Confocal scanning laser microscopy of morphometric human sperm parameters: correlation with acrosin profiles and fertilizing capacity*t

Transcription

1 FERTILITY AND STERILITY Vol. 62, No, 2, August The American Fertility Society Printed on acid-free paper in U. S. A. Confocal scanning laser microscopy of morphometric human sperm parameters: correlation with acrosin profiles and fertilizing capacity*t Nikolaos V. Sofikitis, M.D., Ph.D.:j:, Ikuo Miyagawa, M.D., Ph.D.:j: Panayiotis M. Zavos, Ed.S., Ph.D.11 Toshiko Toda, M.D., Ph.DJr Akihiro lino, M.D., Ph.D.** Naoki Terakawa, M.D., Ph.D.' Tottori University School of Medicine, Yonago, Japan, and University of Kentucky, Lexington, Kentucky. Objectives: To develop quantitative criteria for assessing sperm morphology and to determine the correlation between the percentage of morphologically normal spermatozoa and the outcome of the sperm hypo-osmotic swelling test, sperm acrosin profile, and sperm capacity for fertilization. Design: The maximal length and width of the sperm head, the length of the midpiece and principal piece of the sperm tail, and the ratio of the surface of the acrosomal region to the total surface of the head were determined in specimens obtained from a group of infertile men and a group of fertile men using a confocal scanning laser microscope. Group A consisted of 53 infertile men who were participating in an IVF program, and group B consisted of98 fertile men. The mean ± 2 SD of the morphometric parameters in group B was established as representing the lowest and highest normal values in both groups. A normal spermatozoon was defined as one with morphometric parameters within normal levels. The lowest percentage of morphologically normal spermatozoa, hypo-osmotic swelling test result, and acrosin activity in group B were also taken as the lowest normal values in group A. Setting: In vitro fertilization program at the Tottori University School of Medicine, Yonago, Japan. Main Outcome Measures: Sperm morphometric parameters, percentage of morphologically normal spermatozoa, hypo-osmotic swelling test, and acrosin activity. Results: The length of the midpiece, ratio (X1) of the surface of the acrosomal region to the total surface of the sperm head, percentage of morphologically normal spermatozoa, outcome of hypo-osmotic swelling test, and acrosin activity were significantly higher in group B than in group A. The maximal width of the head was significantly lower in group B than in group A. Strongly positive correlations were observed between percentage of morphologically normal spermatozoa or length of midpiece and the proportion of fertilized oocytes in group A and between ratio (xloo) of the surface of the acrosomal region to the total surface of the head and acrosin activity in groups A and B. Sperm morphology showed high positive and negative predictive values for acrosin activity (normal/abnormal) and fertility potential (present/absent). Conclusions: Using quantitative strict criteria, we found that sperm morphology was an important predictor of sperm fertilizing capacity. The confocal scanning laser microscope provided useful information about the sperm cytoskeleton and its importance in fertilization. Fertil Steril 1994;62: Key Words: Sperm morphology, acrosin, confocal scanning laser microscope, in vitro fertilization Received August 2, 1993; revised and accepted March 10, :j: Department of Urology, Tottori University School of Medicine. * Financed by the annual budget allocation of The Department of Urology, Tottori University School of Medicine, Yonago, Japan, provided by The Japanese Ministry of Education, of Urology, Tottori University School of Medicine, 36 Nishima Reprint requests: Ikuo Miyagawa, M.D., Ph.D., Department Tokyo, Japan. chi, Yonago 683, Japan. t Presented in part at the 87th Annual Meeting of The American Urological Association in Washington, D.C., May 10 to 14, 11 Department of Obstetrics and Gynecology, Tottori Univer II Department of Animal Sciences, University of Kentucky. 1992, and at the 48th Annual Meeting of The American Fertility sity School of Medicine. Society in New Orleans, Louisiana, October 31 to November 5, ** Department of Anatomy, Tottori University SchoolofMedicine Sofikitis et al. New criteria for human sperm morphology

2 The prognostic value of sperm morphology for IVF and the relationship between sperm morphology and spermatozoal function tests (zona-free hamster oocyte sperm penetration test, acrosin assay, hypo-osmotic swelling test, and hemizona attachment assay) are controversial, primarily because there are still no clear, standardized criteria to define a normal spermatozoon. Most studies that have attempted to correlate human sperm morphology with sperm fertilizing capacity and/or spermatozoal function tests have used qualitative, not quantitative, criteria for sperm classification (see Menkveld and Kruger [1] for review). Although under normal circumstances, the morphological phenotype of spermatozoa is genetically determined (2), any stress caused by abnormal physical conditions will be reflected in a reduced percentage of normal forms (1). Therefore, sperm morphology may be regarded as a sensitive parameter for estimating semen quality. We attempted to correlate the morphometric parameters of human spermatozoa determined by the confocal scanning laser microscope with the outcome of the hypo-osmotic swelling test, sperm acrosin activity, and sperm fertilizing capacity. Subjects MATERIALS AND METHODS Fifty-three semen samples were collected