Available via license: CC BY-NC 4.0
Content may be subject to copyright.
Andrology
Identification of two hidden clinical subgroups among
men with idiopathic cryptozoospermia
Lena Charlotte Sch€
ulke
1
, Joachim Wistuba
1
, Verena Nordhoff
2
, Hermann M. Behre
3
, Jann-Frederik Cremers
2
, Sabine Kliesch
2
,
Sara Di Persio
1,†
, and Nina Neuhaus
1,
,†
1
Centre of Reproductive Medicine and Andrology, Institute of Reproductive and Regenerative Biology, University of M€
unster, M€
unster, Germany
2
Department of Clinical and Surgical Andrology, Centre of Reproductive Medicine and Andrology, University of M€
unster, M€
unster, Germany
3
UKM Kinderwunschzentrum, Universit€
atsklinikum M€
unster, M€
unster, Germany
Correspondence address. Centre of Reproductive Medicine and Andrology, University Hospital of M€
unster, Albert-Schweitzer-Campus 1, Building D11,
48149 M€
unster, Germany. E-mail: nina.neuhaus@ukmuenster.de https://orcid.org/0000-0003-0181-6194
†
These authors contributed equally to this work.
ABSTRACT
STUDY QUESTION: Are there subgroups among patients with cryptozoospermia pointing to distinct etiologies?
SUMMARY ANSWER: We reveal two distinct subgroups of cryptozoospermic (Crypto) patients based on testicular tissue composi-
tion, testicular volume, and FSH levels.
WHAT IS KNOWN ALREADY: Cryptozoospermic patients present with a sperm concentration below 0.1 million/ml. While the etiol-
ogy of the severely impaired spermatogenesis remains largely unknown, alterations of the spermatogonial compartment have been
reported including a reduction of the reserve stem cells in these patients.
STUDY DESIGN, SIZE, DURATION: To assess whether there are distinct subgroups among cryptozoospermic patients, we applied
the statistical method of cluster analysis. For this, we retrospectively selected 132 cryptozoospermic patients from a clinical database
who underwent a testicular biopsy in the frame of fertility treatment at a university hospital. As controls (Control), we selected 160
patients with obstructive azoospermia and full spermatogenesis. All 292 patients underwent routine evaluation for endocrine,
semen, and histological parameters (i.e. the percentage of tubules with elongated spermatids). Moreover, outcome of medically
assisted reproduction (MAR) was assessed for cryptozoospermic (n ¼73) and Control patients (n ¼87), respectively. For in-depth
immunohistochemical and histomorphometrical analyses, representative tissue samples from cryptozoospermic (n ¼27) and
Control patients (n¼12) were selected based on cluster analysis results and histological parameters.
PARTICIPANTS/MATERIALS, SETTING, METHODS: This study included two parts: firstly using clinical parameters of the entire co-
hort of 292 patients, we performed principal component analysis (PCA) followed by hierarchical clustering on principal components
(i.e. considering hormonal values, ejaculate parameters, and histological information). Secondly, for histological analyses seminifer-
ous tubules were categorized according to the most advanced germ cell type present in sections stained with Periodic acid Schif. On
the selected cohort of 39 patients (12 Control, 27 cryptozoospermic), we performed immunohistochemistry for spermatogonial
markers melanoma-associated antigen 4 (MAGEA4) and piwi like RNA-mediated gene silencing 4 (PIWIL4) followed by quantitative
analyses. Moreover, the morphologically defined A
dark
spermatogonia, which are considered to be the reserve stem cells,
were quantified.
MAIN RESULTS AND THE ROLE OF CHANCE: The PCA and hierarchical clustering revealed three different clusters, one of them con-
taining all Control samples. The main factors driving the sorting of patients to the clusters were the percentage of tubules with elon-
gated spermatids (Cluster 1, all Control patients and two cryptozoospermic patients), the percentage of tubules with spermatocytes
(Cluster 2, cryptozoospermic patients), and tubules showing a Sertoli cells only phenotype (Cluster 3, cryptozoospermic patients).
Importantly, the percentage of tubules containing elongated spermatids was comparable between Clusters 2 and 3. Additional differ-
ences were higher FSH levels (P<0.001) and lower testicular volumes (P<0.001) in Cluster 3 compared to Cluster 2. In the spermato-
gonial compartment of both cryptozoospermic Clusters, we found lower numbers of MAGEA4
þ
and A
dark
spermatogonia but higher
proportions of PIWIL4
þ
spermatogonia, which were significantly correlated with a lower percentage of tubules containing elongated
spermatids. In line with this common alteration, the outcome of MAR was comparable between Controls as well as both cryptozoo-
spermic Clusters.
LIMITATIONS, REASONS FOR CAUTION: While we have uncovered the existence of subgroups within the cohort of cryptozoosper-
mic patients, comprehensive genetic analyses remain to be performed to unravel potentially distinct etiologies.
WIDER IMPLICATIONS OF THE FINDINGS: The novel insight that cryptozoospermic patients can be divided into two subgroups
will facilitate the strategic search for underlying genetic etiologies. Moreover, the shared alterations of the spermatogonial stem cell
compartment between the two cryptozoospermic subgroups could represent a general response mechanism to the reduced
output of sperm, which may be associated with a progressive phenotype. This study therefore offers novel approaches towards the
understanding of the etiology underlying the reduced sperm formation in cryptozoospermic patients.
Received: July 12, 2023. Revised: December 20, 2023. Editorial decision: January 19, 2024.
# The Author(s) 2024. Published by Oxford University Press on behalf of European Society of Human Reproduction and Embryology.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (https://creativecommons.org/
licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For
commercial re-use, please contact journals.permissions@oup.com
Human Reproduction, 2024, 39(5), 892–901
https://doi.org/10.1093/humrep/deae013
Advance Access Publication Date: February 14, 2024
Original Article
STUDY FUNDING/COMPETING INTEREST(S): German research foundation CRU 326 (grants to: SDP, NN). Moreover, we thank the Faculty
of Medicine of the University of M €
unster for the financial support of Lena Charlotte Sch€
ulke through the MedK-program. We acknowledge
support from the Open Access Publication Fund of the University of M€
unster. The authors have no potential conflicts of interest.
TRIAL REGISTRATION NUMBER: N/A.
Keywords: cluster analysis / cryptozoospermia / piwi like RNA-mediated gene silencing 4 / male infertility / medically assisted repro-
duction / spermatogonial stem cells / meiotic arrest / Sertoli cell only phenotype / testis / spermatogenesis
Introduction
Men with cryptozoospermia have a sperm concentration of 0.1
million per milliliter in the ejaculate, rendering natural concep-
tion almost impossible. It is noteworthy that rare motile sperma-
tozoa can usually only be detected after resuspension of semen
sediments (World Health Organization, 2010), reflecting the
severe impairment of spermatogenesis within testicular tissues.
