The significance of human spermatozoa vacuoles can be elucidated by a novel procedure of array comparative genomic hybridization

Abstract

STUDY QUESTION

Is there an association between spermatozoon genomic stability and vacuolar morphology and location?

SUMMARY ANSWER

The genomic stability of spermatozoa is associated with specific characteristics of vacuolar morphology (depth) and location (cellular compartment, i.e. nucleus and equatorial region).

WHAT IS KNOWN ALREADY

Genetic anomalies in sperm are correlated with semen abnormalities, yet the advantage of morphologically based selection of spermatozoa for IVF according to current criteria is controversial. Selection criteria based on the number of vacuoles and their size have been proposed and are widely applied. Nevertheless, it has not improved the ICSI success rates, suggesting the currently used vacuole criteria are incomplete.

STUDY DESIGN, SIZE, DURATION

Normal sperm according to Motile Sperm Organelle Morphology Examination criteria (MSOME) and common vacuole grading were evaluated. An additional evaluation of sperm vacuole morphology according to novel vacuole criteria (i.e. location and depth) was conducted. An assessment to align these specific vacuolar morphology features with genomic stability was conducted among spermatozoa from infertile patients and healthy fertile donors aged 24–38 between June 2015 and July 2016.

PARTICIPANTS/MATERIALS, SETTING, METHODS

Single spermatozoa (n = 53) from 16 infertile patients and 14 fertile donors were morphologically and genetically evaluated. Each spermatozoon was examined morphologically, by ultra-magnification ×6300, and genetically by a novel comparative genomic hybridization protocol, without the use of reference DNA, to assess chromosomal instability as evident by copy number variations (CNV).

MAIN RESULTS AND THE ROLE OF CHANCE

We established an association between genomic stability and vacuolar morphology as a base for a new classification according to novel vacuolar criteria, specifically depth and location. Genomic instability was found to be related to these two main features of vacuoles and, surprisingly not to the number and size of vacuoles as in the previously proposed classifications. High CNV spermatozoa were characterized by vacuoles located in the nucleus and/or equatorial segment or by deep vacuoles, while, low CNV spermatozoa were characterized by a complete lack of vacuoles or non-deep vacuoles not located in the nucleus/equatorial segment. A putative threshold of ~265 CNV was deduced to distinguish between genetically stable and unstable spermatozoa, and 94% of the tested spermatozoa segregated accordingly.

LIMITATIONS REASONS FOR CAUTION

A relatively small sample of spermatozoa were examined—53 in total. However, the association between vacuoles location and morphology and genomic stability was significant. This is the first study evaluating spermatozoon genomic stability with respect to vacuole morphology according to novel vacuole criteria (i.e. location and depth) and further investigation is warranted to verify the value of these criteria in larger sample size clinical studies.

WIDER IMPLICATIONS OF THE FINDINGS

Our results, which are based on spermatozoon vacuoles morphological classification and genomic parameters, indicate an association between vacuoles morphology and location and genomic stability. The data presented herein suggest the existence of subpopulations of spermatozoa potentially appropriate for IVF–ICSI, as they appear normal according to the current MSOME and vacuoles classification, however they are almost certainly genetically damaged. As current criteria have yet to achieve an unequivocal evaluation of the implantation potential of a given spermatozoon, we propose novel criteria, based on specific vacuolar morphological traits; depth and location, as these were found aligned with genomic findings.

STUDY FUNDING/COMPETING INTEREST(S)

No funding was received for this study. The authors have no conflict of interest to declare.

