Supplementary Appendix

This appendix has been provided by the authors to give readers additional information about their work.

Supplement to: Ackermann M, Verleden SE, Kuehnel M, et al. Pulmonary vascular endothelialitis, thrombosis,
and angiogenesis in Covid-19. N Engl J Med. DOI: 10.1056/NEJMoa2015432
Pulmonary Vascular Endothelialitis, Thrombosis and Angiogenesis
in COVID-19

Supplementary Appendix: Table of Contents

Investigators 2
Materials and Methods
 Autopsy Technique 3
 µCT Analysis 3
 Histopathologic Analysis 3
 Protein Expression Analysis 4
 Transmission Electron Microscopy 4
 Vascular Corrosion Casting, Scanning Electron Microscopy, and Morphometry 5
 Gene Expression Analysis 5
 Statistical Analyses 5
Supplementary Figures
 Figure S1. Relative expression analysis of inflammation-associated genes
in COVID-19 and influenza (A/H1N1) lungs compared to uninfected control
lungs 8
 Figure S2. Histopathology of Influenza (A/H1N1) Lungs 9
 Figure S3. COVID-19-associated thrombosis 10
 Figure S4. Schematic of intussusceptive and sprouting angiogenesis 11
 Figure S5. Anaglyphic 3D-image of a scanning electron micrograph
highlighting the changes in a COVID-19 lung 12
Supplementary Tables
 Table S1A. Clinical, Radiological and Histopathological
Characteristics of COVID-19 Patients 13
 Table S1B. Clinical and Radiological Characteristics
of Influenza (A/H1N1) Patients 15
 Table S2. Relative Counts of ACE2 Positive Cells
of COVID-19, Influenza (A/H1N1) and Control Cases 17

Gene expression data 17

References 18

1
Investigators

Maximilian Ackermann MD,

Stijn E. Verleden PhD,

Mark Kuehnel PhD,

Axel Haverich MD,

Tobias Welte MD,

Florian Laenger MD,

Arno Vanstapel PhD,

Christopher Werlein MD,

Helge Stark PhD,

Alexandar Tzankov MD,

William W. Li MD,

Vincent W. Li, MD, MBA,

Steven J. Mentzer MD, *

Danny Jonigk MD

*Corresponding author.

2
Materials and Methods

To facilitate a comprehensive analysis and comparison of the pulmonary changes found in

COVID-19 and influenza (A/H1N1) patients, we followed a systematic workflow:

Autopsy Technique

We employed a COVID-19 optimized autopsy protocol in line with recently published

recommendations. Autopsies were performed in a room with adequate airflow (>6 air changes
per hour of total room volume) at conditions similar to recommendations for autopsies of
suspected Creutzfeld-Jakob disease (i.e. hazmat suits, boots, goggles, FFP2/3 masks) with an
in-corpore technique analogous to that used in forensic institutions. Thoracic organs were
eviscerated and the heart was separately dissected in the direction of blood flow. Lungs,
trachea and larynx were wholly exenterated and perfused via the trachea with phosphate-
buffered formalin. The trachea was then closed with a clamp and specimens were left in
formalin at room temperature for 72 hours before dissection. The lungs were subsequently cut
into 0.5-1cm parasagittal slices and examined macroscopically. Four sections of each lobe from
the periphery as well as the center of the lobes and the trachea were submitted for histological
analysis1.

µCT Analysis

Lung biopsies were fixed in 6% formaldehyde and dehydrated using increasing ethanol
gradients (30-70-80-90-100%) and 2 hours of hexamethyldisilazane to improve the
contrast, followed by air drying. Samples were fixed on a sample holder and scanned using
a Skyscan 1272 (Bruker, Kontich, Belgium) using the following settings: 200 mA, 45 kV, 0.2°
rotation angle and 3 frames per rotation angle (scan time 45 minutes). Reconstruction of the raw
images was performed using NRecon software (Bruker, Kontich, Belgium). 3D reconstructions
of the bronchovascular bundle in lungs were performed by region growing using ITK-SNAP
3.8.0 software2.