First insights pointing to the stage of germ cell differentiation
failure in cryptozoospermic testes were provided by single cell
RNA-sequencing data (Di Persio et al., 2021; Di Persio and
Neuhaus, 2023). These revealed that the proportions of sperma-
togonia and early meiotic stages were still comparable to the
control situation. Only from pachytene stage onwards did crypto-
zoospermic samples show a reduction of germ cells (Di Persio
et al., 2021). Moreover, these datasets highlighted alterations al-
ready in the spermatogonial stem cell compartment, which
forms the basis of spermatogenesis. Unexpectedly, testicular tis-
sues from cryptozoospermic patients showed increased propor-
tions of piwi like RNA-mediated gene silencing 4 (PIWIL4
þ
)
spermatogonia (Di Persio et al., 2021), which are considered to be
the origin of the spermatogonial differentiation process
(Guo et al., 2018; Sohni et al., 2019; Shami et al., 2020; Di Persio
et al., 2021). In contrast to this, the number of morphologically
defined A
dark
spermatogonia, which have been suggested to
include the reserve stem cell population, was reduced (Clermont,
1963; van Alphen et al., 1988; Caldeira-Brant et al., 2020; Di Persio
et al., 2021). Two possible scenarios can be considered as explana-
tion for these findings. The increased proportion of PIWIL4
þ
spermatogonia may point to a stem cell defect prohibiting
differentiation of the most undifferentiated spermatogonia
and thereby leading to an accumulation of this cell type.
Alternatively, the activation of the spermatogonial stem cell
compartment may be a response mechanism to the reduced pro-
duction of sperm. However, it seems likely that this over activa-
tion of the stem cell pool, which is also at the cost of reserve
stem cell (A
dark
) numbers, may lead to stem cell exhaustion and
a progressive phenotype of cryptozoospermia. Analyses focusing
also on the spermatogonial compartment are therefore essential
to unravel the cellular changes leading to the severely reduced
sperm production in cryptozoospermic men.
Provided that clinical parameters can be identified that distin-
guish subgroups among the highly heterogeneous group of cryp-
tozoospermic patients, these may point to distinct etiologies. One
example demonstrating the successful use of this approach is
the genetic screening of men with meiotic arrest, which led to
the identification of meiosis 1 Associated Protein (M1AP) (mutations
causative for meiotic failure). Specifically, deleterious mutations
in M1AP lead to complete meiotic arrest and azoospermia in
some patients and to cryptozoospermia in others (Wyrwoll et al.,
2020). An analysis of cryptozoospermic patients to unveil poten-
tial subgroups is therefore a necessity to reveal the underlying
etiologies and improve counseling and, prospectively, treatment.
To this end, the statistical approach of cluster analysis can be ap-
plied, which has been successfully used in other medical fields to
improve diagnostics in heterogeneous patient cohorts, building
the basis for specialized treatments and a better prediction of the
prognosis (Burgel et al., 2014; Zhu et al., 2019; Krenz et al., 2021).
This study therefore aimed to assess if subgroups of patients
can be identified in a cohort of 132 cryptozoospermic men con-
sidering clinical/andrological parameters, and to evaluate the
alterations in the spermatogonial compartment and outcome of
medically assisted reproduction (MAR).
Materials and methods
Ethical approval
For the use of testicular tissue samples from patients with ob-
structive azoospermia and cryptozoospermia, ethical approval
was obtained from the Ethics Committee of the Medical Faculty
of M€
unster and the State Medical Board (No. 2008-090-f-S).
Testicular biopsies for research were obtained during microsurgi-
cal testicular sperm extraction and histological analysis was car-
ried out at the Department of Clinical and Surgical Andrology at
the Centre for Reproductive Medicine and Andrology (CeRA),
University Hospital M€
unster following written informed consent.
Selection of the patient cohort
We retrospectively selected all the 3539 patients from the data-
base Androbase
V
R
(T€
uttelmann et al., 2006) who underwent a tes-
ticular biopsy at the CeRA, M€
unster between 1991 and 2020. Only
patients with a complete dataset including physical examination,
hormone analysis (including LH, FSH, testosterone (T), and free
testosterone), ultrasound of the testes, semen analysis according
to the World Health Organization criteria (World Health
Organization, 2010) and if indicated genetic analyses (including
karyotype and azoospermia factor (AZF) deletions, cf transmem-
brane conductance regulator (CFTR) mutations, and single nucle-
otide polymorphism (SNP) FSHB-211 G>T genotyping; Busch et al.,
2015) were included in the study cohorts.
To obtain a cohort of idiopathic cryptozoospermic men,
patients with a sperm concentration 0.1 million per milliliter of
ejaculate were selected. Rare motile spermatozoa can usually
only be detected after resuspension of semen sediments (World
Health Organization, 2010). Exclusion criteria were genetic dis-
eases (i.e. Klinefelter syndrome, AZF deletions), hormonal treat-
ments (i.e. hCG, rFSH, Tamoxifen, Anastrozol), and a history of
cryptorchidism or oncological disease.
For the Control cohort, patients with obstructive azoospermia
(full spermatogenesis, no sperm in the ejaculate) owing to con-
genital bilateral absence of the vas deferens (CBAVD), vaso-
vasostomy, or a physical obstruction in the vas deferens were se-
lected. Exclusion criteria were genetic diseases (i.e. Klinefelter
syndrome, AZF deletions), hormonal treatments (i.e. hCG, rFSH,
Tamoxifen, Anastrozol), and a history of cryptorchidism or onco-
logical disease. Furthermore, patients who only had one testis or
had a FSH concentration outside the reference range (1–7 U/l)
and a Bergmann–Kliesch score below 7 were excluded (Bergmann
and Kliesch, 2010). Applying these criteria, we selected a cohort
Two subgroups among men with cryptozoospermia | 893
of 160 Control and 132 cryptozoospermic patients (Fig. 1,
Supplementary Table S1).
Hormone measurements
Hormone measurements, including gonadotrophins and T, were
performed using blood samples taken in the morning. LH (Cat
No. 02P40-25), FSH (Cat No. 07 K75-25), T (Cat No. 02P13-28), sex
hormone-binding globulin (SHBG, Cat No. 08 K26-20), prolactin
(Cat No. 07 K76-25), estradiol (Cat No. 07 K72-25), and prostate-
specific antigen (PSA, Cat No. 07 K70-25) were measured using
chemiluminescent microparticle immunoassays analyzed with
the Architect i1000 (Abbott Diagnostics, Wiesbaden, Germany).