TRIAL REGISTRATION NUMBER

N/A.

array comparative genomic hybridization, copy number variation, sperm selection, sperm genomic stability, sperm vacuoles, ICSI

Introduction

IVF by ICSI involve the use of specialized micromanipulation and microscopic tools, allowing embryologists to select and pick up one motile spermatozoon for injection into the oocyte. In ICSI cycles, the pregnancy rate per aspiration is only 27.8% (Calhaz-Jorge et al., 2016) and the mean delivery rate per aspiration is 21.1% (Kupka et al., 2014). One of the obstacles in this procedure is defining characteristics of a top-quality spermatozoon. Individual sperm morphology assessed at the moment of ICSI correlates well with fertilization outcome (De Vos et al., 2003). A promising method recently used for observing spermatozoon, termed Motile Sperm Organelle Morphology Examination (MSOME), enables evaluation of the morphology of six subcellular organelles (acrosome, postacrosomal lamina, neck, mitochondria, tail and nucleus) of motile spermatozoon in real-time under ultra-magnification (Bartoov et al., 2002; Setti et al., 2013). Morphological normalcy of the entire spermatozoon, according to MSOME criteria, was positively associated with ICSI fertilization, but not with pregnancy outcome, yet morphological normalcy of the spermatozoon nucleus was positively associated with both the fertilization rate and pregnancy outcome (Bartoov et al., 2002). Based on the MSOME results, intracytoplasmic morphologically selected sperm injection (IMSI) procedure was introduced (Bartoov et al., 2003; Setti et al., 2013). The possible advantage of IMSI over ICSI has been examined and has so far achieved conflicting results. It has been reported to be more beneficial by several investigators while others found the procedures equally efficient (Antinori et al., 2008; Mauri et al., 2010; Balaban et al., 2011; Knez et al., 2011, 2012; Klement et al., 2013). Possibly we do not have full understanding of the spermatozoa structural complexity, consequently leading to an underestimation of the significance of sperm vacuole morphology on implantation rates (Boitrelle et al., 2011; Kacem et al., 2010; Montjean et al., 2012).

It is well known that chromosomal aneuploidies and structural aberrations represent the leading causes of pregnancy loss and the leading genetic causes for developmental disabilities (Jacobs, 1992; Hassold, 1998; Hassold and Hunt, 2001), and thus they present significant limiting factors for the success of both, natural conception and IVF techniques. Moreover, IVF–ICSI is still characterized by low implantation and high abortion rates (Mansour et al., 2014). To date, studies aimed at identifying chromosomal anomalies in spermatozoon are performed using either of the following techniques: karyotyping after fusion of spermatozoon with hamster oocytes followed by fluorescent in-situ hybridization (FISH) with two or more DNA probes on de-condensed spermatozoon nucleus (Templado et al., 2005, 2011), or use of a combination of spermatozoon DNA and reference DNA to perform array comparative genomic hybridization (aCGH) on a bacterial artificial chromosome (BAC) array (Patassini et al., 2013). This new technique has shown to be been advantageous compared to traditional karyotyping as it can provide significantly higher resolution and thus allows for the identification of aneuploidies and unbalanced submicroscopic chromosome alterations. Indeed aCGH is also used in diagnostic clinical practice (such as in pre-natal and pre-implantation genetic diagnosis), yet single nucleotide polymorphism (SNP) aCGH has only recently been reported on a spermatozoon (Patassini et al., 2013; Garolla et al., 2015).

Since the first clinical study (Berkovitz et al., 2006) that investigated the association between sperm nuclear vacuoles and ICSI outcome, it has been continuously investigated and reported on for over a decade (Vanderzwalmen et al., 2008; Hammoud et al., 2013). Vacuoles are normally observed in 96% of patients and donors’ ejaculates, mostly at the tip or middle area of the sperm heads. However, the average incidence of normal sperm exhibiting large vacuoles was 4.6 and 4.2% in the patients and donors, respectively. Yet, no significant chromosomal instability difference was detected between sperm bearing large or small vacuoles (Watanabe et al., 2011). In terms of vacuole location, high heterogeneity has been reported. A significant decrease in the DNA fragmentation rate was observed in MSOME-selected spermatozoa with vacuoles located in the anterior half of the spermatozoa head compared with both MSOME-selected spermatozoa with vacuoles located in the posterior half of the head and unselected spermatozoa (Hammoud et al., 2013). Since the number and size of nuclear vacuoles were negatively associated with blastocyst development, until Day 5, grading was proposed as follows: grade I, no vacuoles; grade II, ≤2 small vacuoles; grade III, ≥1 large vacuole; grade IV, large vacuoles with other abnormalities (Vanderzwalmen et al., 2008). Further research proposed that large vacuoles might have a negative influence on chromatin immaturity (Perdrix and Rives, 2013). Recently, with the aid of aCGH, it was shown that spermatozoa with larger nuclear vacuoles are characterized by higher rates of abnormal karyotypes and other complex chromosomal aberrations (Garolla et al., 2015).

