Ultrastructural changes in acute lung allograft rejection:
Novel insights from an animal study
Wolfgang Jungraithmayr, MD,aAlice Draenert, MD,aKlaus Marquardt,band
Walter Weder, MDa
From theaDivision of Thoracic Surgery, University Hospital Zurich; andbCenter for Microscopy and Image Analysis, University of
Zurich, Zurich, Switzerland.
of the innate and adaptive immune system targeting the engrafted organ. The assessment and classi-
fication of pathologic changes of AR relies essentially on conventional histology. Herein we apply the
technique of scanning electron microscopy (SEM) to identify and characterize ultrastructural changes
of the pulmonary graft after lung Tx.
Orthotopic single-lung Tx was performed between BALB/c (donor) and C57BL/6 (recip-
ient) mice. At Day 5 after Tx, lung allografts were recovered for SEM and for histologic analysis.
Upon Tx, high numbers of leukocytes and thrombocytes were found, showing an activated
surface pattern and a change of their cell body shape. These cells adhered and partly transmigrated through
the endothelium of vessels. Larger vessels were more affected than smaller vessels and the endothelium was
roughened in its surface texture throughout. As a phenomenon, airways were partly covered by activated
dendritic cells. Numerous thrombocytes and macrophages accumulated on the endothelium of the cuff
anastomosis region exposing this area to a particularly higher risk of thrombosis.
SEM allows for detection of morphologic changes during pulmonary allograft rejec-
tion and adds important data to conventional histology when making the diagnosis of acute rejection.
J Heart Lung Transplant 2012;31:94–100
© 2012 International Society for Heart and Lung Transplantation. All rights reserved.
Acute rejection (AR) episodes after lung transplantation (Tx) are orchestrated by cells
Despite remarkable progress in research and treatment of
acute pulmonary rejection (AR), the understanding of the
underlying pathologic mechanisms during the course of AR
remains fragmentary. Unlike ischemia–reperfusion injury,
where cells of the innate immune system flood the graft,
acute rejection episodes are characterized by a strong adap-
tive immune response. These episodes are initiated and
maintained by key immune cells such as lymphocytes and
professional antigen-presenting cells, namely dendritic cells
(DCs), which trigger the immune response. The diagnosis of
AR commonly relies on histologic findings according to the
revised classification of the International Society for Heart
and Lung Transplantation (ISHLT).1
In contrast to conventional histology, the technique of
scanning electron microscopy (SEM) has been used to more
accurately characterize the morphology of lung compart-
ments and also of interactions between residing cells and
passenger leukocytes. SEM was employed to investigate the
pathophysiology of pulmonary cells in vitro,2but this tech-
nique was also used in in vivo studies of rat lung transplan-
tation (Tx).3Applying the orthotopic lung Tx model, the
investigators showed characteristic alterations of the allo-
graft that were induced by the allogeneic antigen stimulus.
Although orthotopic Tx of rat lung has served as a valuable
tool to study immunologic and non-immunologic events
after Tx, the emergence of the mouse lung Tx model4offers
the potential for studying a wide range of genetic modifi-
cations and transgenic models.
The purpose of this investigation was to analyze AR
phenomena in mice, specifically in orthotopically trans-
planted pulmonary allografts when compared with the right,
Reprint requests: Wolfgang Jungraithmayr, MD, Division of Thoracic
Surgery, University Hospital Zurich, Raemistrasse 100, 8091 Zurich, Swit-
zerland. Telephone: ?41-44-2558802. Fax: ?41-44-2558805.
E-mail address: email@example.com
1053-2498/$ -see front matter © 2012 International Society for Heart and Lung Transplantation. All rights reserved.
non-transplanted lung. This model is currently being ap-
plied more frequently in other studies. Elucidation of patho-
logic events of the AR process by additional techniques
could lead to a better understanding of the underlying im-
Animals and experimental setting
Specific pathogen-free inbred male mice (Fuellinsdorf, Zurich,
Switzerland), weighing 25 to 30 g, were used in this experiment.