from 53 infertile men 24 to 38 years old (group A) attending our IVF clinic. The duration of infertility was >4 years. A female infertility factor in the wives of these men had been ruled out by pelvic ultrasonography, follicular monitoring, daily measurements of urinary estrogen, peripheral serum levels oflh and serum FSH, GnRH challenge test, thyrotropin-releasing hormone challenge test, Rubin's test, sperm antibody detection assays, measurement of BBT, and sometimes by laparoscopy. The mean age of the women who participated in this study was 28.4 years (range, 22 to 36 years). An additional 98 semen samples were collected from 98 fertile men (group B) who had fathered a child within the past 2 years and who did not suffer from any urogenital or systemic diseases. All samples were collected by masturbation after 36 to 48 hours of sexual abstinence and were allowed to liquefy at room temperature. Ten minutes after sperm liquefaction, sperm concentration, semen volume, percentage of motile spermatozoa, and speed of forward progression were assessed using methods recommended by the World Health Organization (3). The percentage of dead cells in each sample was assessed by Trypan blue staining (4). Observation by Confocal Scanning Laser Microscope We previously used the confocal scanning laser microscope to identify glomerular bleeding (5, 6) and to observe biological processes in living specimens, such as the division of metaphase chromosomes in cultured lymphocytes (7). Unlike the conventional light microscope, the confocal scanning laser microscope produces sharp images free of outof-focus artifacts that can be observed on a television monitor and recorded on videotape. The confocal scanning laser microscope was developed to examine semiconductors; it provides an automatic system for the instantaneous measurement of distance. This technique has the capacity to provide three-dimensional images of cells at 1,0 to 20,0X magnification without requiring staining of materials. However, when cells are observed within viscosed material, for instance, spermatozoa in seminal plasma, confocal scanning laser microscope provides high resolution and clear images up to magnification of 8,0. Therefore, it is possible to determine morphometric parameters of undisturbed motile cells. A microdrop of each sample was placed on a slide and covered by a coverglass and observed under a confocal scanning laser microscope (1 LM 01;Lasertec, Yokohama, Japan). Motility of spermatozoa, the maximal length and width of the sperm head, and the lengths of the midpiece and principal piece of the sperm tail were routinely evaluated at magnifications ranging from 6,0 to 8,0. In most of the samples, morphometric parameters were determined at a magnification of 7,0 because the clearest images were obtained at this magnification. Maximal length ofthe sperm head was defined as the distance from the point of connection of the neck region of the head with the midpiece to the top of the acrosomal region. Maximal width of the sperm head was defined as the maximal width of the sperm head at the vertical axis of maximal length ofthe sperm head. The principal piece ofthe tail can easily be differentiated from the end piece by the confocal scanning laser microscope because the end piece lacks a fibrous sheath. The confocal scanning laser microscope was connected to a Macintosh 512k computer (Apple Computers, Inc., Vol. 62, No.2, August 1994 Sofikitis et al. New criteria for human sperm morphology 377

3 Figure 1 Confocal scanning laser microphotographs of spermatozoa. (A), Morphologically normal motile spermatozoon (magnification was X8,6); the midpiece (-). (B, C, D), Morphologically abnormal motile spermatozoa (magnification was X6,3, X7,lOO, and X7,2, respectively). Cupertino, CA) that calculated the surface of the acrosomal region and the total surface of the sperm head. The acrosomal region of the sperm head was. clearly distinguished from the dark postacrosomal region. More than 1 randomly selected spermatozoa (motile and immotile) were analyzed in each sample (Fig. 1). To confirm the accuracy of the measurements provided by the confocal scanning laser microscope, the morphometric parameters of a large number of semen samples were determined using the confocal scanning laser microscope and transmission electron microscope (8). This previous study confirmed that the values of sperm parameters obtained with the confocal scanning laser microscope-computer-assisted system were reliable measurements of the actual anatomic values (8). Hypo-Osmotic Swelling Test The hypo-osmotic solution was prepared as previ0usly described (9). A preparation consisting of Sofikitis et al. New criteria for human sperm morphology ml of each sample mixed with 1.0 ml of the hypoosmotic solution was incubated at 37 C for 1 hour. Then the spermatozoa were examined for typical tail-swelling abnormalities, as described by Jeyendran and co-workers (10). Acrosin Assay The total acrosin activity of spermatozoa was measured as previously described by Tummon and co-workers (11). An aliquot of semen diluted with physiological saline was placed in each well of a membrane-bottom, 96-well microtiter plate. Then, seminal plasma and saline were removed from the spermatozoa. A detergent solution and then a substrate solution (N-a-benzoyl-DL-arginine p-n~tro anilide hydrochloride; 5.5 M) were added to each well, and the preparation was incubated for 90 minutes at room temperature. Optical density (OD) was measured at 405 nm in a spectrophotometer.