The objective of this study was to investigate the association between human spermatozoa vacuolar characteristics and genomic stability, at a single cell resolution. We examined single spermatozoa derived from both infertile patients who went through ICSI procedure and from fertile donors. Various vacuole traits, including previous (number and size of vacuoles) and novel (location and morphology, i.e. depth) traits, were assessed morphologically by ultra-magnification. A novel aCGH protocol (following PCR amplification) was developed, optimized and applied to comprehensively screen the spermatozoon’s chromosomes and evaluate copy number variations (CNV).

Materials and Methods

Study population and sperm morphological classification

Up to two single spermatozoa were collected and isolated from 16 infertile patients that underwent an ICSI procedure (total of 31 spermatozoa). Up to two single spermatozoa were also collected from 14 healthy, proven fertile volunteers (total of 22 spermatozoa) and were examined in parallel for comparison purposes. Spermatozoa that were included in this study were those that follow our previous morphological criteria and MSOME (Berkovitz et al., 2006), such as normal head configuration without any other abnormality in the neck, mitochondria, postacrosomal lamina or tail. They were then also graded according to their vacuoles structural examination, i.e. grades I–III, according to previous studies (Vanderzwalmen et al., 2008). A total of 53 spermatozoa from 30 patients of two sperm source groups were analyzed.

Sperm cell observation

Each spermatozoon was examined under immersion oil by an inverted microscope (Nikon Eclipse Ti-S) equipped with Nomarski optics using UPlan Apo ×100/1.4 NA objective lens (note: spermatozoa were not fixed during morphological inspection). The images were captured by an Invenio 3SII (Deltapix) 3.1MP digital CMOS color camera with 6.6 mm × 4.9 mm sensor size and 3.2 × 3.2 µm pixel size. A camera adapter with a magnification of ×0.5 was used. Images were viewed on a 24′ monitor with a resolution of 1920 × 1200 pixels and a pixel size of 0.27 mm. Digital zoom-in of ×1.5 was used to facilitate the observation. Under these conditions spermatozoa were displayed on the screen with a total magnification of 6300 as determined by the following equation:

$$\begin{array}{c}{\mathrm{M}}_{\mathrm{Dis}}{=\mathrm{M}}_{\mathrm{ob}}\times {\mathrm{M}}_{\mathrm{adapter}}\times \mathrm{Pixel}\mathrm{ratio}\times \mathrm{Digital}\mathrm{zoom}\hfill \\ =\mathrm{100}\times \mathrm{0}\mathrm{.5}\times \left(\mathrm{0}\mathrm{.27}/\mathrm{0}\mathrm{.0032}\right)\times \mathrm{1}\mathrm{.5}=\mathrm{6328}\mathrm{.}\hfill \end{array}$$

Morphological observation was done under ultra-magnification ×6300 independently by two different skilled observers who assessed each spermatozoon image separately with regard to the number of vacuoles and the size of each vacuole. Each image was also graded according to previous vacuole classification (Vanderzwalmen et al., 2008).

The vacuole morphological assessment was also performed according to our novel vacuole criteria: location in the cell and texture (depth). The location of vacuoles was classified as: nuclear, equatorial segment and acrosomal. The texture of vacuoles was classified as ‘deep’ if an internal vacuolar shadow was evident. Observers thereafter discussed assessments until full agreement was reached.

Sperm management

The entire sample of analyzed spermatozoa was saved as a collection of images with a unique serial number for each spermatozoon in the andrology laboratory prior to the genomic analysis. Spermatozoa were transferred from the hospital’s andrology laboratory to the genetic laboratory with addition of separation medium in PCR tubes covered with paraffin paper.