Animals received adequate care according to The Principles of
Laboratory Animal Care (National Institutes of Health, promul-
gated in 1985, most recently revised in 1996). Orthotopic, alloge-
neic lung Tx (n ? 4) between C57BL/6 mice (H-2Kb) as recipients
and BALB/c mice (H-2Kd) as donors was performed as described
elsewhere.4For the vein, slip knots and clamps were used in an
alternating manner. The right, non-transplanted lungs served as
controls. The transplanted graft and the right, non-transplanted
lung were analyzed at Day 5 post-Tx.
Animals were intubated and anesthetized, and a laparosternotomy
was performed as described elsewhere.4Ventilation was main-
tained by applying a tidal volume of 0.8 ml with a positive
end-expiratory pressure of 1.5 mbar, and a respiratory rate of
120/min. The lungs were first flushed with 10 ml of 0.9% normal
saline solution at a pressure of 5 cm H2O via the pulmonary artery,
followed by perfusion of 10 ml of a 2% glutaraldehyde-containing
phosphate buffer solution based using the method described by
Fehrenbach et al5and Bachofen et al.6Ventilation was stopped
during glutaraldehyde fixation. Removal of the heart and lung was
performed en bloc with lungs inflated at two-thirds. The fixation
procedure was maintained for 24 hours by immersion in an equiv-
right, non-transplanted lung at Day 5 after Tx. The respiratory tract
was largely unaffected (A) (scale bar: 200 ?m) compared with the
non-transplanted, right lung (B) (scale bar: 100 ?m). Cilia were
covered by mucus plaques (C) (white arrow) (scale bar: 2 ?m),
which were less frequent in the right lung (D) (scale bar: 2 ?m).
Cilia displayed multiple vesicles along the shank of the cilia (E)
(scale bar: 200 ?m), and showed less expression in the right lung
(F) (scale bar: 1 ?m).
Bronchial tree from the transplanted (Tx) lung and the
(scale bar: 10 ?m). Long extensions, called filopodia, extended out of
the cell body (B) (white arrows) (scale bar: 3 ?m). Macrophages and
lymphocytes were numerous in the alveolar and interstitial space, also
showing podia extending out of the cell body (C) (white arrow) (scale
bar: 2 ?m).
Dendritic cells covering the epithelium of bronchi (A)
95 Jungraithmayr et al. SEM in Acute Lung Rejection
alent 2% glutaraldehyde-containing phosphate buffer solution, as
just described, and stored at 4°C.
Light microscopy and histology
Specimen preparation was accomplished by using a photo-documen-
tation system (Leica Z16 APO; Leica Microsystems, Heerbrugg,
Switzerland). The entire lung was cut into half in the frontal plane, in
both the transplanted and the non-transplanted lung. Sections were
taken from peripheral and central parts of the lung to ensure objective
and accurate sampling with equal probability of the samples. After
completion of the SEM process, specimens were embedded in paraf-
fin and stained for hematoxylin and eosin (H&E).
Scanning electronic microscopy
After glutaraldehyde perfusion, the whole heart–lung block was
soaked and stored in glutaraldehyde for 24 hours. Both the trans-
planted and the right, non-transplanted lung were cut into halves
through the frontal plane. Sections were dehydrated in ethanol,
using increasing concentrations from 10% to 100%, with four final
washing steps in 100% anhydrous ethanol. Each washing step
required 30 minutes. Next, critical-point drying was performed in
a critical-point dryer (CPD 030; Leica Microsystems). Dried sam-
ples were sputter-coated with 60 Å of platinum in a sputter-coating
apparatus (SCD 500; Leica Microsystems) and then analyzed by
SEM (Zeiss Supra 50 VP; Carl Zeiss NTS GmbH, Oberkochen,
Germany) at a 5- to 10-kV acceleration voltage using the SE2
Airways of Tx lung and right lung at Day 5
The diameter of the trachea and the bronchi were of similar
size when comparing the Tx lung (Figure 1A) with the right
lung (Figure 1B). One particular specimen showed the di-
ameter of the entire respiratory tree to be enlarged on both
sides. When looking at the surface of the main bronchi, the
segmental bronchi and also the bronchioles, ciliate cells were
identifiable, although rare. The cilia and bronchial epithelium
were covered by a mucus layer that in part hid the underlying
respiratory epithelium (Figure 1C). Between the ciliated cells,
these mucus plaques were loosely distributed and were cover-
ing and adhering to the cilia, so that the motility of the cilia in
the transplanted lung seemed to be impaired (Figure 1C, white
arrow) compared with the cilia of the right lung (Figure 1D).