4 The total acrosin activity of spermatozoa for each sample was calculated as follows: OD (diluted sperm) X 1012/OD (positive control) X (number of sperm/ml). This measurement was defined as the acrosin activity index and represents a relative value for total acrosin activity per spermatozoa. Assay kits were obtained from OEM Concepts Inc. (Toms River, NJ). The advantage of the present method was the low number of spermatozoa necessary for acrosin measurements. Semen Processing for IVF An aliquot of semen from group A was mixed with two volumes of Biggers, Whitten, and Whittingham (BWW) medium (12) preincubated for 3 hours at 37 C (5% CO2 environment) and supplemented with 30% human serum albumin and 10% patient's serum and centrifuged at 1,0 X g for 20 minutes. The sperm pellet was resuspended in 2 ml of medium and recentrifuged at 250 X g for 10 minutes. The supernatant was removed, and the pellet was layered under 1 ml of medium. The spermatozoa were allowed to swim into the medium at 37 C in an atmosphere of 5% CO2 for 20 to 30 minutes. The top 0.5 ml of medium was recovered. Sperm solutions were incubated under the above conditions for 3 hours and were then used for insemination. Oocytes were inseminated with 0.15 X 10 6 motile spermatozoa per oocyte. Ovarian Stimulation and Oocyte Recovery The wives of men in group A were stimulated with a standard regimen for their initial cycle, as we previously described (13). The regimen consisted of 1 mg/d of clomiphene citrate (Clomid; Shionogi and Co., Ltd., Osaka, Japan) for 5 days, starting on day 3 of the menstrual cycle, and an 1M injection of 150 IV of hmg (Humegon; N.V. Organon, Tokyo, Japan) every other day, beginning on day 5 of the cycle. The dose and interval of hmg were adjusted on days 7 and 9 of the cycle according to the ovarian response. A dose of 5,0 IV of hcg (Mochida; Mochida Pharmaceutical Co., Ltd., Tokyo, Japan) was administered to induce final maturation when the dominant follicle reached 20 mm in three dimensions and the concentration of serum E2 exceeded 2 pg/ml (734.2 pmoljl) per large follicle. Transvaginal ultrasound-guided follicular aspiration for oocyte pick-up was performed 36 hours after the administration of hcg. Large follicles (> 15 mm in mean diameter) were irrigated one or two times with the modified BWW medium if an oocyte-cu- mulus complex did not exist in the follicular fluid. The treatment was discontinued when a single follicle developed, when the serum E2 concentration was <183.5 pg/ml (673.4 pmol/l) on day 9 of the cycle, or when the serum P concentration increased to > 1.0 ng/ml in association with significant LH release, signifying premature luteinization. Collected oocytes were classified morphologically as immature, mature (stage 1), mature (stage 2), or atretic, according to our modified criteria for oocyte maturity (13), originally defined by Ben-Rafael and co-workers (14). The mature (stages 1 and 2) and immature oocytes were preincubated in vitro for 4 to 6 hours or for at least 12 hours, respectively. The wives of men in group A participated in one IVF cycle. Oocytes were stripped of adherent cumulus cells with a fine-bore pipette 16 to 18 hours. after insemination and examined for the presence of pronuclei. Embryos were examined 48 to 60 hours after oocyte retrieval and were transferred to the uterus using a Tomcat catheter (Sherwood Medical, St. Louis, MO). Vp to four embryos were inserted per transfer. Degenerated and one-cell embryos were discarded. Details of the frequency of blood collection and the performance of hormone assays in our IVF clinic have been previously described (13). Statistical Analysis and Definition of Normal Values and Correlations Maximal length of the sperm head, maximal width of the head, length of the midpiece, length of the principal piece of the sperm tail, and ratio (XI) of the surface of the acrosomal region to the total surface of the head were calculated for each sample in groups A and B. The mean ± 2 SD of each morphometric parameter was determined for group B samples. This value range was defined as the lowest and highest normal values for the respective parameter in both groups A and B. A spermatozoon was defined as normal in groups A and B if all the morphometric parameters fell within normal ranges. The percentage of morphologically normal spermatozoa was determined in each sample from groups A and B. The lowest values for the percentage of morphologically normal spermatozoa, the outcome of the hypo-osmotic swelling test, and the total acrosin activity in group B were defined as the lowest normal values for the respective parameters in group A. The proportion of oocytes fertilized in group A was considered to indicate fertility potential. Fertilization of at least one oocyte was consid- Vol. 62, No.2, August 1994 Sofikitis et al. New criteria for human sperm morphology 379