Free testosterone was calculated considering the total testoster-
one and the fraction bound to SHBG using the published
formula (https://www.issam.ch/freetesto.htm; Vermeulen et al.,
1999). Moreover, ACTIVE
V
R
dihydrotestosterone (DHT) radio-
immunoassay (DSL-9600 Beckmann-Coulter, Krefeld, Germany),
was used to measure DHT. The assay shows less than 2% cross-
reactivity with T and is calibrated against a standard of LCMS-MS
with an accuracy of <15% within the range of 0.1 to 5 nmol/l.
Intra-assay co-efficient of variation (CV) is 3.5%, mean inter-
assay CV is 7% for the DHT assay.
Principal component analysis and
hierarchical clustering
The principal component analysis (PCA) and clustering were per-
formed using the hierarchical clustering on principal compo-
nents (HCPC) method. R 4.0.5 and the R packages FactoMineR
(Version 2.4) (L^
e et al., 2008) and factoextra (Version 1.0.7) were
used. For the cohort of 292 patients, the variables included in the
analyses were: age of the patients at the time of surgery, LH, FSH,
T, free testosterone, ejaculate volume, ejaculate pH, sperm con-
centration, total sperm count, and testicular volume (Busch et al.,
2019; Krenz et al., 2021). In addition, the histological information
(percentage of tubules with elongated spermatids, round sper-
matids, spermatocytes, spermatogonia as the most advanced
germ cell types or displaying Sertoli cell only (SCO) phenotype or
tubular shadows) was taken into consideration. An overview of
all clinical parameters is provided in Supplementary Table S1.
The PCA was performed to reduce the dimensionality of the data
into few continuous variables using the function PCA() from the
FactoMineR (Version 2.4) package. Prior to the analysis, the data
were scaled to avoid dominance by variables with large measure-
ment units. Five dimensions were retained in the output. The
results of the PCA were visualized using the factoextra (Version
1.0.7) package. The hierarchical clustering was performed using
the Ward’s criterion on the selected principal components using
the function HCPC() from the FactoMineR (Version 2.4) package.
The partition in different clusters was initially performed by cut-
ting the hierarchical tree, and the K-means clustering was then
used to improve the initial partition. The function fviz_cluster()
in factoextra (Version 1.0.7) package was used to visualize indi-
vidual clusters.
Analysis of ICSI results
Data reporting the outcome of MAR treatments and ICSI proce-
dures performed at the UKM Kinderwunschzentrum and the
Centre of Reproductive Medicine and Andrology were available
for 160 of the 292 patients included in the study. The percentages
of embryo transfer, clinical and biochemical pregnancies, and
live birth were calculated for each Cluster based on the total
number of patients for which ICSI data were available. The out-
come was considered positive if a patient had at least one em-
bryo transfer, a biochemical or clinical pregnancy, or a live birth,
respectively, and results were not normalized by number of
ICSI cycle.
Periodic acid-Schiff staining and histological
evaluation of human testicular tissue sections
For routine histological evaluation, testicular tissues were proc-
essed and stained according to previously described protocols
(Di Persio et al., 2021). To assess the histological status, periodic
acid-Schiff-stained sections from two independent biopsies per
testis were graded using the Bergmann–Kliesch (BK) score
(Bergmann and Kliesch, 2010). A score from 0 to 10 was assigned
to each patient based on the percentage of tubules containing
elongated spermatids, with a score of 1 indicating 10% of the
tubules with elongated spermatids. Additionally, the percen-
tages of tubules with round spermatids, spermatocytes, and
spermatogonia as the most advanced germ cell type as well
as the percentage of SCO tubules and tubular shadows
were evaluated.
Paents with a tescular biopsy
n= 3539
Azoospermic paents
(0M sperm /ml)
n=1987
Obstrucve azoospermia
(Normal)
n=160
Paents with genec
diseases, maldecensus,
tumors, medicaon, BK-
score <7,FSH >7
n=1827
Cryptozoospermic paents
(< 0.1M sperm/ml)
)
)
)
n=414
Cryptozoospermia
(Crypto)
n=132
Paents with genec
diseases, maldecensus,
tumors, medicaon
n=282
Paents with > 0.1M
sperm/ml
n=323
Paents without sperm
paramenters
n=815
Figure 1. Selection of the study cohort. Flow chart depicting the selection of 160 Control (teal) and 132 cryptozoospermic (Crypto, purple) patients from
the database Androbase. Exclusion criteria are provided in grey boxes. M, million; ml, milliliter; BK-Score, Bergmann–Kliesch Score.
894 | Sch€
ulke et al.
Immunohistochemical staining of human
testicular tissue sections
For immunohistochemical analyses, we randomly selected 13
samples from Cluster 1 (including 12 Control patients and one
cryptozoospermic patient), 13 cryptozoospermic samples from
Clusters 2 and 3. The samples from Clusters 2 and 3 were
matched based on the BK score to ensure an equal representation
of tubules with full spermatogenesis in the two groups. The im-
munohistochemical stainings were performed as previously pub-
lished (Di Persio et al., 2021). Mouse monoclonal anti-melanoma-
associated antigen 4 (MAGEA4) (provided by Prof. G.C. Spagnoli,
University Hospital of Basel, Switzerland; dilution 1:20) and rab-
bit polyclonal anti-PIWIL4 (PIWIL4, Cat No. HPA036588, Atlas
Antibodies, Bromma, Sweden; dilution 1:50) were used as pri-
mary antibodies. As negative controls, sections were incubated
with species-specific immunoglobulin G (Mouse-IgG Control
Antibody, unconjugated, Cat No. 15381, Sigma Aldrich, St Louis,
MO, USA; dilution 1:100 and Rabbit-IgG Control Antibody, uncon-
jugated, Cat No. 15006, Sigma Aldrich; dilution 1:100) or without
primary antibody. After overnight incubation, the sections were
incubated with a corresponding secondary antibody (Goat F(ab’)2
Anti-Mouse IgG – (Fab’)2 (Biotin), Cat No. ab5886, Abcam,
Cambridge, UK; dilution 1:100 and Goat F(ab’)2 Anti-Rabbit
IgG—H&L (Biotin), Cat No. ab6012, Abcam; dilution 1:100) and
subsequently with streptavidin-peroxidase (Cat No. S5512, Sigma-
Aldrich; dilution 1:500). 3,3’-diaminobenzidine tetrahydrochloride
solution (Cat No. A0596.0001, Applichem, Darmstadt, Germany)
was used to detect the peroxidase activity. Lastly, the nuclei were
counterstained with Mayer’s hematoxylin (Cat No. 1.092.490.500,
Sigma-Aldrich). Stained sections were digitized employing the
Precipoint M8 Microscope and Scanner (Precipoint, Freising,
Germany), using a 60objective and were analyzed using the
Precipoint software Viewpoint (Precipoint, Freising, Germany).