Along the surface of the cilia, multiple vesicles appeared as a
sign of exocytosis and active secretion (Figure 1E). They were
more numerous than in the right lung (Figure 1F).
One of these specimens showed a large number of cells
with a specific cell-body shape characteristic of dendritic cells
(DCs), which were on top of and also embedded into the
epithelial layer (Figure 2A). The mean diameter of these DCs
was 30 ?m and they displayed multiple extensions, named
filopodia or pseudopodia (Pp), which were expanding out of
were found within the alveolar space, but lymphocytes were
alveolar wall (Figure 2C). They were partly flattened in shape
and showed long protrusions that were similar to the Pp seen
in DCs (Figure 3A), keeping in close contact with the base. In
contrast, almost none of these cells were found in the alveolar
space of the right lung (Figure 3B). Some of the Pp of mac-
rophages reached the surface of the alveolar epithelium and
were in close contact with these cells, particularly Type I
pneumocytes (PI) (Figure 3C). They seemed to interact with
and were attached to the epithelium by multiple adhesions
from filopodia and lamellipodia. Furthermore, small vesicles
developed and accumulated at the end of the Pp, similar to a
the end of the filopodia, they developed bunches of vesicles (white arrow). They left stripes behind, likely adhesion molecules (C) (dashed
white arrow), whereas the right lung revealed any immune cells within the alveolar space (B) (scale bars: 2 ?m).
Alveolar macrophages attached to pneumocytes and developed long filopodia (A) that extended out for pneumocytes (C). At
96 The Journal of Heart and Lung Transplantation, Vol 31, No 1, January 2012
cluster of grapes (Figure 3C, arrow). Macrophages left stripes
behind likely to be adhesion or chemotactic molecules (Figure
3C, dashed arrow). Hemotoxylin–eosin histology confirmed
the preserved epithelial lining and the accumulation of numer-
ous leukocytes around vessels in the transplanted graft
Although the morphology of the alveolar space appeared
intact, some portions of the alveolar structure were filled with
degradation. The alveolar surface was slightly shrunken when
compared with the right lung, where the alveolar structure
appeared normal (Figure 3B). Also, the inter-alveolar septa
macrophages were seen in the right lung at Day 5 post-Tx.
Endothelium of Tx and right lung at Day 5
The space around vessels, mainly arteries, was widened
(Figure 5A), and leukocytes, such as lymphocytes, were
numerous in this compartment when compared with the
right lung (Figure 5B). The surface of the endothelium of
the transplanted graft appeared roughened, rendering the
endothelium prone to adhesion of cells (Figure 5C, E, F),
whereas the surface of the endothelium of the right lung was
regular and smooth (Figure 5D). Here, the borders of the
endothelial cells appeared clear as a regular endothelial
lining (Figure 5G). Overall, on the endothelium of the right
lung, cells were rare (Figure 5B, D, G). Those that were
present did not adhere to the endothelium, although some of
these leukocytes even showed a change in surface pattern.