DNA extraction, fragmentation and amplification

Whole genome amplification (WGA) was performed directly on the PCR tubes with the separation medium to avoid possible loss of biological materials. The genomic analysis was performed on a spermatozoon with possible abnormal and normal morphological characteristics utilizing a blind experimental design (namely, the genetic laboratory staff were unaware of the morphological classification of each spermatozoon, previously determined in the andrology laboratory).

DNA extraction, fragmentation and amplification on each spermatozoon cell was conducted with the ‘Genomeplex single cell WGA kit’ (Sigma-Aldrich, St. Louis, MO) according to the instruction manual without any modifications. In brief, a single cell WGA process is divided into lysis and fragmentation (with a lysis specific buffer, Proteinase K and heating procedure), OmniPlex library generation (with a single cell library preparation buffer and PCR as follows: 16°C for 20 min, 24°C for 20 min, 37°C for 20 min, 75°C for 5 min and finally hold at 4°C) and PCR amplification (with amplification master mix and PCR as follows: initial denaturation at 95°C for 3 min, then 25 cycles of denaturation at 94°C for 30 s and annealing/extension at 65°C for 5 min, and finally a hold at 4°C). The quality of the WGA DNA was determined by loading 5–10% (4–8 μL) of the final reaction volume onto a 1.5% agarose gel. The DNA fragments size ranged between 300 and 1000 bp.

CGH array and analysis

DNA from each preparation was measured for quantity and quality with Nano-Drop 1000 spectrophotometer (Thermo Fisher Scientific, Dreieich, Germany). A 60 ng quantity from each sample were used in the aCGH protocol. The aCGH was performed according to Affymetrix® CytoScan™ Assay protocol with no deviations. Briefly, this protocol includes a WGA followed by digestion of the amplified DNA, ligation of an adaptor to the digested DNA, amplification by PCR, fragmentation, TdT labeling, hybridization to Affymetrix gene chip apparatus, and running on LD CytoScan array which contains 750 kb marks. In contrast to common protocols in the field (Bejjani and Shaffer, 2006) and to previous reports of aCGH applied to spermatozoon (Hornak et al., 2014; Garolla et al., 2015), in our protocol no diploid DNA was added to the spermatozoon haploid DNA. The Affymetrix Chromosome Analysis Suite (GRCh 37/hg19) Version cytoB-N1 was used to analyze the results. In addition, results were analyzed using Chas Browser (NetAFFX32 (hg19)) with the following filters: deletions/insertions over 50 kb and at least 25 markers, at known cytogenetic regions and deletions/insertions over 400 kb and at least 50 markers at other genomic regions.

Statistical analysis

All results are expressed as the average value + SD, and categorical variables are expressed as a percentage. Comparison between groups was performed by Student’s t test. P-values of <0.05 were considered to be statistically significant. Data analysis was performed using Microsoft Excel (2010).

Ethical approval

The participants were volunteers and each signed an informed consent. The study was approved by the Assuta Medical Center Institutional Review Board (IRB) ethical committee, approval number ASMC-0082-17.

Results

Spermatozoon isolation and vacuole morphological analysis

Single spermatozoa (n = 53) from 16 infertile patients who underwent an ICSI procedure (n = 31) and from 14 fertile donors (n = 22) were collected, isolated and morphologically characterized using high magnitude microscopic resolution software as described in the Materials and Methods. Each spermatozoon was graded according to previous vacuole classification (Vanderzwalmen et al., 2008), and according to specific vacuole parameters: the exact number, location and morphology (i.e. size and depth). Schematic illustration and exemplary high magnification pictures of vacuole locations and depth morphology are provided in Fig. 1.

Figure 1

Spermatozoa’ vacuole location and morphology (depth). Upper panel: A schematic illustration of single spermatozoon and its major compartments. Lower panel: Sample images of spermatozoa exemplary of diverse vacuole characteristics and genomic stability (A–D). For each image, description of spermatozoon’s vacuole size, number and texture (indicated as ‘deep’, if internal vacuolar shadow was evident) is given, alongside indication of suspicion of nuclear damage and total CNV. Major compartments are indicated by the symbols given in the upper panel. For each exemplar spermatozoon, an ultra-magnification (×6300) image is provided, as well as description of the vacuole characteristic, indication of nuclear damage and total CNV.