5 Table 1 Standard Semen Parameters, Morphometric Sperm Parameters, Hypo-Osmotic Swelling Test, and Acrosin Index in Infertile and Fertile Men* Outcome of IVF Parameter Sperm concentration (X10 6 /ml) Motile spermatozoa (%) Speed of forward progression (0 to 4) Morphologically normal spermatozoa (%) Maximal length of sperm head (/Lm) Maximal width of sperm head (/Lm) Length of sperm midpiece (/Lm) Length of principle piece of the sperm tail (/Lm) Acrosomal region of sperm head:total surface of sperm head (X1) Semen volume (ml) Hypo-osmotic swelling test Acrosin index * Values are means ± SD. Infertile (n = 53) 23 ± 6t 46 ± 9t 2.1 ± 0.3t 19 ± 8t 5.4 ± ± 0.2t 4.3 ± 0.2t 51 ± ± 0.2t 2.9 ± ± 9t 4 ± 3t Fertile Positive Negative (n = 98) (n = 27) (n = 14) 38 ±8 28 ± 4t 18 ±4 71 ±7 52 ± 6t 37 ±5 2.8 ± ± 0.2t 1.9 ± ±6 26 ± 4t 11 ±4 5.4 ± ± ± ± ±0.lt 3.4 ± ± ± O.lt 4.2 ± ±5 52 ±3 51 ± ± ± 0.2t 49.7 ± ± ± ± ± 5 58 ± 4t 46 ± 6 13 ± 3 5 ± 2t 3±2 t Parameters in fertile men were compared with those in infertile men, and the values in cycles with fertilization were compared with those in cycles without fertilization. P < ered to indicate a normal result with fertility potential. Failure to fertilize any oocyte was considered an abnormal result, indicating absence of fertility potential. We assessed the ability ofthe percentage of morphologically normal spermatozoa to predict the outcome of the hypo-osmotic swelling test (normal/abnormal), the outcome of the acrosin assay (normal/abnormal), and the fertility potential (presence/absence). We determined the correlations between each morphometric parameter, the percentage of morphologically normal spermatozoa, and the remaining semen analysis parameters in groups A and B, and the outcome of the hypo-osmotic swelling test, the outcome of the acrosin assay, and the proportion of fertilized oocytes (fertilized oocytes/inseminated oocytes) using Pearson's correlation coefficients. We also determined correlations between length of the midpiece or length of the principal piece of the tail and the percentage of motile sperm or speed of forward progression (qualitative motility). The reproducibility of measurements of morphometric sperm parameters by confocal scanning laser microscope was tested by analyzing different samples of the same ejaculate. Differences in each morphometric spermatozoal parameter, the standard parameters of the semen analysis, the outcome of the hypo-osmotic swelling test, and the acrosin profiles between groups A and B and between cycles with fertilization and cycles without fertilization within group A were analyzed by parametric tests (Student's t-test) when the data were distributed normally and by nonparametric tests (Wilcoxon's test) when a normal distribution was not present. Values are expressed as means ± SD. A probability of P < 0.05 was considered statistically significant. RESULTS Morphometric Sperm Parameters The length of the midpiece was significantly lower, and the maximal width of the sperm head was significantly higher in group A than in group B (Table 1). There were no significant differences in maximal length of the sperm head or length of the principal piece of the tail between groups A and B. The ratio (X1) of the surface of the acrosomal region to the total surface of the head was significantly higher in group B than in group A. Each morphometric parameter showed an essentially normal distribution. The mean ± 2 SD of each morphometric parameter in group B was used to define the normal range (i.e., 5.0,urn ;?; maximal length of the head;?; 3.3,urn, 4.5,urn ; length of the midpiece ; 4.9,urn, 40,urn ; length of the principal piece of the tail ; 60,urn, and 49.5 ; ratio (X1) of the surface of the acrosomal region to the total surface ofthe head; 51.5). In group A, there were no significant differences in maximal length of the head and length of the principal piece of the tail values 380 Sofikitis et a1. New criteria for human sperm morphology

6 7 A ~ ~---- B ~ o ----:: ( ----;;; o.- IIIIC.ITIY! llCATIY! 4 C :: ;; 51 D o ~~ o 1e8Il"'49.7 PlEITID Figure 2 (A), Individual maximal length of sperm head (MLH) for IVF cycles with and without fertilization; (B), Individual length of sperm midpiece (LMP) for IVF cycles with and without fertilization; (C), Individual maximal width of sperm head (MWH) for IVF cycles with and without fertilization; (D), Individual ratio of acrosomal region: total surface of sperm head (ARHS/TSH) for IVF cycles with and without fertilization. between cycles with fertilization of at least one oocyte and cycles without fertilization (Fig. 2A). However, length of the midpiece was significantly higher (Fig. 2B), and maximal width of the head was significantly lower (Fig. 2C) in cycles with fertilization than in cycles without fertilization. The ratio (XI) of the surface of the acrosomal region to the total surface of the head was significantly higher in cycles with fertilization than in cycles without fertilization (Fig. 2D). Semen Parameters There was no significant difference in semen volume between groups A and B or between cycles with fertilization and those without fertilization within group A. However, sperm concentration, speed of forward progression, percentage of motile spermatozoa, and percentage of morphologically normal spermatozoa were significantly higher in fertile men compared with infertile men and in cycles with fertilization compared with cycles without fertilization (Table 1). The lowest percentage of morphologically normal spermatozoa in group B was 22%; values ~22% were considered normal in groups A and B. The percentage of dead spermatozoa was considered negligible: 2.3% ± 1.1% in group A and 2.7% ± 1.3% in group B. Hypo-Osmotic Swelling Test and Acrosin Assay The results of the hypo-osmotic swelling test and the acrosin assay were significantly higher