In contrast, numerous leukocytes accumulated on the endo-
thelium of the transplanted lung (Figure 5A, C, E, F). The
endothelium of the larger vessels (Figure 6A, B) in the graft
revealed larger numbers of adhering cells than did the
smaller vessels (Figure 7A–D).
appeared in different stages of activation and were thus distin-
guishable (Figure 8A). These cells consisted of mainly lym-
phocytes, but macrophages and dendritic cells were also pres-
ent. Some of these cells were in a state of adhesion (Figure 8B,
arrow), some transmigrated as doublets through the endothe-
lium (Figure 8B, dashed arrow), and others engaged in cell–
cell contact, possibly as antigen-presenting cells to lympho-
cytes (Figure 8C). Activated lymphocytes showed increased
activity characterized by the development of exocytotic vesi-
cles on their surface (Figure 8B, bold arrow). Of note, throm-
bocytes that changed cell-body shape became closely attached
to the endothelium (Figure 8C, arrow).
Different cells communicated with each other in a dis-
tinct manner. Immune cells were accompanied by thrombo-
cytes (Figure 9A, B) and appeared to be interacting with
each other (Figure 9A, arrow) while changing their cell-
body shape. Lymphocytes and also macrophages developed
lamellipodia, extending out to the endothelium and estab-
lishing cell–cell contact (Figure 9C, D). These cells also
diapedesed and transmigrated through the endothelium
(Figure 9E). Lymphocytes, or possibly dendritic cells pre-
senting antigen, extended filopodia (Figure 9F), and reached
out to interact with other immune cells.
Apart from the findings from the endothelium of intragraft
vessels, we found a high density of clustering thrombocytes
that were adherent on the recipient vessel wall at the site of
clamping and slip knotting (Figure 10A). These cells were
responding samples from SEM showed massive accumulation of
immune cells around a vessel, with the alveolar lining well pre-
served (scale bar: 200 ?m).
Histology, hematoxylin and eosin (H&E) stain of cor-
a slightly enlarged perivascular space (A) (scale bar: 100 ?m)
when compared with the right, non-transplanted lung (B) (scale
bar: 100 ?m). Increasing amounts of leukocytes accumulated on
the endothelium [scale bars: 10 ?m in (C), 2 ?m in (E), 2 ?m in
(F)]. However, endothelium of the right lung lacked any cells (D)
(scale bar: 20 ?m). Cell borders of the endothelial cells were
properly identifiable (G) (black arrow) (scale bar: 1 ?m).
Endothelium from the transplanted (Tx) lung showing
97Jungraithmayr et al. SEM in Acute Lung Rejection
additionally covered by numerous macrophages. These acti-
vated macrophages were positioned in gaps between a dis-
rupted layer of mucus on the endothelium (Figure 10B).
In this study we analyzed ultrastructural changes of acute lung
rejection in mouse lung transplants by SEM. Several important
details of allograft rejection were found distinct from conven-
tional histology, such as an accumulation of leukocytes and
thrombocytes, which appeared in different states of activation.
These cells adhered and partly transmigrated through the en-
dothelium, and the endothelial surface displayed a roughened
texture. While many cells were accumulating in large vessels,
smaller vessels were less affected. As a phenomenon, we
found a large number of maturing dendritic cells that covered
the respiratory epithelial lining, whereas the corresponding
right lung remained almost unaffected.
We chose Day 5 for the study of the allogeneic response
because, at this time-point, characteristic features of acute
rejection are detectable by conventional histology. Changes
of the alloimmune response at earlier time-points after Tx
are rather difficult to distinguish from phenomena related to
ischemia–reperfusion (I/R), as I/R injury dominates the
morphologic pattern during the first 24 to 28 hours after Tx.
Indeed, when looking at specimens on Day 1 after Tx, we
realized that the pathomorphology was similar between syn-
geneic and allogeneic transplants (data not shown). These
changes were clearly related to I/R injury in which immune
cells and thrombocytes showed low numbers, with a mildly
activated surface, less adherence to the endothelium and less
cell–cell interaction. Also, the endothelial lining was less
roughened when compared with Day 5 after Tx, and the
surface of the endothelium displayed more microvilli, a
phenomenon almost absent at Day 5.