Quantifications of spermatogonial
subpopulations
The numbers of MAGEA4
þ
and PIWIL4
þ
spermatogonia per tu-
bule were quantified in two independent testicular tissue sec-
tions and 20 round tubules per sample, which was the maximum
number that could be reached in some cryptozoospermic sam-
ples owing to the limited number of tubules containing germ
cells. Tubules were regarded as round when the ratio of the two
diameters was between 1 and 1.5. To obtain the proportion of
PIWIL4
þ
spermatogonia among the entire spermatogonial popu-
lation, the number of PIWIL4
þ
spermatogonia was normalized to
the number of MAGEA4
þ
spermatogonia. A
dark
spermatogonia
were identified based on morphological criteria (Clermont, 1966).
The A
dark
spermatogonia are characterized by an unstained rare-
faction zone in the center of a nucleus with uniformly dark
stained chromatin. The proportion of A
dark
spermatogonia was
quantified per 100 MAGEA4
þ
spermatogonia.
Statistical analyses and data presentation
For the clinical and histological data, normality (Shapiro–Wilk
test), homogeneity of variances (Levene and Fligner tests), and
homoscedasticity (Breusch–Pagan) tests were performed for all
variables. For normally distributed data, the one-way ANOVA
test followed by pairwise Student’s t-test was performed. In case
of data not normally distributed, the Kruskal–Wallis rank sum
test was used to compare the values among the three patient
groups, followed by pairwise Wilcox tests. Prior to correlation
analysis, the data were checked for normality using the Shapiro–
Wilk test. Based on the normality test results, the correlation
analyses were performed using the Spearman correlation
method. Statistical analyses were executed using R 4.1.2 and
specifically the package stats (Version 4.1.2) (R a language and
environment for statistical computing, 2010). Details regarding
the statistical analysis of each variable are provided in
Supplementary Tables S2 and S3.
Plots were generated using GraphPad Prism
V
R
Version 5.0
(GraphPad Software, Inc., San Diego, CA, USA). Box plots ele-
ments are defined as follows: center line represents the median;
box limits depict upper and lower quartiles and whiskers
the 1.5interquartile range. Outliers are shown as individ-
ual points.
Results
PCA and hierarchical clustering based on clinical
data of cryptozoospermic men
To assess if there are different subgroups of cryptozoospermic
patients, we screened 3.539 patients who underwent a testicular
biopsy and full clinical evaluation at the CeRA. We selected
160 patients with obstructive azoospermia (Control) and
132 cryptozoospermic patients, applying the criteria outlined
in Fig. 1.
The PCA based on clinical parameters showed a clear separa-
tion of the Control and the cryptozoospermic patient groups on
the first principal component (PC1), which accounts for 30.3% of
the total variation (Supplementary Fig. S1A). The variation on the
PC1 was mostly due to the following variables: sperm concentra-
tion in the ejaculate, percentage of tubules with elongated sper-
matids, and FSH concentration (Supplementary Fig. S1B).
In order to evaluate whether the cryptozoospermic
patients can be separated from the Controls and whether they
can be divided into subgroups, we applied HCPC method
and identified three different clusters. The first cluster contained
all 160 Control patients and 2 cryptozoospermic patients
(Cluster 1). The remaining cryptozoospermic patients were
equally subdivided into Clusters 2 (n ¼65) and 3 (n ¼65), respec-
tively (Fig. 2).
(n=162)
(n=65)
(n=65)
−2
0
2
4
6
−2 0 2 4
Dim1 (30.3%)
Dim2 (13.2%)
Cluster
1
2
3
Figure 2. Hierarchical clustering on principal components revealed
three distinctive clusters among the 292 included patients. The
clustering is shown in a principal component analysis (PCA) space that
was based on the clinical and histological parameters of each patient.
The cluster analysis identified three different clusters. The first cluster
contained all 160 Control patients and 2 cryptozoospermic patients
(Cluster 1). The remaining cryptozoospermic patients were equally
subdivided into Clusters 2 (n ¼65) and 3 (n¼65), respectively. Each dot
represents one patient and the lines that group the dots show the
clustering. The percentages on the x- and y-axes show the variation of
the data.
Two subgroups among men with cryptozoospermia | 895
Sub-classification of the cryptozoospermic
patient cohort
The main factors (leading factors) determining the three clusters
were the percentage of tubules with elongated spermatids for
Cluster 1, the percentage of tubules with spermatocytes for
Cluster 2, and the percentage of tubules containing solely Sertoli
cells for Cluster 3 (Fig. 3A, Table 1). The composition of testicular
tissues in these clusters is therefore fundamentally different,
with the majority of tubules showing an arrest at the spermato-
cyte level (median: 51.5%) in Cluster 2 compared to the majority
of tubules devoid of germ cells (median SCO: 63.5%; median TS:
11%) in Cluster 3 (Fig. 3B, Table 1). Importantly, although these
represent the predominant tubular phenotype in each cluster,
the percentage of tubules with full germ cell differentiation (up
to the elongated spermatids) was comparable in cryptozoosper-
mic Clusters 2 (median: 4%) and 3 (median: 3.5%), respectively
(Fig. 3C, Table 1).
Comparison of clinical parameters between
patient subgroups
In line with the comparable percentage of tubules containing
elongated spermatids in Clusters 2 and 3, we found comparable
sperm retrieval rates after testicular sperm extraction (TESE)
between the two cryptozoospermic groups (Supplementary
Fig. S2A). Additional clinical parameters, apart from the leading
factors, that showed differences between the two cryptozoospermic
bbb
ccc
aaa
bbb
ccc
bbb
bbb
ccc
aaa
aaa
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
20
40
60
80
100
% tubules with ES
Cluster 1 Cluster 2 Cluster 3
Cluster 2
Cluster 1
aaa
bbb
ccc
ccc
aaa
aaa
bb
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
20
40
60
80
100
% tubules with RS
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
20
40
60
80
100
% tubules with SC
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
20
40
60
80
100
% tubules with SG
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
20
40
60
80
100
% tubules with SCO
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
20
40
60
80
100
% tubules with TS
6%
2%
9% 3%
61%
19%
Elongated spermatids
Round spermatids
Spermatocytes
Spermatogonia
Sertoli cell only
Tubular shadows
Clu
Clu
ste
ste
r
1
1
Clu
Clu
ste
ste
r
2
Percentage of tubules with:
Cluster 3
A
B
C
Figure 3. Comparison of the histological status of the testis tissues from patients in the three clusters. (A) Boxplots comparing the percentage of the
most advanced germ cell types in the seminiferous tubules of each cluster. Cluster 1 (teal): n¼162 (160 Control and 2 cryptozoospermic patients),
Cluster 2 (orange): n¼65 cryptozoospermic patients, Cluster 3 (purple): n¼65 cryptozoospermic patients. ES, elongated spermatids; RS, round
spermatids; SC, spermatocytes; SG, spermatogonia; SCO, Sertoli cell only; TS, tubular shadow. Boxplot elements are defined as follows: center line:
median; box limits: upper and lower quartiles; whiskers: 1.5interquartile range; points: outliers. Statistically significant differences between the
clusters are shown by letters: (a) Cluster 1 is different from Cluster 2, (b) Cluster 1 is different from Cluster 3, (c) Cluster 2 is different from Cluster 3.