Histopathologic Analysis

After appropriate fixation with buffered formalin under a controlled inflation pressure of 55 – 65
cm H20 via the trachea representative sections of the lung were submitted for histological
analysis. As pulmonary specimens can contain a wide variety of injury patterns, patient as well
as control lungs were systematically sampled and morphologically comparable areas
systematically assessed by experienced pulmonary pathologists. According to established
guidelines, we analyzed the central and peripheral compartments of the upper, middle and
lower lungs, examining the airways and vasculature, including capillaries as well as pre- and
post-capillary vessels by Hematoxylin/Eosin (HE), Elastica van Giesson (EvG), and Periodic
Acid Schiff (PAS) staining, excluding areas of necrosis. Histopathological features of COVID-19
infected lungs and influenza lungs were scored semi-quantitatively with regard to their site of

3
manifestation.3 For qualitative and quantitative assessment of vascular thrombi by conventional
histology HE, PAS and EvG stains were scanned at high magnification. Three hot spot areas of
comparable morphology for each lobe of a cubic length of 1 centimeter were analyzed at high
magnification (400x) using immunohistochemical stains of CD61 and fibrinogen as additional
adjuncts. Thrombi in veins and arteries were counted as events per high power field, capillary
thrombi as events per alveolar wall segment. Results were normalized to 1 square centimeter.

Protein Expression Analysis

Complementary, compartment-specific detection of expressed proteins was performed to

characterize the inflammatory cell composition, potential virus entry sites, lymphatics and
presence of thrombi, using the multiplex labelling OPAL 7 system. The OPAL 7-plex
fluorescence system performs subsequent immunohistochemical antibody staining with
intermittent washing steps using the same tissue section and acquiring data by counting
fluorescence signals. By this approach co-localization of differentiation markers and spatial
relation of cells can be studied. To study the binding of multiple Abs to the same cells, OPAL 7-
plex fluorescence IHC staining was performed using the below mentioned antibodies in a single
tissue section. An optimal staining protocol was determined through titrating individual
antibodies and using different antigen retrieval conditions.

Briefly, composition of inflammatory cells was analyzed with regard to their localization to the
vasculature and alveolar walls using the following antibodies: lymphocytes (CD3), T-cells (CD4
and CD8), neutrophils (CD15), B-cells (CD20) and macrophages (CD68).

Specifically, the following antibodies were used: CD3 (1:200, Agilent Dako [GA50361-2]), CD4
(1:50, Zytomed Systems [503-3354], CD8 (1:600, Agilent Dako [M 7103]), CD15 (1:200, BD
biosciences [347420]), CD20 (1:1000, Agilent Dako [M 0755]), CD68 (1:1000, Agilent Dako [M
0876]), Podoplanin (1:25, Zytomed [MSK057]). Virus entry sites: (ACE2) and transmembrane
serine protease 4 (TMPRSS2): ACE2 (1:50 ABCAM [AB108252]), TMPRSS2 (1:500 ABCAM
[AB92323]). Thrombi were analyzed in pre and post capillary vessels: Fibrinogen (1:12000
DAKO [A0080]), CD61 (1:50 DAKO [M7103])4.

Transmission Electron Microscopy

Lung tissue was processed according to standard protocols and embedded in Epon. Briefly,
lung tissue samples were fixated overnight in 1.5% Glutaraldehyde and 1.5% Paraformaldehyde
in150mM Hepes pH7.3. After several washes with hepes buffer, specimens were fixed with 1%
(w/v) osmium tetroxide in Hepes buffer for 1 h at room temperature. Subsequently, samples
were dried in an increasing Aceton series and were embedded in EPON (Serva, Heidelberg,
Germany) following the instructions of the manufacturer. Semi-thin sections (0.5 μm) were
stained with tolouidine blue (Sigma Aldrich, Munich, Germany) and analyzed with a Zeiss
Axiophot microscope (Zeiss, Oberkochen, Germany). 700 Å ultrathin sections were
counterstained with 0.5% uranyl acetate (pH 4.5) for 30 min at 40°C and with 1% lead citrate for
30 min at 20°C (Ultrostainer; Leica) and analyzed using a Leo 906 digital transmission electron
microscope (Leo, Oberkochen, Germany) with an acceleration voltage of 80kV5.