At Day 5, we expected that cells of the adaptive immune
system would be increasingly accumulating and would thus
replace cells of the innate immune system, such as neutrophils
and macrophages. However, large numbers of leukocytes,
likely macrophages, were still seen in these specimens on the
endothelium but also in the alveolar space, suggesting that
the innate immune system is still involved in modulating of the
rejection response. These macrophages displayed protrusions
that are recognized as Pp, as a sign of activation, and were thus
morphologically clearly distinguishable from lymphocytes.
Lymphocytes develop lamellipodia that can extend to pseudo-
podia, macrophages instead show filigree and slender filopo-
dia. In contrast to resting macrophages, activated macrophages
and thrombocytes that were adherent to endothelium [scale bars:
30 ?m in (A), 10 ?m in (B)].
Large vessels revealed large numbers of leukocytes
bars: 20 ?m in (A), 2 ?m in (B)], and less tendency toward
adherence and transmigration of cells [scale bars: 10 ?m in (C), 2
?m in (D)].
Small vessels showed scattered cell distribution [scale
types (A) (scale bar: 1 ?m). Locomotion of these cells revealed
firm adherence to the endothelium (B) (arrow) (scale bar: 2 ?m),
transmigration through the endothelium (B) (dashed arrow) (scale
bar: 2 ?m), exocytosis by an immune cell (B) (bold arrow) (scale
bar: 2 ?m), and close interaction between two cells, but also an
attachment of thrombocytes to endothelium (C) (arrow) (scale bar:
Leukocytes appeared in different activation pheno-
98The Journal of Heart and Lung Transplantation, Vol 31, No 1, January 2012
change their morphology in response to pathogens by flatten-
ing and by increasing the number of filopodia and cell surface
ruffles, as shown in an in vitro study.7This observation was
also shown in a model of Type II cell–mediated macrophage
activation.8We detected these filopodia that extended out to-
ward cells of the respiratory epithelium (Type II cells) to gain
contact with neighboring cells for possible antigen uptake or
presentation. The reason for the formation of the observed
vesicle accumulation at the end of the Pp remains speculative,
but it is most likely due to exocytosis for cytokine release.
In contrast to macrophages, lymphocytes showed a dif-
ferent pattern of roughened surface as a sign of activation
and were found on the endothelium in different morpho-
logic and locomotion states. On the one hand, folding of the
lymphocyte surface is a sign of changed motility9; on the
other hand, it helps the cell to arrest, adhere and eventually
transmigrate through the endothelium to induce a specific
local inflammatory immune response.10
As platelets were increasingly found to be adherent to the
altered surface of the endothelium, the formation of partial
thrombosis seems to be a reasonable explanation for the de-
velopment of thrombi within small vessels that was observed
in some areas of the engrafted organ. To our surprise, throm-
bocytes also changed their surface phenotype and got in close
contact with other immune cells. It remains speculative as to
whether these interactions help in the formation of thrombi or
if they participate in the promotion of the rejection process.
The area where the vessel was clamped for outflow occlusion
during completion of the anastomosis was also prone to a
dense accumulation of thrombocytes, regardless of whether a
slip knot or a clamp was used for occlusion.
Under steady-state conditions, pulmonary DCs reside in
respiratory airways between the basal membrane and the
epithelial layer. The maturation of DCs results in pheno-
typic changes, linked to enhanced capacity to process anti-
gens and to prime T cells. These phenotypic changes in-
clude the protrusion of the Pp between the respiratory
epithelial cells into the airway to gain access to the luminal
site. This phenomenon was nicely shown in a triple-cell
co-culture model of epithelial airway barrier.11In one of our
specimens, we found DCs to move from their original lo-
cation toward the surface of the respiratory epithelial lining,
suggesting either an advanced state of activation as reflected
by increased mobility, or an apparent dislocation due to a
vanishing of ciliated and goblet cells. However, we cannot
rule out the possibility that this phenomenon developed as
an artifact of the transplanted specimen. Accordingly, the
respiratory epithelium could be lost due to a lack of ade-
quate arterial perfusion and as a result of DCs migrating
toward the vanishing epithelial layer. The enlargement ob-
served in the airways was likely a result of mild overpres-
sure of ventilation prior to heart–lung removal. Human
donor lungs are usually kept inflated at a mild positive
airway pressure to ensure some residual oxygen supply.