DNA content of spermatozoon

To evaluate the stability of the genome of each morphologically normal spermatozoon, a novel protocol of aCGH was applied. The protocol includes WGA, which was performed directly on each single spermatozoon, as the first step, followed by CGH array to analyze total CNV. All spermatozoa, regardless of morphological classification, underwent proper WGA. Detailed information regarding the morphological classification, grading, vacuoles locations and total CNV of each spermatozoon are provided in Table I.

Table I

Sperm vacuoles (number, compartment, size and classification) and suspicion of nuclear damage.

Sperm source	Total CNV	Vacuoles: Number, compartment, size, suspicion of nuclear damage and classification
		No.	Acrosome	Nucleus	Equatorial segment	Classificationa	Suspected of nuclear damage
Fertile donor	171	Single	1 × Small			2	No
Fertile donor	173	–				1	No
Fertile donor	177	–				1	No
Fertile donor	187	Single	1 × Small			2	No
Fertile donor	202	–				1	No
Fertile donor	204	–				1	No
Fertile donor	215	Single	1 × Small			2	No
Fertile donor	242	–				1	No
Fertile donor	243	Single	1 × Small			2	No
Fertile donor	260	Single	1 × Small			2	No
Fertile donor	235	Single	1 × Small and deep			2	Yes: depth
Fertile donor	252	Three	1 × Large and deep + 2 × Small			3	Yes: depth
Fertile donor	288	Single	1 × Small and deep			2	Yes: depth
Fertile donor	304	Single	1 × Large and deep			3	Yes: depth
Fertile donor	274	Single			1 × Small	2	Yes: location
Fertile donor	297	Two			2 × Small	2	Yes: location
Fertile donor	300	Multiple	Multiple, small		1 × Large	3	Yes: location
Fertile donor	304	Single			1 × Small	2	Yes: location
Fertile donor	372	Multiple	2 × Small	1 × Small	1 × Small	3	Yes: location
Fertile donor	496	Single			1 × Small	2	Yes: location
Fertile donor	409	Three	2 × Small		1 × Small and deep	3	Yes: location and depth
Fertile donor	415	Three			1 × Large and deep + 2 × Small	3	Yes: location and depth
Infertile patient	151	Single	1 × Very large			3	No
Infertile patient	176	–				1	No
Infertile patient	200	–				1	No
Infertile patient	212	–				1	No
Infertile patient	220	Multiple	Multiple, small			3	No
Infertile patient	234	–				1	No
Infertile patient	237	Multiple	Multiple, small			3	No
Infertile patient	244	Single	1 × Small			2	No
Infertile patient	245	Single	1 × Small			2	No
Infertile patient	248	–				1	No
Infertile patient	283	Single	1 × Small			2	No
Infertile patient	304	Multiple	5 × Small, 3 are deep			3	Yes: depth
Infertile patient	310	Single	1 × Small and deep			2	Yes: depth
Infertile patient	321	Single	1 × Large and deep			3	Yes: depth
Infertile patient	396	Two	2 × Large and deep			3	Yes: depth
Infertile patient	440	Single	1 × Small and deep			2	Yes: depth
Infertile patient	273	Single			1 × Small	2	Yes: location
Infertile patient	285	Two			2 × Small	2	Yes: location
Infertile patient	294	Two		1 × Small	1 × Small	2	Yes: location
Infertile patient	391	Multiple	3 × Small	Multiple, small	1 × Large	3	Yes: location
Infertile patient	421	Multiple	Multiple, large		Multiple, large	3	Yes: location
Infertile patient	429	Single			1 × Large	3	Yes: location
Infertile patient	431	Two	1 × Small	1 × Intermediate		2	Yes: location
Infertile patient	493	Multiple		1 × Small	4 × Small	3	Yes: location
Infertile patient	771	Multiple	2 × Small	1 × Small	1 × Small	3	Yes: location
Infertile patient	335	Three	2 × Small and deep		1 × Small	3	Yes: location and depth
Infertile patient	376	Three	1 × very Large		2 × Small and deep	3	Yes: location and depth
Infertile patient	460	Two		1 × Small and deep	1 × Small	2	Yes: location and depth
Infertile patient	460	Single			1 × Large and deep	3	Yes: location and depth
Infertile patient	528	Multiple	Multiple, small	3 × Small and deep		3	Yes: location and depth
Infertile patient	744	Single		1 × Small and deep		2	Yes: location and depth