Significance levels are shown by the number of letters, e.g. a¼P<0.05, aa ¼P<0.01, aaa ¼P<0.001. (B) Representative images of periodic acid-Schiff-
stained tubules with full spermatogenesis (Cluster 1), meiotic arrest at the spermatocyte stage (Cluster 2), and Sertoli cell only phenotype (Cluster 3).
Scale bar: 100mm. (C) Pie charts showing the means of the percentage of the most advanced germ cell type in the seminiferous tubules of patients from
each cluster.
896 | Sch€
ulke et al.
groups are outlined below (Fig. 4, Table 1) and an overview of the
results of all the parameters analyzed can be found in
Supplementary Table S2.
Statistical analyses revealed significantly larger testicular vol-
umes in the Cluster 2 patients compared to the Cluster 3 patients
(Fig. 4A). Moreover, LH levels were significantly lower in Cluster 2
patients (median: 4.8 U/l) compared to Cluster 3 (median: 6.7 U/l,
Fig. 4B). While the LH levels in both cryptozoospermic groups
were still within the internal reference values, the FSH values in
Cluster 3 patients (median: 20.8 U/l) were significantly higher
compared to the reference. The FSH values in Cluster 2 were sig-
nificantly lower (median: 6.4 U/l, Fig. 4C, Supplementary Table
S2). The T and free testosterone levels were significantly higher
in the Cluster 2 compared to Cluster 3 patients but still within
the reference values (Fig. 4D and E, Supplementary Table S2). In
contrast to the differences in hormonal values, there was no en-
richment of specific FSHB-211 SNPs between the groups
(Supplementary Fig. S2B). Interestingly, a significant difference
could be found regarding the age of the patients with Cluster 2
patients being slightly younger compared to Control patients in
Cluster 1 but not Cluster 3 (Fig. 4F).
In order to evaluate whether the three groups had different
outcomes with regard to MAR treatment, we analyzed the ICSI
results of 160 patients included in our study (n¼88 Cluster 1,
n¼33 Cluster 2, n ¼39 Cluster 3). Interestingly, the three
groups showed comparable percentages of embryo transfers,
biochemical and clinical pregnancies, and live births (Fig. 5,
Supplementary Table S4).
Histological analysis of the spermatogonial
compartment
As alterations in the stem cell compartment of cryptozoospermic
patients have been found, specifically a reduction of A
dark
sper-
matogonia and an increased number of PIWIL4
þ
cells (Di Persio
et al., 2021), we evaluated whether these observed changes are
shared between the two clusters of cryptozoospermic patients.
We first examined the spermatogonial compartment by analyz-
ing the pan-spermatogonial marker MAGEA4 (Fig. 6A and B,
Supplementary Table S3). This analysis revealed a significant re-
duction of MAGEA4
þ
spermatogonia per tubule in both crypto-
zoospermic groups compared to the Control cluster but did not
show a significant difference between the two cryptozoospermic
clusters. Furthermore, Spearman correlation analysis revealed a
significant positive correlation between the number of spermato-
gonia per tubule and the percentage of tubules containing elon-
gated spermatids represented as BK-Score (Fig. 6C).
To assess changes in the population of reserve spermatogonial
stem cells, we quantified the number of the morphologically de-
fined A
dark
spermatogonia (Fig. 6D, Supplementary Table S3).
While we found significantly lower numbers of the A
dark
sperma-
togonia in the cryptozoospermic clusters compared to the
Table 1. Hormonal and histological parameters of patients from the three clusters.
Patient groups
Hormone parameters (normal range) Histological parameters of tubules
Testicular volume
(12–30 ml) FSH (1–7 U/l) LH (2–10 U/l) % ES % SC % SCO
Cluster 1 (n¼162) 24.88 (20, 29.5) 3.35 (2.5, 4.38) 2.7 (1.9, 3.9) 81 (77, 86) 10.25 (7, 13.5) 0 (0, 0.5)
Cluster 2 (n¼65) 14.5 (11.5, 18.5) 6.4 (4.3, 13) 4.8 (3, 5.8) 4 (0, 24.5) 51.5 (34.5, 66.5) 3.5 (1, 13.5)
Cluster 3 (n¼65) 11.5 (8.5, 15) 20.8 (15.8, 24.1) 6.7 (4.8, 8.8) 3.5 (0, 9.5) 6 (0.5, 13.5) 63.5 (43, 82)
For each cluster (Cluster 1: 160 Control and 2 cryptozoospermic patients, Cluster 2 and 3: cryptozoospermic patients) and parameter, the median is shown, as
well as the 25th and 75th percentile in brackets (median, 25th, 75th percentile). ES, elongated spermatids; SC, spermatocytes; SCO, Sertoli cell only; ml, milliliter; l,
liter; U, units.
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
20
40
60
Testosterone (nmol/l)
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
500
1000
1500
Free Testosterone (pmol/l)
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
20
40
60
Age (years)
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
20
40
60
Testicular volume (ml)
bbb
ccc
aaa
aa
bbb
ccc
cc
aaa
bbb
bbb
bb
ccc
c
aaa
A
B
C
D
E
F
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
10
20
30
40
50
FSH (U/l)
Cluster 1 Cluster 2 Cluster 3
Cluster 1 Cluster 2 Cluster 3
0
5
10
15
20
LH (U/l)
Figure 4. Comparison of clinical parameters among the three patient clusters. Boxplots comparing the (A) testicular volume, (B) LH, (C) FSH, (D)
testosterone, (E) free testosterone values, and (F) age of the patients of the three clusters. Cluster 1 (teal): n ¼162 (160 Control and 2 cryptozoospermic
patients), Cluster 2 (orange): n ¼65 cryptozoospermic patients, Cluster 3 (purple): n ¼65 cryptozoospermic patients. Boxplot elements are defined as
follows: center line: median; box limits: upper and lower quartiles; whiskers: 1.5interquartile range; points: outliers. Statistically significant
differences between the clusters are shown by letters: (a) Cluster 1 is different from Cluster 2, (b) Cluster 1 is different from Cluster 3, (c) Cluster 2 is
different from Cluster 3. Significance levels are shown by the number of letters, e.g. a ¼P<0.05, aa ¼P<0.01, aaa ¼P<0.001.