4
Vascular Corrosion Casting, Scanning Electron Microscopy, and Morphometry

The afferent vessels of lung specimen were cannulated with an olive-tipped cannula. The
vasculature was flushed with saline (at body temperature) followed by glutaraldehyde fixation
solution (2.5%, pH 7.4, Sigma Aldrich, Munich, Germany). Fixation was followed by injection of
prepolymerized PU4ii resin (VasQtec, Zurich, Switzerland) mixed with a hardener (40% solvent)
and blue dye as casting medium. After curing of the resin, the lung tissue was macerated in
10% KOH (Fluka, Neu-Ulm, Germany) at 40°C for 2 to 3 days. Specimens were then rinsed with
water and frozen in distilled water. The casts were freeze-dried and sputtered with gold in an
argon atmosphere and examined using a Philips ESEM XL-30 scanning electron microscope
(Philips, Eindhoven, Netherlands). Vascular morphometry of variants of angiogenesis were then
assessed: high power images of the capillary plexuses were scanned (at least 10 images per
lung) in COVID-19 lungs, influenza (A/H1N1) lungs and age-matched controls. Intussusceptive
angiogenesis was identified via the occurrence of tiny holes with a diameter of 2-5 µm in
scanning electron micrographs of microvascular corrosion casts. Sprouting angiogenesis was
identified as extra luminal tiny sprouts of 2-3µm. The quantitation of these features was
expressed as numerical density per vessel area, FOV: 0.0168 mm². The data shown in Figure 4
represents numerical density from comparable areas in every lung6.

Gene Expression Analysis

Gene expression analysis was assessed by detection of individual copy numbers of RNA
molecules, using the NanoString nCounter technology. This approach uses a digital color-coded
barcode technology that is based on direct multiplexed measurement, thereby avoiding the
drawbacks of conventional real-time polymerase chain reaction7. For this, RNA was isolated
using the Maxwell® RNA extraction system (Promega, Madison, Wisconsin) and following
quality control via Qubit analysis (ThermoFisher, Waltham, Massachusetts) used for further
analysis. Both, RNA extraction and statistical analysis using the nCounter Analysis System
(NanoString Technologies, Seattle, WA) have been performed as previously described with the
following modifications: for this study, we used the commercially available PanCancer
Progression and Inflammation Panel and normalized samples to negative control lanes
(arithmetic mean background subtraction), positive control lanes (geometric mean normalization
factor) and all reference genes present on each panel (geometric mean normalization factor)
using the nSolver Analysis Software version 4.0. Statistical testing (Student’s T-tests) and a
subsequently calculated false discovery rate (FDR) threshold of 0.05 were used to identify
significant changes in gene expression4.

Statistical Analyses

All comparisons of numerical variables (including the gene expression analysis) were conducted
using Student’s t-test. Family-wise error rates (FWER) due to multiplicity were set at 0.05 by
using Benjamini-Hochberg method of controlling of false discovery rates (FDR). All confidence
intervals have been calculated based on the t-distribution as well. Calculation of t-tests and FDR
were conducted using the “stats” package from the statistical programming language R version
3.4.4 (https://www.r-project.org/) and, in the case of gene expression data, the nSolver Analysis
Software version 4.0.

5
Families of statistical tests (i.e. separate statistical analyses with separate control of the FWER)
are listed below. The analyses include all calculated p-values (ordered by size) and the maximal
FDR thresholds for which each p-value would still be considered significant (0.05 used in our
manuscript, all p-values above the FDR threshold are marked in red).

The analysis of gene expression data is excluded from this section due to its magnitude, more
details on it can be found in the section ‘Gene Expression Data’.