^aBased on Vanderzwalmen et al. (2008), and in agreement with Garolla et al. (2015).

A wide range of total CNVs was observed among spermatozoa from both infertile patients and fertile donors (Table I). We investigated which morphological characteristics might explain this heterogeneity and distinguish spermatozoa with high CNVs from those with low CNVs.

Association between sperm vacuoles and genomic stability, and potential relevance of vacuole size and number

We used a widely accepted classification of spermatozoa based on vacuole size and number (Vanderzwalmen et al., 2008). Indeed, in each of the two sperm source groups (infertile patients and fertile donors), the higher grade corresponded to a clear trend showing higher average total CNVs. In the infertile patients group, average CNV values in spermatozoa with no vacuoles were 214 ± 28, through to 364 ± 149 in spermatozoa with ≤2 small vacuoles, to 389 ± 148 in spermatozoa with ≥1 large vacuole. The corresponding values in the fertile donors group were: 200 ± 28, 270 ± 87 and 342 ± 66, respectively. Moreover, within each of the two sperm source groups, the sperm with vacuoles (grades II and III) both had significantly higher average CNVs compared to spermatozoa with no vacuoles (P < 0.01 and P < 0.001 in the infertile patients group; P < 0.03 and P < 0.003 in the fertile donors group, respectively). Nevertheless, within each of the two sperm source groups, the average CNVs of spermatozoa graded II or III was not significantly different from each other (P = 0.683 and P = 0.078, in the infertile patients and fertile donors’ groups, respectively). Taken together, the results highlight the importance of vacuoles to sperm genomic stability, yet question the association between vacuole number and size and sperm genomic stability.

Association between sperm vacuole morphology and location and genomic stability

We postulated that other spermatozoa vacuolar characteristics might be related to genomic stability, prompting us to further examine the morphology and location of the vacuoles to determine if the vacuoles were deep and/or at either the nucleus or equatorial segment. When vacuoles fulfilled either of these two morphological criteria of specific location or depth, the spermatozoon was marked as suspected of nuclear damage.

Within the infertile patients group, the average total CNVs in spermatozoa with no suspicion of nuclear damage was 223 ± 37 whereas that of spermatozoa suspected of nuclear damage was 423 ± 136 (Fig. 2). Similarly, within the fertile donors group, the average total CNVs in spermatozoa with no suspicion of nuclear damage was 207 ± 32 whereas that of spermatozoa suspected of nuclear damage was 329 ± 78 (Fig. 2). Within each spermatozoa source group, both infertile patients and fertile donors, the average total CNVs was significantly higher (P < 0.0001 and P < 0.0002, respectively). Of note, even if omitting the results of the two spermatozoa with more than 700 CNVs (both suspected of nuclear damage, from infertile patients), a statistical significance would still be evident within the infertile patients group, between spermatozoa with (average total CNV of 386 ± 77) and without suspicion of nuclear damage (P < 0.0001). Thus, the association between sperm vacuole location and depth and genomic stability was demonstrated, regardless of sperm source, i.e. in both infertile patients and fertile donors.

Figure 2

Total CNV of single spermatozoa from fertile donors and infertile patients. Total CNVs of spermatozoa from fertile donors (rhombus) and infertile patients (circles), stratified to these spermatozoa suspected (white) or not suspicious (black) of nuclear damage. Average CNVs for each of the subgroups is given in white and black boxes, respectively. A putative threshold of ~265 total CNVs is denoted by a dashed line.