in fertile men than in infertile men and in cycles with fertilization than in cycles without fertilization. The lowest values for the hypo-osmotic swelling test and the acrosin assay in group B were 61 % and 6, respectively. Values ~ 61 % and 6 for these respective parameters were considered normal in groups A andb. In Vitro Fertilization There were no significant differences in mean age or mean duration of infertility between men and women who demonstrated cycles with fertilization and those who demonstrated cycles without fertilization. Forty-three retrievals were attempted in 43 women. In 10 women, the cycles were canceled. Failure to collect even 1 oocyte occurred in 2 cases, which were excluded from statistical analysis. The Vol. 62, No.2, August 1994 Sofikitis et al. New criteria for human sperm morphology 381

7 Table 2 Correlations Between Semen and Morphometric Parameters With the Outcome of Hypo-Osmotic Swelling Test, the Acrosin Index, and the Percentage of Fertilized Oocytes in Infertile and Fertile Men* Outcome of IVF Infertile Fertile Positive Negative (n = 53) (n = 98) (n = 27) (n = 14) Hypo- Percentage Hypo- Hypo- Percentage Hypoosmotic of osmotic osmotic of osmotic swelling Acrosin fertilized swelling Acrosin swelling Acrosin fertilized swelling Acrosin Parameter test index oocytes test index test index oocytes test index Sperm concentration (X10 6 /ml) Motile spermatozoa (%) 0.34t 0.30t 0.34t 0.31t 0.23t t Speed of forward progression (0 to 4) 0.37t 0.29t OA2t 0.35t 0.25t OA7t Morphologically normal spermatozoa (%) 0.53t 0.74t 0.61t OA2t 0.61t OA8t O.73t 0.64t 0.56t 0.75t Maximal length of sperm head (I'm) t t Maximal width of sperm head (I'm) -0.38t -0.58t -OA3t -0.29t -0.54t t -OA5t -.54t -0.64t Length of sperm midpiece (I'm) t 0.58t Length of principal piece of the sperm tail (I'm) Acrosomal region of sperm head:total surface of sperm head (X1) t 0.34t Semen volume t t t t t 0.38t t * Values are Pearson's correlation coefficients. t Significant correlation (P < 0.05). mean number of oocytes recovered was 5.1 ± 2.2. A total of 213 oocytes were collected and inseminated; fertilization occurred in 102 cases. In the 41 cases in which 1 or more oocytes were obtained, the proportion of inseminated oocytes that became fertilized was 0.47 ± In 27 women, at least 1 oocyte was fertilized. In 14 women, no fertilization occurred. In the 27 cases in which 1 or more oocytes were fertilized, the proportion of inseminated oocytes that became fertilized was 0.71 ± In all cases, cleaving embryos that developed from fertilized - cytes were transferred. Correlations The sperm concentration, semen volume, and length of the principal piece of the tail were not significantly correlated with the outcome of the hypo-osmotic swelling test, the sperm acrosin profile, or the proportion of fertilized oocytes in groups A and B (Table 2). The percentage of motile spermatozoa and speed of forward progression showed significant positive correlations with the hypo-osmotic swelling test, acrosin assay, and the proportion of fertilized oocytes in both fertile and infertile men, although the correlation coefficients were not particularly high (r < 0.50). There were significant positive correlations between the percentage ofmorphologically normal spermatozoa and the hypo-osmotic swelling test, acrosin profile, and proportion of fertilized oocytes in groups A and B; the correlation coefficient for the acrosin assay was relatively high (r = 0.74) in infertile men. Maximal length of the head was positively correlated with the hypoosmotic swelling test, acrosin profile, and percentage of fertilized oocytes in both groups. In groups A and B, length of the midpiece showed a significant positive correlation with the percentage of motile spermatozoa (r = 0.48 and r = 0.54, respectively), speed of forward progression (r = 0.57 and r = 0.64, respectively), proportion of fertilized oocytes and sperm acrosin profiles. The ratio (X1) of the surface of the acrosomal region to the total surface of the head was significantly and positively correlated with the proportion offertilized oocytes and acrosin profile in groups A and B; the correlation coefficient for sperm acrosin profiles was markedly high in group A (r = 0.84). The maximal width of the head was significantly and negatively correlated with the hypo-osmotic swelling test, acrosin profile, 382 Sofikitis et a1. New criteria for human sperm morphology

8 and proportion of fertilized oocytes in groups A and B. In group A, the proportion of fertilized oocytes was correlated most strongly with the percentage of morphologically normal spermatozoa and length of the midpiece. Length of the principal piece of the tail was positively, but not significantly, correlated with the percentage of motile spermatozoa and the speed of forward progression in group A (r = 0.11 and r = 0.19, respectively) and group B (r = 0.13 and r = 0.18, respectively). Predictive Value of the Percentage of Morphologically Normal Spermatozoa for the Hypo-Osmotic Swelling Test Outcome When the percentage of morphologically normal spermatozoa was ~22%, the outcome of the hypoosmotic swelling test was normal (~61%) in 101 men (groups A and B combined) and abnormal ( <61 %) in 18 men. When the percentage of morphologically normal spermatozoa was <22%, the outcome of the hypo-osmotic swelling test