Lung weights

Comparison p-value FDR

COVID-19 vs. Controls 1.62E-08 4.86E-08
Influenza vs. Controls 2.93E-03 4.39E-03
COVID-19 vs. Influenza 4.48E-02 4.48E-02

Immunohistochemical analysis of ACE2 expression (see Table S2)

Cell type Comparison p-value FDR

Endothelial cells Influenza vs. Controls 5.52E-12 3.31E-11
Alveolar cells Influenza vs. Controls 1.66E-08 4.84E-08
Lymphocytes Influenza vs. Controls 2.42E-08 4.84E-08
Endothelial cells COVID-19 vs. Controls 1.08E-07 1.62E-07
Alveolar cells COVID-19 vs. Controls 6.01E-06 7.21E-06
Lymphocytes COVID-19 vs. Controls 1.81E-05 1.81E-05

Immunohistochemical analysis of CD3-, CD4-, CD8-, CD15, and CD20 positive cells

Cell type Comparison p-value FDR

CD15 COVID-19 vs. Influenza 1.55E-03 6.19E-03
CD8 COVID-19 vs. Influenza 7.81E-03 1.56E-02
CD4 COVID-19 vs. Influenza 3.75E-02 4.99E-02
CD3 COVID-19 vs. Influenza 9.43E-02 9.43E-02

Thrombosis

Compartment Comparison p-value FDR

Capillary COVID-19 vs. Influenza 1.85E-03 5.54E-03
Vene COVID-19 vs. Influenza 1.82E-02 2.74E-02
Artery COVID-19 vs. Influenza 1.88E-01 1.88E-01

6
Quantification of angiogenic features

Angiogenesis Comparison p-value FDR

Intussuscpetion COVID-19 vs. Controls 4.67E-09 1.87E-08
Intussuscpetion COVID-19 vs. Influenza 2.58E-06 5.15E-06
Sprouting COVID-19 vs. Controls 3.69E-05 4.93E-05
Sprouting COVID-19 vs. Influenza 8.31E-03 8.31E-03

Correlation of angiogenic features with hospitalization time

Angiogenesis Disease p-value FDR

Intussuscpetion COVID-19 1.28E-04 5.12E-04
Sprouting COVID-19 0.0108 2.16E-02
Sprouting Influenza 0.19 0.25
Intussuscpetion Influenza 0.52 0.52

7
Supplementary Figures

Figure S1. Relative expression analysis of inflammation-associated genes in COVID-19

and influenza (A/H1N1) lungs compared to uninfected control lungs.

RNA was isolated from sections sampled directly adjacent to those used for complementary
histology and immunohistochemistry. RNA was isolated using the Maxwell® RNA extraction
system (Promega, Madison, Wisconsin) and following quality control via Qubit analysis
(ThermoFisher, Waltham, Massachusetts) used for further analysis. During the NanoString
procedure, individual copies of all RNA molecules were labelled with gene specific bar codes
and counted individually using the nCounter Analysis System (NanoString Technologies,
Seattle, WA). The expression of inflammation-associated genes was measured using the
NanoString nCounter Inflammation Panel (all 249 genes present on this panel). The resulting
gene expression data was normalized to negative control lanes (arithmetic mean background
subtraction), positive control lanes (geometric mean normalization factor) and all reference
genes present on the panel (geometric mean normalization factor) using the nSolver Analysis
Software version 4.0. Displayed in the Venn diagram are only genes that are statistically
differentially expressed compared to controls in both disease groups (Student’s T-Test, FDR ≤
0.05). Up- and downregulation of genes is indicated by colored arrows suffixed to the gene
symbols (green = upregulation, red = downregulation).

8
Figure S2. Histopathology of an Influenza (A/H1N1) lung.

Diffuse alveolar damage (DAD) as the histomorphological hallmark of severe acute respiratory
distress syndrome (ARDS) in an influenza A infected lung: there was increased cellularity with
abundant interstitial and intraalveolar inflammatory cells, cellular debris and edema. In the
alveolar lumen, the ubiquitously exudated proteins aggregated i) together with loose connective
tissue and inflammatory cells into acute fibrinous organizing pneumonia (AFOP) and ii) into
hyaline membranes, superimposed on the alveolar lining, the latter with hyperplastic type II
alveolar cells and in part desquamated. (HE stain, 50x magnification, scale bar = 200 µm).