Two subgroups among men with cryptozoospermia | 897
Control samples (Fig. 6E), no significant difference between the
two cryptozoospermic clusters was observed.
In order to assess if the reduction of reserve spermatogonial
stem cells is linked to the spermatogenic state, we performed a
correlation analysis. This showed that the number of A
dark
sper-
matogonia is positively correlated with the percentage of tubules
containing elongated spermatids (Fig. 6F).
Interestingly, the proportion of PIWIL4
þ
cells, which are con-
sidered the most undifferentiated spermatogonial stem cell type
(Guo et al., 2018; Sohni et al., 2019; Shami et al., 2020; Di Persio
et al., 2021), among the MAGEA4
þ
cells is increased in both cryp-
tozoospermic groups (Fig. 6G and H, Supplementary Table S3).
Moreover, the percentage of PIWIL4
þ
spermatogonia was nega-
tively correlated with the percentage of tubules containing elon-
gated spermatids (Fig. 6I). Hence, both cryptozoospermic groups
displayed shared alterations of their spermatogonial compart-
ments with regard to the lower number of A
dark
and higher num-
ber of PIWIL4
þ
spermatogonia, respectively.
Discussion
Heterogeneity of clinical parameters—such as testicular tissue
composition and endocrine levels—in cryptozoospermic men has
hampered the identification of the etiology behind this infertility
phenotype (Di Persio et al., 2021). The almost complete absence of
sperm from the ejaculate often necessitates TESE to fulfill the
wish for fatherhood in these patients, providing the unique op-
portunity to evaluate alterations in the spermatogenic process.
First analyses point to changes already in the spermatogonial
compartment indicating stem cell exhaustion and the existence
of distinct severity phenotypes. Considering the heterogeneity of
cryptozoospermic patients, we sought to assess if subgroups can
be identified by cluster analysis of clinical parameters, which is
the prerequisite to unveil the etiology of cryptozoospermia.
Cluster analyses have been successfully used in other medical
fields for the identification of clinical subgroups in heteroge-
neous patient cohorts, building the basis for specialized treat-
ments and a better prediction of the prognosis (Burgel et al., 2014;
Zhu et al., 2019). The PCA-based cluster analysis revealed two
subgroups of cryptozoospermic patients characterized by a mei-
otic arrest- and SCO phenotype in the majority of testicular tis-
sues, respectively. We consider the following scenario which
could lead to the two clusters of cryptozoospermic patients.
Cryptozoospermia may be a progressive disease, which develops
from a meiotic arrest to a SCO phenotype, potentially caused by
stem cell exhaustion over time. Also, these two histologically dis-
tinct phenotypes may be caused by distinct underlying genetic
causes. It is of note that these two scenarios are not necessarily
mutually exclusive. The findings supporting the first scenario are
the lower testicular volume, higher FSH levels, and higher num-
ber of tubules displaying a SCO phenotype in Cluster 3 compared
to Cluster 2. While we found no significant association between
the patient age of cryptozoospermic patients and the cluster as-
signment, which may be expected if patients would transition
from one phenotype to the other over time, the start of germ cell
differentiation failure and the speed of its progression may be
different among cryptozoospermic men compared to others.
However, a study performed on men who underwent multiple
TESE owing to obstructive and non-obstructive azoospermia
reported that the sperm retrieval rate was comparable in all the
attempts independent of the patient diagnosis, suggesting that
the spermatogenic rate in these men did not change over time
(Vernaeve et al., 2006). For further validation of this hypothesis
close monitoring of sperm parameters in individual oligozoosper-
mic and cryptozoospermic patients over time is necessary. If this
Cluster 1Cluster 2Cluster 3
Embryo
transfer
97.7%
Embryo
transfer
90.9%
Embryo
transfer
87.2%
No embryo transfer 2.3%
No embryo transfer 9,1%
No embryo transfer 12.8%
No
pregnancy
59.1%
No
pregnancy
51.5%
No
pregnancy
51.3%
Clinical pregnancy
35.2%
Clinical pregnancy
36.4%
Clinical pregnancy
30.8%
Biochemical pregnancy 3.4%
Biochemical pregnancy 3%
Biochemical pregnancy 5.1%
No live birth
8%
No live birth
6.1%
No live birth
10.3%
Live birth
27.3%
Live birth
30.3%
Live birth
20.5%
Figure 5. Sankey diagram summarizing the ICSI results in the patient cohort. From left to right the percentage of patients with or without embryo
transfer, clinical and biochemical pregnancies, no pregnancy as well as live birth or no live birth are depicted for each cluster of patients (Cluster 1
(teal), Cluster 2 (orange), Cluster 3 (purple)). Numbers represent the percentage of patients in each category. Absolute numbers are shown in
Supplementary Table S4.
898 | Sch€
ulke et al.
hypothesis can be confirmed, these patients may benefit from
cryopreservation of sperm early after the diagnosis, avoiding
the surgical approach of TESE in order to fulfill their wish
for fatherhood.
The existence of two subgroups, largely presenting a meiotic-
arrest or SCO phenotype, could also support the second scenario
of distinct underlying genetic causes in cryptozoospermic
patients. To confirm this hypothesis, genetic analyses of candi-
date genes previously associated with these histological pheno-
types (Oud et al., 2019) are required. One example demonstrating
that this approach can be successful involves the recently
revealed mutations in the M1AP gene, which result in phenotypes
ranging from azoospermia owing to complete meiotic arrest to
cryptozoospermia (Wyrwoll et al., 2020). For counseling of respec-
tive patients, it will be relevant to assess whether those sperm
that are formed are also carrying respective loss of function var-
iants and could be passed on to the offspring.
We tested for an association with the FSHB-211 G>T polymor-
phism between Clusters 2 and 3, as patients with unexplained
azoospermia presenting the TT-genotype in the FSHB-211 locus
show decreased testicular volume and increased FSH levels
compared to GG patients (Busch et al., 2019). However, we could
not find a specific enrichment for this TT-genotype in Cluster 3,
despite the similarities regarding testicular volume and FSH lev-
els. In line with our findings, cluster analysis of patients with ob-
structive and non-obstructive azoospermia also revealed that
cluster formation did not depend on FSHB genotypes (Schubert
et al., 2020), although among men with unexplained infertility
this parameter was the strongest segregation marker (Krenz
et al., 2021).