was normal in 10 men and abnormal in 22 men. The sensitivity, specificity, positive predictive value, and negative predictive value were 0.55, 0.90, 0.68, and 0.84, respectively. Predictive Value of the Percentage of Morphologically Normal Spermatozoa for Acrosin Assay When the percentage of morphologically normal spermatozoa was ~22%, the acrosin index was normal (~6) in 114 men and abnormal «6) in 5 men. When the percentage of morphologically normal spermatozoa was <22%, the acrosin index was normal in 7 men and abnormal in 25 men. The sensitivity, specificity, positive predictive value, and negative predictive value were 0.83, 0.94, 0.78, and 0.95, respectively. Predictive Value of the Percentage of Morphologically Normal Spermatozoa for Fertility Potential When the percentage of morphologically normal spermatozoa was ~22%, fertilization occurred in 25 of 26 cases. When the percentage of morphologically normal spermatozoa was <22%, fertilization occurred in 2 of 15 cases. The sensitivity, specificity, positive predictive value, and negative predictive value of sperm morphology were 0.92, 0.92, 0.86, and 0.96, respectively. Reasonably consistent results and very low coefficients of variation (CV s) (range, 0. to 0.04) were obtained when different samples of the same ejaculate were tested. DISCUSSION Sperm morphology has been regarded as a descriptive parameter because of the lack of quantitative criteria to define a morphologically normal spermatozoon. Investigations have attempted to identify spermatozoal morphological features useful for establishing a reference standard (15-17), but most ofthese attempts involved qualitative criteria as well as the need for spermatozoal staining. Qualitative criteria are highly variable in terms of subject, observer, and procedure. Because cellular staining changes cellular osmolarity, the morphological features observed may not be present in normal motile spermatozoa. In addition, such staining kills the cell. Thus, the characteristics observed refer to a dead, not a living, cellular population. Most investigators have proposed descriptive criteria to characterize normal spermatozoa that do not represent biological spermatozoal parameters. In addition, most available sperm classifications do not correlate sperm anatomy and function. Another drawback to available systems is that some of the qualitative criteria used to define normal sperm morphology are based on selected spermatozoal populations. For instance, Kruger and co-workers (15) proposed a classification of spermatozoa based on the appearance of selected spermatozoa in postcoital periovulatory mucus. However, the probability that spermatozoa with morphological characteristics different from those found in postcoital periovulatory mucus can fertilize oocytes cannot be excluded. Therefore, in our attempt to define the normal ranges of morphometric parameters, we used a fresh spermatozoal population obtained from fertile men and not selected spermatozoa, such as those bound to zona pellucida (ZP) or those found in postcoital periovulatory mucus. Although only motile spermatozoa are expected to fertilize oocytes, we did not limit our analysis to motile cells because sometimes sudden stroilg forward movements were observed in previously immotile spermatozoa after they had been observed for a few minutes. Our proposed criteria are quantitative, not qualitative. Intraobserver, interobserver, and intraprocedure variations should be very low because spermatozoa were not subjected to procedures such as washing or staining that can affect results before microscopy and because morphometric parameters Vol. 62, No.2, August 1994 Sofikitis et al. New criteria for human sperm morphology 383

9 were calculated automatically by the confocal scanning laser microscope-computer-assisted system itself, which does not require specially trained staff. We found very low CV s for sperm morphometric parameters when we evaluated the reproducibility of the sperm morphology assay, indicating that the sperm morphometric measurements obtained with the confocal scanning laser microscope were highly accurate. Sperm staining was unnecessary, making it possible to observe characteristics of living spermatozoa, both motile and immotile. In contrast to other sperm classification systems (1, 15), our criteria included measurements of biological parameters or their ratios. The abnormal spermatozoa identified by our classification appear to have some abnormal functions that result in a lower capacity for fertilization related to an anatomical abnormality. Therefore, our criteria for normal sperm morphology correlate sperm anatomy, sperm function, and fertilizing potential. The strongly positive and significant correlation found between the ratio (XI) of the surface of the acrosomal region to the total surface of the head and the acrosin index suggests that the larger the region of the spermatozoal head occupied by the acrosomal portion, the higher the amount of sperm acrosin. Thus, spermatozoa with abnormally low ratios of the surface of the acrosomal region to the total surface of the head may have a low acrosin content and, therefore, a low fertilization potential because sperm acrosin content is known to correlate with the proportion of oocytes fertilized in vitro (11). The ratio of the surface of the acrosomal region to the