9
Figure S3. COVID-19-associated thrombosis. (A,B): In the inflamed vessels, there
were multifocal thrombi (*) with (sub) total vascular occlusion of both pulmonary arteries
and veins as visualized by scanning electron microscopy (A) and conventional
histopathology (B) (scale bar = 100um). In the scanning electron microscopy image, the
thrombus is pseudocolored pink and the infiltrating lymphocytes pseudocolored yellow.
(C, D): µCT-based 3D reconstruction of subsegmental pulmonary arteries (red) and
airways (blue) demonstrated (sub)total occlusion of the arteries in COVID-19-lungs (C),
as compared to uninfected controls (D) (scale bar = 300 µm).

10
Figure S 4. Schematic of intussusceptive and sprouting angiogenesis. (A) The regular
vasculature consists of an afferent (a.v.) and an efferent vessel (e.v.) and the intervening
pulmonary alveolar capillary plexus (diameters not to scale). (B) Intussusceptive angiogenesis is
found in COVID19-pulmonary plexuses characterized by the separation of preexisting blood
vessels into two via the formation of an intravascular pillar (p). This allows a rapid expansion of
the vascular plexuses to inflammatory and metabolic adaptations, and recruitment and
incorporation of circulating progenitor cells from the bone marrow due to mechanosensitive
regulations of shear stress and blood flow velocity. (C) The hallmark of sprouting angiogenesis
is the formation of new “daughter” channels (s) composed of endothelial cells directly from an
existing vessel extending to form tip and stalk cells, regulated by hypoxia and growth factors.

11
Figure S5. Anaglyphic 3D-image of a scanning electron micrograph highlighting
the changes in a COVID-19 lung. This represents a 3D reconstruction of a scanning
electron micrograph of the vasculature and small airways of a COVID-19 lung. Pairs of stereo
images were taken by scanning electron micrographs using a tilt angle difference of exactly 6
degree (Philips XL30, Eindhoven, The Netherlands). The images were stored and reconstructed
three-dimensionally by image-processing software (KS300, Kontron Zeiss, Eching, Germany).
When viewed with red-green glasses, an expanded interstitium with interspersed inflammatory
cells and aerated alveoli can be seen in 3D. In the alveolar lumina, numerous lymphocytes are
attached to the epithelial lining. On the left hand a (sub) total vascular occlusion by a fibrin
thrombus is demonstrated (scale bar = 100 µm).

12
 Supplementary Tables
Table S1A
Clinical, radiological and (histo-) pathological characteristics of COVID-19 patients

Patient ID 1 2 3 4 5 6 7
Age (Years) 68 86 96 78 66 74 81
Gender F M M M M M F
BMI (kg/m²) 35 26 23 44 29 27 26
Mechanical
No No No Yes Yes No No
Ventilation
PC
PC (myelo- Intubation
Reason non- PC (multiple PC (elderly + (metastasized
dysplastic - - refused by
ventilation sclerosis) M. Parkinson) prostate
syndrome) patient
cancer)
Ventilation
No No No NIV NIV No No
mode
PEEP
- - - 17 18 - -
(cmH2O)
FiO2 - - - 0.80 0.90 - -
Hospital-
isation (days 9 5 3 3 9 3 4
at death)
Cause of
RF RF RF CRF RF RF RF
death
Smoker No No No Yes Yes Yes Yes
Hypertension Yes Yes Yes Yes Yes Yes Yes
Diabetes
No Yes No Yes No Yes No
Type II
Immunosuppr
No Yes No No No No No
ession
RAAS
interacting Yes Yes Yes Yes Yes Yes Yes
drugs
Ground glass
Yes No Yes Yes Yes Yes Yes
(CT)
Consolidation
No Yes Yes Yes Yes Yes No
(CT)
Lung weight
1660 1740 1780 1600 1650 1720 1620
(g)
Arterial
2
Thrombi/cm 1.25±5.59 None None 1.25±5.59 1.25±5.59 None 2.51.25±7.70
tissue area
Capillary
2
Thrombi/cm 98.96±56.73 102±47.28 165±47.57 234.21±67.29 276.32±108.81 161.25±51.60 78.75±45.36
tissue area
Venous
2
Thrombi/cm None 6.25±13.75 15±18.85 43.75±24.16 12.5±17.21 11.25±18.98 None
tissue area
DAD Yes Yes Yes Yes Yes Yes Yes