Intriguingly, despite these highly distinct histological pheno-
types between the two cryptozoospermic clusters, the analysis of
the spermatogonial compartment has revealed shared altera-
tions. In both clusters, we found a reduction of A
dark
spermatogo-
nia and an increase in PIWIL4
þ
spermatogonia, in line with
previous data (Di Persio et al., 2021). This suggests that the altera-
tions of the spermatogonial compartment are a general response
mechanism to the increased loss of more advanced germ cells—
that is shared in both cryptozoospermic clusters irrespective of
the potentially distinct underlying causes and the elevated FSH
levels in Cluster 3. Moreover, we found a correlation between the
reduction of the A
dark
spermatogonia and the percentages of
Figure 6. Histological evaluation and comparison of patient samples. From Cluster 1 (teal, n¼13 (12 Control and 1 cryptozoospermic patients)),
Cluster 2 (orange, n¼13 cryptozoospermic patients), and Cluster 3 (purple, n ¼13 cryptozoospermic patients). Significance levels: P<0.05, P<0.01,
P<0.001. (A) Representative melanoma-associated antigen 4 (MAGEA4)immunohistochemical stainings in tubules from a Control and a
cryptozoospermic (Crypto) sample. Scale bars: 100 mm. (B) Boxplot comparing the quantification of MAGEA4
þ
spermatogonia per tubule. (C) Scatterplot
and linear regression showing the correlation between the Bergmann Kliesch (BK)-Score (0–10, indicating 0–100% of tubules with elongated spermatids)
and the number of MAGEA4
þ
spermatogonia per tubule. (D) Representative MAGEA4 stainings in tubules from a Control and Crypto sample. Inlay
showing a representative A
dark
spermatogonium, discernible by the rarefaction zone in the nucleus. Scale bar: 100mm. (E) Boxplot comparing the
quantification of A
dark
spermatogonia per MAGEA4
þ
spermatogonia. (F) Scatterplot and linear regression showing the correlation between the BK Score
(0–10, indicating 0–100% of tubules with elongated spermatids) and the proportion of A
dark
spermatogonia per MAGEA4
þ
spermatogonia. (G)
Representative Piwi Like RNA-Mediated Gene Silencing 4 (PIWIL4) stainings in tubules from a Control and Crypto sample. Scale bar: 100mm. (H) Boxplot
showing the percentage of PIWIL4
þ
spermatogonia per MAGEA4
þ
spermatogonia after quantification. (I) Scatterplot and linear regression showing the
correlation between the BK Score (0–10, indicating 0–100% of tubules with elongated spermatids) and the percentage of PIWIL4
þ
spermatogonia per
MAGEA4
þ
spermatogonia. Boxplot elements are defined as follows: center line: median; box limits: upper and lower quartiles; whiskers: 1.5
interquartile range; points: outliers. Details about the absolute numbers and statistical analyses are provided in Supplementary Table S3.
Two subgroups among men with cryptozoospermia | 899
tubules containing elongated spermatids, which ranged from 1%
to 46% in both clusters. Those patients with a lower percentage
of tubules with elongated spermatids are also those with the low-
est numbers of reserve stem cells (A
dark
); the latter are required
to safeguard genetic integrity of the germline and to ensure sus-
tainability of spermatogenesis. Consequently, in cryptozoosper-
mic men the genetic integrity of sperm may worsen over time.
Provided that this can be demonstrated in follow-up studies,
cryopreservation of sperm should be offered to this patient co-
hort early after diagnosis. Importantly, the comparative evalua-
tion of MAR results between the two cryptozoospermic groups
does not point to an obvious impairment of sperm integrity based
on the within range percentages of clinical pregnancies and live
births (Nordhoff et al., 2015; Corona et al., 2019; ESHRE Clinic PI
Working Group et al., 2021).
While we have uncovered the existence of subgroups within
the cohort of cryptozoospermic patients, a limitation of this
study is that follow-up analyses are required to uncover the un-
derlying etiologies. As this study was limited to clinical data from
humans, we do not have the opportunity to assess testicular bi-
opsies obtained at consecutive time points. However, this would
be the prerequisite to draw immediate conclusions on the pro-
gressiveness of cryptozoospermia. Another limitation of the cur-
rent study is that comprehensive genetic analyses remain to be
performed to uncover further loss of function variants underly-
ing cryptozoospermia.
In conclusion, employing cluster analysis, we have identified
that cryptozoospermic patients can be divided into two distinct
subgroups. The leading parameters to distinguish these two clus-
ters are the histological composition of testicular tissues, testicu-
lar volumes, and FSH levels, which may reflect distinct stages of a
progressive disease phenotype. In line with this, we have shown
that the alterations in the spermatogonial compartment are
shared in all cryptozoospermic patients and that they become
more severe with a lower number of tubules containing full sper-
matogenesis. This distinction of two clinical phenotypes of crypto-
zoospermia is, therefore, the first step to evaluate if
cryptozoospermia constitutes a progressive disease and to assess
potentially distinct underlying genetic causes in follow-up studies.
Supplementary data
Supplementary data are available at Human Reproduction online.
Data availability
The data underlying this article are available in the article and in
its online supplementary material.
Acknowledgements
We thank Heidi Kersebom and Elke K€
oßer for histological evalua-
tion of testicular tissues. We also thank Sabine Forsthoff and
Nicole Terwort for endocrinological measurements and the
andrological laboratory for semen analyses. Finally, we thank Dr
Sandra Laurentino and Dr Maria Schubert for their expert opin-
ion on this manuscript.
Authors’ roles
Study conception and design: SDP, NN; Supervision: SDP, NN;
Acquisition and evaluation of clinical data: J-FC, SK, VN, HMB;
Lab work: LCS, JW; Data and bioinformatic analyses: LCS, SDP;
Writing-Original Draft; LCS, SDP, and NN; All authors were in-
volved in editing, read and approved the final version of
the manuscript.
Funding
This work was supported by the German Research Foundation
(grant number CRU326 to SDP and NN). We thank the Faculty of
Medicine of the University of M €
unster for the financial support
through the MedK-program. We acknowledge support from the
Open Access Publication Fund of the University of M€
unster.
Conflict of interest
The authors declare no conflict of interest.
References
Bergmann M, Kliesch S. Testicular biopsy and histology. In:
Nieschlag E, Behre H, Nieschlag S (eds). Andrology. Berlin,
Heidelberg: Springer, 2010,155–167.
Burgel P-R, Paillasseur J-L, Roche N. Identification of clinical pheno-
types using cluster analyses in COPD patients with multiple
comorbidities. Biomed Res Int 2014;2014:420134.