To further investigate the association of these two parameters with spermatozoa’ genomic stability, spermatozoa were analyzed based on both the common classification (criteria: vacuole size and number) (grades I–III, as proposed by Vanderzwalmen et al., 2008) and the new classification (criteria: location and morphology). In each of the grades, the average total CNVs was significantly higher in spermatozoa suspected of nuclear damage compared to those with no such suspicion. In spermatozoa with ≤2 small vacuoles (grade II), the average total CNVs were: 367 ± 136 and 231 ± 38 (P < 0.003), and in spermatozoa with ≥1 large vacuole (grade III): 404 ± 116 and 203 ± 46 (P < 0.001), respectively. Accordingly, spermatozoa within each grade can thus be separated, based on vacuole size and number, into different cell populations with different levels of genomic stability, based on vacuole location and morphology. Altogether, these findings demonstrate the importance of the vacuolar location and depth to genomic stability, regardless of the number and size of vacuoles.

A putative threshold of ~265 (261–272) total CNVs was deduced to distinguish between spermatozoa with and without suspicion of nuclear damage, regardless of sperm source, based on genomic stability (Fig. 2, dashed horizontal line). The total CNV of almost all spermatozoa without suspicion of nuclear damage was lower than the putative threshold, whereas that of almost all spermatozoa suspected of nuclear damage was above this threshold. Only three spermatozoa (6% of total spermatozoa tested) slightly deviated from the putative threshold of total CNVs: two spermatozoa (from fertile donors) suspected of nuclear damage with 235 and 252 CNVs, only slightly below the putative threshold, and one (from an infertile patient) with 283 CNVs although vacuole morphology and location indicated no suspicion of nuclear damage (Fig. 2). Any putative threshold that differed from 261 to 272 yielded a higher rate of deviation.

Discussion

Sperm genetic anomalies and morphological abnormalities are correlated with decreased fertility (Zini et al., 2002; Moskovtsev et al., 2009; Patassini et al., 2013; Belloc et al., 2014; Garolla et al., 2015; Jenkins et al., 2016; Zidi-Jrah et al., 2016). Yet, the advantage of morphologically based selection of spermatozoa for IMSI according to current criteria is controversial, as contradictory findings concerning its superiority over alternative methods have accumulated (Antinori et al., 2008; Kacem et al., 2010; Mauri et al., 2010; Balaban et al., 2011; Boitrelle et al., 2011; Knez et al., 2011, 2012; Menkveld et al., 2011; Montjean et al., 2012; Klement et al., 2013; Setti et al., 2013; Garolla et al., 2015). Moreover, 15 years after the introduction of MSOME, there is still no consensus regarding MSOME criteria for evaluating normal or abnormal spermatozoon morphology. Recent studies focusing on sperm vacuoles or head morphologies contradicted some of the MSOME veteran assumptions and even suggested against the use of MSOME (De Vos et al., 2013; Gatimel et al., 2014; Fortunato et al., 2016). While more prospective randomized trials are still required to confirm whether IMSI is superior over conventional ICSI, it is widely accepted that selecting the best spermatozoon is an obligatory prerequisite towards a successful fertilization (Setti et al., 2013).

To this end, we aimed to explore whether there is an association between sperm morphology according to current criteria under ultra-magnification and genomic stability, while taking an even closer look at the complexity of vacuolar morphology and location, two characteristics that were previously proposed to be correlated with nuclear damage in the case of large vacuoles (Perdrix and Rives, 2013).