total surface of the head was significantly and positively correlated with fertility potential in the infertile men. Furthermore, this ratio was significantly higher in the fertile men than in the infertile men and in cycles with fertilization than in cycles without fertilization, which suggests that the ratio ofthe surface of the acrosomal region to the total surface of the head is an important morphometric parameter of the fertilization potential of spermatozoa. Length of the midpiece is another biological parameter of sperm that is important in the fertilization process. This suggestion is supported by the significant positive correlations of length of the midpiece with the percentage of motile spermatozoa, speed of forward progression, and the proportion of fertilized oocytes. It is also supported by the significantly higher values of length of the midpiece in the fertile men than in the infertile men and in cycles with fertilization than in cycles without fertilization. This suggests that spermatozoa with an ab- normally short midpiece may be less likely to approach and to fertilize oocytes because the midpiece is the site of energy production of sperm and may be regarded as the "battery" ofthe spermatozoon. The mean length of the midpiece was >4.21Lm (Fig. 2B) in all IVF trials that resulted in fertilization. When the mean length of the midpiece was <4.2 llm, not even one oocyte was fertilized (Fig. 2B). The mean length of the midpiece was >4.2 llm in all men of group B. Further study is needed to determine if a mean length of the midpiece <4.2 llm is associated with poor prognosis for IVF. A length of the midpiece ~4.2 llm was defined as abnormal in our study because it was below the lowest normal value. Although identification ofthe lowest and highest normal values for each morphometric parameter based on the mean ± 2 SD in fertile men was subjective, only a small percentage of the spermatozoa of a given specimen in the fertile group (38% ± 6%) was characterized as normal when a normal spermatozoon was defined as one with all five morphometric parameters within the normal range. Using our quantitative criteria, we found that the percentage of morphologically normal spermatozoa was significantly lower in infertile men than in fertile men and in cycles without fertilization compared with cycles with fertilization. Among semen parameters, sperm morphology was most strongly correlated with the outcome of the hypo-osmotic swelling test and acrosin assay, which are significantly correlated with IVF outcome (11, 18) and with the proportion of oocytes fertilized by infertile men. When the lowest normal percentage of morphologically normal spermatozoa was defined as 22%, sperm morphology showed high positive and negative predictive values for fertility potential. These findings indicate the importance of sperm morphology for estimating the fertility potential. Although some investigators have not found a definite influence of sperm morphology on fertility potential, their results may be due to the qualitative criteria used to evaluate sperm morphology (16, 17). When Kruger and co-workers (15) used strict qualitative criteria to evaluate sperm morphology, they found that the percentage of morphologically normal spermatozoa did have a prognostic value for estimating human fertility potential. The percentage of morphologically normal spermatozoa in both groups A and B showed significant positive correlations with the hypo-osmotic swelling test and the acrosin index. These findings are compatible with previous reports in which qualitative criteria for sperm morphology were used (10, 384 Sofikitis et al. New criteria for human sperm morphology

10 19-21), although the correlation coefficients obtained in our study were higher than those found in previous studies (10, 19). The differences in results may be related to the strict morphometric criteria we used to evaluate sperm morphology. In addition, the percentage of morphologically normal spermatozoa showed high positive and negative predictive values for the hypo-osmotic swelling test and acrosin profiles, suggesting that a morphologically and morphometrically abnormal spermatozoon is associated with an abnormal membrane function and an aberrant packaging of acrosin. Results of the hypo-osmotic swelling test and the acrosin assay were significantly lower in infertile men than in fertile men and in cycles without fertilization compared with cycles with fertilization, confirming previous findings that suggested that abnormal sperm membrane function or low acrosin profiles may contribute to development of infertility (10, 11, 19-21). The significant positive correlations observed between parameters of sperm motility and the hypoosmotic swelling test and acrosin assay are consistent with findings of previous studies (10, 19). The sperm maximal width ofthe head was significantly higher in infertile men than in fertile men and in cycles without fertilization than in cycles with fertilization. In addition, significant negative correlations were noted between the maximal width of the head and fertility potential in infertile men and between maximal width of the head and the hypo-osmotic swelling test and acrosin index in both fertile and infertile men, which suggests that maximal width of the head is another important morphometric parameter of