Busch AS, T€
uttelmann F, Cremers J-F, Schubert M, Nordhoff V,
Sch€
uring AN, Zitzmann M, Gromoll J, Kliesch S. FSHB-211 G>T
polymorphism as predictor for TESE success in patients with
unexplained azoospermia. J Clin Endocrinol Metab 2019;
104:2315–2324.
Busch AS, T€
uttelmann F, Zitzmann M, Kliesch S, Gromoll J. The
FSHB −211G>T variant attenuates serum FSH levels in the supra-
physiological gonadotropin setting of Klinefelter syndrome. Eur J
Hum Genet 2015;23:700–703.
Caldeira-Brant AL, Martinelli LM, Marques MM, Reis AB, Martello R,
Almeida FRCL, Chiarini-Garcia H. A subpopulation of human
Adark spermatogonia behaves as the reserve stem cell.
Reproduction 2020;159:437–451.
Clermont Y. Renewal of spermatogonia in man. Am J Anat 1966;
118:509–524.
Clermont Y. The cycle of the seminiferous epithelium in man. Am J
Anat 1963;112:35–51.
Corona G, Minhas S, Giwercman A, Bettocchi C, Dinkelman-Smit M,
Dohle G, Fusco F, Kadioglou A, Kliesch S, Kopa Z et al. Sperm re-
covery and ICSI outcomes in men with non-obstructive azoosper-
mia: a systematic review and meta-analysis. Hum Reprod Update
2019;25:733–757.
Di Persio S, Neuhaus N. Human spermatogonial stem cells and their
niche in male (in)fertility: novel concepts from single-cell RNA-
sequencing. Hum Reprod 2023;38:1–13.
Di Persio S, Tekath T, Siebert-Kuss LM, Cremers J-F, Wistuba J, Li X,
Meyer zu H€
orste G, Drexler HCA, Wyrwoll MJ, T€
uttelmann F et al.
Single-cell RNA-seq unravels alterations of the human sperma-
togonial stem cell compartment in patients with impaired sper-
matogenesis. Cell Rep Med 2021;2:100395.
ESHRE Clinic PI Working Group, Vlaisavljevic V, Apter S, Capalbo A,
D’Angelo A, Gianaroli L, Griesinger G, Kolibianakis EM, Lainas G,
Mardesic T, Motrenko T et al. The Maribor consensus: report of an
expert meeting on the development of performance indicators
for clinical practice in ART. Hum Reprod Open 2021;2021:hoab022.
Guo J, Grow EJ, Mlcochova H, Maher GJ, Lindskog C, Nie X, Guo Y,
Takei Y, Yun J, Cai L et al. The adult human testis transcriptional
cell atlas. Cell Res 2018;28:1141–1157.
900 | Sch€
ulke et al.
Krenz H, Sansone A, Kliesch S, Gromoll J, Schubert M. FSHB genotype
identified as a relevant diagnostic parameter revealed by cluster
analysis of men with idiopathic infertility. Front Endocrinol
(Lausanne) 2021;12:780403.
L^
e S, Josse J, Husson F. FactoMineR: an R package for multivariate
analysis. J Stat Soft 2008;25:1–18.
Nordhoff V, Fricke RK, Sch€
uring AN, Zitzmann M, Kliesch S.
Treatment strategies for severe oligoasthenoteratozoospermia
(OAT) (<0.1 million/mL) patients. Andrology 2015;3:856–863.
Oud MS, Volozonoka L, Smits RM, Vissers LELM, Ramos L, Veltman
JA. A systematic review and standardized clinical validity assess-
ment of male infertility genes. Hum Reprod 2019;34:932–941.
R Development Core Team. R: A Language and Environment for
Statistical Computing: Reference Index. Vienna: R Foundation for
Statistical Computing, 2010.
Schubert M, Kaldewey S, P�
erez Lanuza L, Krenz H, Dugas M, Berres S,
Kliesch S, Wistuba J, Gromoll J. Does the FSHB c.-211G>T poly-
morphism impact Sertoli cell number and the spermatogenic po-
tential in infertile patients? Andrology 2020;8:1030–1037.
Shami AN, Zheng X, Munyoki SK, Ma Q, Manske GL, Green CD,
Sukhwani M, Orwig KE, Li JZ, Hammoud SS. Single-cell RNA se-
quencing of human, macaque, and mouse testes uncovers con-
served and divergent features of mammalian spermatogenesis.
Dev Cell 2020;54:529–547.e12.
Sohni A, Tan K, Song H-W, Burow D, Rooij DGd, Laurent L, Hsieh
T-C, Rabah R, Hammoud SS, Vicini E et al. The neonatal and adult
human testis defined at the single-cell level. Cell Rep 2019;26:
1501–1517.e4.
T€
uttelmann F, Luetjens CM, Nieschlag E. Optimising workflow in
andrology: a new electronic patient record and database. Asian J
Androl 2006;8:235–241.
van Alphen M, van de Kant HJG, de Rooij DG. Depletion of the sper-
matogonia from the seminiferous epithelium of the rhesus mon-
key after X irradiation. Radiat Res 1988;113:473–486.
Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of
simple methods for the estimation of free testosterone in serum.
J Clin Endocrinol Metab 1999;84:3666–3672.
Vernaeve V, Verheyen G, Goossens A, Van Steirteghem A, Devroey P,
Tournaye H. How successful is repeat testicular sperm extraction
in patients with azoospermia? Hum Reprod 2006;21:1551–1554.
World Health Organization. WHO Laboratory Manual for the
Examination and Processing of Human Semen. Geneva: World Health
Organization, 2010.
Wyrwoll MJ, Temel S¸G, Nagirnaja L, Oud MS, Lopes AM, Heijden
GWvd, Heald JS, Rotte N, Wistuba J, W€
oste M et al.; GEMINI
Consortium. Bi-allelic mutations in M1AP are a frequent cause of
meiotic arrest and severely impaired spermatogenesis leading to
male infertility. Am J Hum Genet 2020;107:342–351.
Zhu H, Wu C, Jiang N, Wang Y, Zhao J, Xu D, Wang Q, Li M, Zeng X.
Identification of 6 dermatomyositis subgroups using principal
component analysis-based cluster analysis. Int J Rheum Dis 2019;
22:1383–1392.
Two subgroups among men with cryptozoospermia | 901
# The Author(s) 2024. Published by Oxford University Press on behalf of European Society of Human Reproduction and Embryology.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (https://creativecommons.
org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly
cited. For commercial re-use, please contact journals.permissions@oup.com
Human Reproduction, 2024, 39, 892–901
https://doi.org/10.1093/humrep/deae013
Original Article