aCGH is an automated method for an effective and comprehensive chromosomal screening of the whole genome. We developed a simple protocol that allows a genetic screen based on aCGH of a single cell (in this case a human spermatozoon) without the need for reference DNA. To the best of our knowledge, this is the first protocol of its kind. Furthermore, the use of aCGH based on affymetrix 750 kb (LD) CGH array enabled us to determine in detail, the genomic stability of each spermatozoon, which could not previously be determined by other CGH arrays such as (50 kb) Blue Genome (Garolla et al., 2015). Genomic analysis of single spermatozoa, which were morphologically characterized by MSOME and particularly by the vacuole morphology and location, enabled us to discover an association between specific spermatozoon vacuole characteristics (deep vacuoles and nuclear or equatorial region vacuoles) and genomic stability (CNVs). These vacuolar characteristics can define the cell as ‘suspicious’ of nuclear damage. We have demonstrated that regardless of sperm source (infertile patients or fertile donors), as all participating spermatozoa passed the MSOME criteria test, spermatozoa that are not suspicious of nuclear damage, according to the new vacuole criteria, exhibit relatively low chromosomal instability (i.e. bearing a significantly lower CNV) while others that passed MSOME morphological criteria and are suspicious of nuclear damage acquire high chromosomal instability (i.e. bearing a significantly higher CNV). A putative threshold of ~265 (261–272) total CNVs was deduced to distinguish between spermatozoa with and without suspicion of nuclear damage, based on genomic stability. Only 6% of the spermatozoa (3 of 53) slightly deviated from this putative threshold: two spermatozoa (from fertile donors) suspicious of nuclear damage with 235 and 252 CNVs, only slightly below the putative threshold and one (from an infertile patient) with 283 CNVs although vacuole morphology and location indicated no suspicion of nuclear damage (Fig. 2). These slight deviations might be explained by additional non-morphological sperm characteristics, such as levels of reactive oxygen species and seminal oxidative stress that affect sperm DNA integrity (Aitken et al., 2010; Dorostghoal et al., 2017; Gosalvez et al., 2017).

Surprisingly, our findings only partially aligned with the previous hypothesis that vacuole size is related to chromosomal instability (Perdrix and Rives, 2013; Vanderzwalmen et al., 2008). There was no significant difference in the average CNVs between spermatozoa of ≤2 small vacuoles and ≥1 large vacuoles (grades II or III, respectively, under previous classification), weakening the link between vacuoles size or number and the spermatozoon genomic stability. The significantly higher average total CNVs in spermatozoa suspicious of nuclear damage compared to those with no such suspicion, albeit within the same grade of previous vacuolar classification, highlights the importance of the vacuoles specific location and depth to the genomic stability of the single spermatozoon (Fig. 2). Implementing these criteria enabled us to differentiate 94% of spermatozoa with higher CNVs from those with lower CNVs, regardless of their source (i.e. from fertile donors or infertile patients).

Examining each spermatozoon separately enabled the identification of subpopulations of single spermatozoa that appear normal according to current criteria and even vacuole grading, yet they are genetically damaged, and vice versa (Table I). These incidences have not been recognized in previous studies aimed to evaluate genetic abnormalities in spermatozoa by using standard vacuole classifications (Garolla et al., 2015; Vanderzwalmen et al., 2008).

In conclusion, this study, with the appropriate cautions due to sample size limitations, has proposed novel criteria to assess vacuole morphology (i.e. depth, location) for selecting the appropriate spermatozoon for IVF–ICSI, which should be applied in addition to previous MSOME criteria. The implications of this emerging morphological classification for future clinical practice will enable a higher level of discrimination and can thus be used in selection of the spermatozoon with the most stable genome for ICSI. Selection of the most suitable and genetically stable spermatozoa, based on MSOME as a base with the additional implementation of the proposed novel vacuole criteria, may further improve ICSI outcomes, regarding the ‘take home healthy baby’ rate. Future studies should explore the possible association of these novel criteria with IVF–ICSI outcomes, in terms of embryo development, implantation rate and pregnancy outcomes.

Acknowledgements

The authors express their gratitude to Ms Dalia Dawn Orkin for her important English language contributions and professional editing services.

Authors’ roles

A.B. and D.B. participated in the study conception and design; Y.M., Y.D., A.B. and D.B. participated in analysis and interpretation of results and in writing and critical discussion of the manuscript. A.B., Y.D. and Y.M. performed the morphological analysis. D.B., R.G. and S.B. performed the genomic experimental work and results analysis. A.B. recruited the patients and collected the samples used for this study. All authors critically reviewed and approved the final manuscript.

Funding

No funding was received for this study.

Conflict of interest

The authors have no conflict of interest to declare.