fertilization. Spermatozoa with abnormally high values of maximal width of the head may have defective membrane function and/or abnormal acrosin content, contributing to their lower capacity for fertilization. These spermatozoa may represent an immature sperm population. The confocal scanning laser microscope is capable of characterizing "round cells" and distinguishing immature germ cells, polymorphonuclear leucocytes, and lymphocytes without staining. The present study represents the first application of quantitative criteria for evaluation of sperm morphology without the use of staining. Our criteria were based on confocal scanning laser microscope observation of spermatozoa obtained from fertile men. Although the confocal scanning laser microscope is a relatively expensive instrument, its use in infertility clinics and in research in the field of andrology is definitely recommended because it contributes to a better understanding of the sperm cytoskeleton and its importance in fertilization. The use of confocal scanning laser microscope in infertility clinics may not be limited to evaluation of sperm morphology; it may also be useful for the observation and analysis of sperm movement and the process of sperm penetration into the ZP. REFERENCES 1. Menkveld R, Kruger TF. Laboratory procedures: review and background. In: Acosta AA, Swanson RJ, Ackerman SB, Kruger TF, van Zyl JA, Menkveld R, editors. Human spermatozoa in assisted reproduction. Baltimore: Williams and Wilkins, 1990: Wyrobek AJ, Bruce WR. Induction of sperm-shape abnormalities in mice and humans. In: Hollander A, de Serres FJ, editors. Chemical mutagens. Vol 5. New York: Plenum, 1978: World Health Organization: WHO Laboratory mannual for the examination of human semen and semen-cervical mucus interaction. 2nd ed. Cambridge: The Press Syndicate of the University of Cambridge, 1987; Talbot P, Chacon R. A triple-stain technique for evaluating normal acrosome reactions of human sperm. J Exp Zool 1981;215: Hyodo T, Ono K, Koumi T, Sofikitis N, Watanabe N, Miyagawa I, et al. Application of the confocal scanning laser microscope for observing biological specimens: detection of glomerular bleeding. Nephron 1991;59: Miyagawa I, Hyodo T, lino A. Detection of glomerular bleeding by the confocal scanning laser microscope. Aktuel Urol 1993;24(1 Suppl): Iino A, Inaga S. Application of color laser microscope for observing living biological specimens. Cytobios 1989;60: Sofikitis N, Zavos PM, Toda T, Hirakawa S, Miyagawa I, Mio Y. Morphometric parameters observed by confocal scanning laser microscopy versus transmission electron microscopy. J Androl 1993;14 Suppl 1:P Sofikitis N, Miyagawa I, Toda T, Harada T, Mio Y, Terakawa N. The effect of kallikrein on human sperm membrane function. Tohoku J Exp Med 1991;164: Jeyendran RS, Van der Ven HH, Perez-Pelaez M, Crabo BG, Zaneveld LJD. Development of an assay to assess the functional integrity of the human sperm membrane and its relationship to other semen characteristics. J Reprod Fertil 1984;70: Tummon IS, Yuzpe AA, Daniel SAJ, Deutsch A. Total acrosin activity correlates with fertility potential after fertilization in vitro. Fertil Steril 1991;56: Biggers JD, Whitten WK, Whittingham DG. The culture of mouse embryos in vitro. In: Daniel DJ, editor. Methods in mammalian embryology. San Francisco: Freeman, 1971: Mio Y, Sekijima A, Iwabe T, Onohara Y, Harada T, Terakawa N. Subtle rise in serum progesterone during the follicular phase as a predictor of the outcome of in vitro fertilization. Fertil SteriI1992;58: Ben-Rafael Z, Kopf GS, Blasco L, Flickinger GL, Tureck Vol. 62, No.2, August 1994 Sofikitis et al. New criteria for human sperm morphology 385

11 RW, Strauss JF, et al. Follicular maturation parameters associated with the failure of oocyte retrieval, fertilization, and cleavage in vitro. Fertil Steril1986;45: Kruger TF, Acosta AA, Simmons KF, Swanson RJ, Matta JF, Oehninger S. Predictive value of abnormal sperm morphology in in vitro fertilization. Fertil Steril1988;49: Alper MM, Lee GS, Seibel MM, Smith D, Oskowaity SP, Rausil BJ, et al. The relationship of semen parameters to fertilization in patients participating in a program of in vitro fertilization. J in Vitro Fert Embryo Transf 1985;2: Tarlatzis BC, De Cherney AH, Amster S, Laufer N, Graebe RA, Boyers S, et al. Semen characteristics and in-vitro fertilization of human oocytes. J Androl 1985;6:N Van der Ven HH, Jeyendran RS, Al-Hasani S, Perez-Pelaez M, Diedrich K, Zaneveld LJD. Correlation between human sperm swelling in hypoosmotic medium (hypoosmotic swelling test) and in vitro fertilization. J Androl 1986;7: Goodpasture JC, Zavos PM, Zaneveld LJD. Relationship of human sperm acrosin and proacrosin to semen parameters. II. Correlations. J Androl1987;8: Goodpasture JC, Zavos PM, Cohen MR, Zaneveld LJD. Relationship of human sperm acrosin and proacrosin to semen parameters. I. Comparisons between symptomatic men of infertile couples and asymptomatic men, and between different split ejaculate fractions. J Androl1982;3: Zaneveld LJD, Jeyendran RS, Krajeski P, Coetzee K, Kruger TF, Lombard CJ. Hypoosmotic swelling test. In: Acosta AA, Swanson RJ, Ackerman SB, Kruger TF, van Zyl JA, Menkveld R, editors. Human spermatozoa in assisted reproduction. Baltimore: Williams and Wilkins, 1990: Sofikitis et ai. New criteria for human sperm morphology