contact and even attachment of thrombocytes to larger leukocytes
(A) (arrow) (scale bar: 1 ?m), development of lamellipodia search-
ing for contact to the endothelium [scale bar: 1 ?m in (B), 300 nm
in C)], interacting immune cells [scale bar: 1 ?m in (D)] or
immune cells that were transmigrating [scale bar: 2 ?m in (E)],
and extending filopodia out to allow contact [scale bar: 1 ?m in
(F)]. Images (B) and (C) are magnifications of Figure 5E, and (F)
is a magnification of Figure 5F.
Detailed view of cell–cell interactions showed close
outflow occlusion. A field of thrombocytes covered the endothelium
of this particular portion of vessel (A) (arrow) (scale bar: 10 ?m) that
was interspersed with macrophages (B) (arrow) (scale bar: 2 ?m).
Endothelium from vessels at the region of clamping for
99Jungraithmayr et al.SEM in Acute Lung Rejection
Compared with human tissue, mouse lung airways are more
fragile and less resistant to minor changes in ventilation
pressure, giving rise to further extension of airways.
In conclusion, some of the results from this study coincide
with features of conventional histology. In contrast, some of
our other findings, such as the differential activation status of
leukocytes and the presence of thrombocytes and their inter-
action, have not been reported so far and require further study.
When considering the SEM approach as a supporting tech-
nique, adding precise and detailed changes to histology, the
current findings may help to better understand the phenomena
of acute rejection in murine allograft rejection.
The authors have no conflicts of interest to disclose. The first two
authors (W.J. and A.D.) contributed equally to this study. We
thank Urs Ziegler and Andres Kaech for critically reading the
manuscript and for their helpful suggestions.
1. Stewart S, Fishbein MC, Snell GI, et al. Revision of the 1996 working
formulation for the standardization of nomenclature in the diagnosis of
lung rejection. J Heart Lung Transplant 2007;26:1229-42.
2. Chowdhury F, Howat WJ, Phillips GJ, et al. Interactions between
endothelial cells and epithelial cells in a combined cell model of
airway mucosa: effects on tight junction permeability. Exp Lung Res
3. Semik M, Schnabel R, Bruske T, et al. Ultrastructural studies of acute
rejection following single lung transplantation in the rat—histological
and immunohistological findings. Thorac Cardiovasc Surg 1994;42:
4. Jungraithmayr WM, Korom S, Hillinger S, et al. A mouse model of
orthotopic, single-lung transplantation. J Thorac Cardiovasc Surg
5. Fehrenbach H, Schepelmann D, Albes JM, et al. Pulmonary ischemia/
reperfusion injury: a quantitative study of structure and function in
isolated heart–lungs of the rat. Anat Rec 1999;255:84-9.
6. Bachofen H, Ammann A, Wangensteen D, et al. Perfusion fixation of
lungs for structure–function analysis: credits and limitations. J Appl
7. Haranaga S, Yamaguchi H, Ikejima H, et al. Chlamydia pneumoniae
infection of alveolar macrophages: a model. J Infect Dis 2003;187:
8. Farberman MM, Demello DE, Hoffmann JW, et al. Morphologic
changes in alveolar macrophages in response to UVEC-activated pul-
monary Type II epithelial cells. Tissue Cell 2005;37:213-22.
9. Nobile C, Rudnicka D, Hasan M, et al. HIV-1 Nef inhibits ruffles,
induces filopodia, and modulates migration of infected lymphocytes.
J Virol 2010;84:2282-93.
10. Campbell JJ, Butcher EC. Chemokines in tissue-specific and microen-
vironment-specific lymphocyte homing. Curr Opin Immunol 2000;12:
11. Gehr P, Blank F, Rothen-Rutishauser BM. Fate of inhaled particles
after interaction with the lung surface. Paediatr Respir Rev 2006;
100 The Journal of Heart and Lung Transplantation, Vol 31, No 1, January 2012