Clinical, radiological and (histo-) pathological characteristics of COVID-19 patients. Patients

were evaluated for age, gender, BMI, smoking status, hypertension, Diabetes, medication intake

13
(immunosuppression or RAAS interacting drugs), as well as days of hospitalization prior to
death, administration of mechanical ventilation and cause of death according to results of the
autopsy. Number of present thrombi was examined in multiple comparable (e.g. exclusion of
necrosis) field of views and normalized to 1 cm² tissue area. Values are given as means per cm²
tissue ± standard deviation. PC – palliative care, NIV – non-invasive ventilation, RF – respiratory
failure, CRF – cardiorespiratory failure, DAD – diffuse alveolar damage

14
 Table S1B
Clinical and radiological characteristics of Influenza patients (top) and control lungs (bottom)
Influenza (A/H1N1) patients
Patient ID 1 2 3 4 5 6 7
Age (Years) 74 53 52 66 53 45 59
Gender M M M F M M F
BMI (kg/m²) 41.1 (p.m.) 21 46.7 (p.m.) 27 35.2 48.5 (p.m.) 28.6
Mechanical
yes yes yes yes Yes Yes Yes
Ventilation
Ventilation mode PCV PCV NIV NIV PCV PCV PCV
PEEP (cmH2O) 16 20 20 15 18 16 13
FiO2 (%) 100 70 80 50 65 80 80
Hospitalization (days) 9 12 4 17 11 8 20
CRF, Myocarditis,
Cause of death consecutive CRF consecutive MOF CRF CRF RF
MOF MOF
Smoker yes yes no no Yes No No
Hypertension yes yes yes yes Yes Yes No
Diabetes Type II No No No No No No No
Immunosuppression No No No No No No No
RAAS interacting
Yes Yes Yes No No No No
drugs
Ground glass (CT) No No No No No No Yes
Consolidation (CT) Yes Yes No Yes Yes Yes No
Lung weight (g) 2680 2820 1400 1800 3230 3200 1700
Arterial Thrombi 6.25±11.11 None 2.5±7.70 3.75±9.16 None None 3.75±9.16
Capillary Thrombi 20±22.82 3±10.99 15±20.52 6±14.93 9±20.26 51.25±34.86 9±15.94
Venous Thrombi 55±33.05 17.5±20.03 37.5±34.89 18.75±35.24 37.5±35.82 25±34.41 58.75±38.28
DAD Yes Yes Yes Yes Yes Yes Yes
Control patients
Patient ID 1 2 3 4 5 6 7 8 9 10
Age (Years) 77 69 75 79 83 62 79 69 62 82
Gender M F M F M F M F F M
BMI (kg/m²) 29.39 22.31 25.395 23.66 27.78 25.39 23.88 23.88 22.95 20.57
Mechanical
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Ventilation
Ventilation mode VCV VCV VCV VCV VCV VCV VCV VCV VCV VCV
PEEP (cmH2O) 10 5 4 5 5 6 5 5 5 5
FiO2 (%) 70 100 100 100 50 35 100 40 35 100
Hospitalization (days) 3 3.5 25.8 1.8 2.9 2.3 2.3 2.9 5.9 1.75
Cardiac
Cause of death Trauma Trauma Trauma CVA Trauma CVA Trauma CVA CVA
arrest
Smoker No No Yes No No Yes No Yes No No
Hypertension No No Yes Yes Yes Yes No Yes No Yes
Diabetes Type II No No No No Yes Yes No No No No
Immunosuppression No No No No No No No No No No

15
RAAS interacting
No No No Yes No No No No Yes No
drugs
Ground glass (CT) Mild No No No No No No No No No
No No No No No No No No No Mild
Lung weight (g) 1175.5 967.7 1091.1 881.5 1085.5 967.5 961.9 1011.29 926.5 1382.94

