The cannulation port contains junctions for
arterial, venous, tracheal, and drainage/pressure adjust-
ment lines. This pleural assembly is one of the key com-
ponents of the design, which, along with the artificial
diaphragm, sets this design significantly apart from
other whole-lung perfusion apparatus.
The ventilation apparatus (B) cyclically pumps fluid
into and out of the fluid reservoir in the top plate (A.1).
Driven by a DC motor coupled to a hydraulic cylinder
through a variable-swing eccentric disc (B.2, B.4, and
B.5, respectively), the apparatus is capable of pumping
fluid volumes between 10 and 1,000 mL in a sinusoidal
pattern at rates between 1 and 15 cycles per minute,
thereby providing a means to breathe both human- and
large animal-sized lungs at physiological rates and vol-
umes. The expansion and contraction of the diaphragm
membrane (A.3) drives volume changes in the jar,
which subsequently cause the silicone pleura and encased
organ to expand or contract, pulling and pushing fluid
from an attached reservoir connected to the trachea.
Vascular perfusion is driven by a bellows pump
(C.1–C.6) that advances fluid around a vascular loop
that includes both a media reservoir and the vascular
conduits of the organ. The pump is capable of perfus-
ing 0–55 mL per stroke at rates between 0 and 94 cycles
per minute, thereby encompassing most human phys-
iological conditions for lung perfusion. Two one-way
silicone poppet/duckbill valves (C.3) direct media flow.
During operation, it is most useful to look at the reactor
as a set of five distinct fluid compartments. Supplemen-
tary Figure S1 shows a schematic that groups these five
compartments, and Supplementary Table S1 details the
specific location and purpose of each. The schematic
also includes various flow paths, relief valves, check
valves, and control consoles necessary for operation .
Bioreactor construction
Organ chamber.
Borosilicate glass (26.5 or 17 L;
Simax) jar (A.10) was purchased from Friedrich &
Dimmock. Top plate (A.1) and bottom plate (A.11)
were laser cut from 1/8† and 1/4† aluminum, respec-
tively. Diaphragm dome was hand turned on a lathe
out of aluminum sheet to match a machined mandrel
designed to have a volume of 0.75 L. Dome and top
cap foundation were welded together to create an air-
tight seal, and ½† national pipe thread (NPT) fittings
attached. All ports were designed to accommodate ½†
NPT to valved quick-connect fittings (A.4; Colder),
accommodating ½† internal diameter tubing. A dia-
phragm (A.3) was cut from 1/8† thick platinum-
cured silicone to match the top plate port pattern. A
sealing ring (A.5) was machined out of high-
temperature polysulfone (Radel) to hold this dia-
phragm to the top plate (A.1), which isolates the
space between the dome and the diaphragm while
still allowing unimpeded flow through the quick-
connect ports; this ring is secured against the top cap
by machine screws passing through small holes in the
Raredon, et al.; BioResearch Open Access 2016, 5.1
http://online.liebertpub.com/doi/10.1089/biores.2016.0006
75
diaphragm. The drop bars and thumbscrews (A.2) were
machined from brass and electroplated with nickel.
Stainless steel handles were attached to both the top
and bottom cap to allow physical transport of the bio-
reactor unit, and exposed aluminum pieces were pow-
der coated to prevent wear and oxidation (blue in
Figs. 2 and 5). The interior anchoring structure (A.9),
which would be in contact with fluid, was made either
from high-density polyethylene or 316 stainless steel. It
consists of four threaded rods that attach through
FIG. 2. Setup, perfusion, and ventilation. A lung is isolated, cannulated for arterial, venous, and tracheal flow,
attached to a cannulation port, and sealed within an appropriately sized silicone pleura (A). Corresponding
flow lines are attached, and the assembled construct is placed within the organ chamber and suspended from
the internal architecture (B). Arterial perfusion of the construct displays pressure profiles similar to those found
in vivo, complete with dicrotic notch valve closing (C). Venous outflow displays regular smooth pressure curves
(D). Pressure within the tra chea flow line oscillates due to ne gative-pressure ventilation (E). Pressure in the
arterial line during simultaneous perfusion and ventilation can be seen in (F), with high frequen cy oscillations
corresponding to 20 bpm perfusion and the lower sinusoidal modulation corresponding to 2 bpm ventilatio n.
Raredon, et al.; BioResearch Open Access 2016, 5.1
http://online.liebertpub.com/doi/10.1089/biores.2016.0006
76