Clinical, and radiological characteristics of influenza (A/H1N1) (top) and control (bottom)
patients. Patients were evaluated for age, gender, BMI, smoking status, hypertension, Diabetes,
medication intake (immunosuppression or RAAS interacting drugs), as well as days of
hospitalization prior to death, administration of mechanical ventilation and cause of death
according to results of the autopsy. Number of present thrombi was examined in multiple
comparable (e.g. exclusion of necrosis) field of views and normalized to 1 cm² tissue area.
Values are given as means per cm² tissue ± standard deviation. P.m. – values collected post
mortem, PCV - pressure control ventilation, NIV – non-invasive ventilation, RF – respiratory
failure, CRF – cardiorespiratory failure, MOF – multi-organ-failure, CVA - Cerebrovascular
accident

16
Table S2

Relative counts of ACE2 positive cells of COVID-19, Influenza (A/H1N1) and control cases

Cell Type COVID-19 Influenza Controls

Alveolar cells 0.25 (±0.14)*** 0.35 (±0.15)*** 0.053 (±0.03)

Endothelial cells 0.49 (±0.28)*** 0.55 (±0.11)*** 0.066 (±0.03)

Lymphocytes 0.22 (±0.18)*** 0.15 (±0.09)*** 0.000 (±0.00)

Histological sections of paraffin-fixed formalin-embedded tissue from COVID-19, Influenza

(A/H1N1) and control patients were stained for ACE2 via immunofluorescence using the OPAL
7-plex system. Multiple fields of view per case centered on pre- and post-capillary blood vessels
were analyzed for relative counts of ACE2 positive alveolar cells, endothelial cells and
lymphocytes. Mean counts (± standard deviation) of evaluable cells per field of view were,
except for the absent lymphocytes in control lungs, comparable between COVID-19, Influenza
and control lungs: 26.04±6.47 alveolar cells, 15.29±7.09 endothelial cells and 27.83±30.62 for
lymphocytes. Values are listed as means of relative counts per field of view (± standard
deviation). Significance of results was determined using the Student's t‐test controlled for the
family-wise error rate using a Benjamini-Hochberg false discovery rate threshold of 0.05.
*** p ≤ 0.001; significant difference against control.

Gene Expression Data

All gene expression and associated data has been uploaded on the Vivli platform
(https://vivli.org/) and can be requested via the following digital object identifier (DOI):
https://doi.org/10.25934/00005576.

17
References

1. Menter T, Haslbauer J, Nienhold R, et al. Post-mortem examination of COVID19 patients

reveals diffuse alveolar damage with massive capillary congestion and variegated

findings of lungs and other organs suggesting vascular dysfunction. Histopathology

2020;HISTOP-04-:1–57.

2. Verleden SE, Vasilescu DM, Willems S, et al. The site and nature of airway obstruction

after lung transplantation. Am J Respir Crit Care Med 2014;189(3):292–300.

3. Jonigk D, Stark H, Braubach P, et al. Morphological and molecular motifs of fibrosing

pulmonary injury patterns. J Pathol Clin Res 2019;5(4):256–71.

4. Neubert L, Borchert P, Stark H, et al. Molecular Profiling of Vascular Remodeling in

Chronic Pulmonary Disease. Am J Pathol [Internet] 2020;Available from:

https://doi.org/10.1016/j.ajpath.2020.03.008

5. Ackermann M, Houdek JP, Gibney BC, et al. Sprouting and intussusceptive angiogenesis

in postpneumonectomy lung growth: Mechanisms of alveolar neovascularization.

Angiogenesis 2014;17(3):541–51.

6. Ackermann M, Konerding MA. Vascular morphogenesis: Methods and protocols. Vasc

Morphog Methods Protoc 2014;1214:1–272.

7. Geiss GK, Bumgarner RE, Birditt B, et al. Direct multiplexed measurement of gene

expression with color-coded probe pairs. Nat Biotechnol 2008;26(3):317–25.

18