Tygerberg Hospital and University of Stellenbosch, Tygerberg, South Africa, and Free University of Amsterdam, Amsterdam, The Netherlands

Transcription

1 FERTILITY AND STERILITY Copyright 1996 American Society for Reproductive Medicine Printed on acid-free paper in U. s. A. Acrosomal morphology as a novel criterion for male fertility diagnosis: relation with acrosin activity, morphology (strict criteria), and fertilization in vitro* Roelof Menkveld, Ph,D. t+ Johan P. T. Rhemrev, M.D. II Daniel R. Franken, Ph.D. t Jan P. W. Vermeiden, Ph.D.11 Thinus F. Kruger, M.D.t Tygerberg Hospital and University of Stellenbosch, Tygerberg, South Africa, and Free University of Amsterdam, Amsterdam, The Netherlands Objective: To determine the relationships between sperm acrosin activity, sperm morphology evaluated according to strict criteria, visually observed acrosomal morphology, and IVF rates. Design: Prospective analytic study. Acrosin activity was determined on all semen samples together with a standard semen analysis. Emphasis was placed on sperm morphology and especially a novel criterion viz acrosome morphology (acrosome index) as recorded with bright field microscopy. Setting: University-based tertiary care center. Patients: Thirty-three couples undergoing IVF or GIFT with two or more metaphase II ova inseminated in vitro. Main Outcome Measure: In vitro fertilization rates of inseminated ova. Results: Strong correlations were found between acrosome index, normal sperm morphology, and IVF rates. An acrosome index cutoff value could be established at > 10% normal acrosomes for IVF rates of 2:50% (sensitivity and specificity = 100%) and an acrosin activity cutoff value at > 18 JLIU/10 6 sperm. A multiple linear regression analysis showed that the acrosome index and acrosin activity added a significant contribution to the explanation of the variation in the fertilization rates. Conclusions: A strong positive correlation was found between acrosome index and IVF rates. Although the numbers of the study are small, the results indicate that the acrosome index possibly may be regarded as an additional tool in the prediction of IVF outcome and especially may be of value in the group of men with severe teratozoospermia, i.e., :=;4% morphologically normal spermatozoa. Fertil Steril 1996; 65: Key Words: Sperm morphology, strict criteria, acrosome index, acrosin activity, IVF rates Development of assisted reproductive techniques has focused the attention of investigators on the Received April 17, 1995; revised and accepted August 30, * Preliminary results presented at the 10th Annual Meeting of the European Society of Human Reproduction and Embryology (ESHRE), Brussels, Belgium, June 26 to 29, t Reproductive Biology Unit, Department of Obstetrics and Gynaecology, Tygerberg Hospital and University of Stellenbosch. :j: Supported by the South African Medical Research Council, Tygerberg, South Africa. Reprint requests: RoelofMenkveld, Ph.D., Andrology Laboratory E3, Tygerberg Hospital, 7505 Tygerberg, South Africa. (FAX: ). II In Vitro Fertilization Center, Department of Obstetrics and Gynaecology, Academic Hospital, Free University of Amsterdam. functional potential of spermatozoa in comparison with the results obtained from standard semen analyses in the evaluation of male fertility potential. This partly was due to the onset of IVF, as this caused an enhancement in the reports on gamete interaction. The search for functional sperm tests led to the establishment of new tests or the reinvestigation of existing tests, e.g., the determination of sperm acrosin activity (1, 2). Recent literature has focused on the relationship between acrosin activity and IVF. In this regard the role of acrosin is not defined clearly because contradictory results have been reported (2-7). Acrosin is an unique neutral proteinase enzyme found mostly in the acrosomal region of the sperm Menkveld et ai. Acrosome morphology and NF 637

2 head in an inactive zymogen form called proacrosin. With release from the acrosome, proacrosin is converted to acrosin (1). Therefore, it may be expected that acrosin activity may be dependent on the percentage of normal intact acrosomes in a specific semen sample. In fact, low or no acrosin activity has been observed in the presence of small or the absence of acrosomes (1, 8). A variety of techniques have been used to study acrosome status, e.g., triple stain (9), immunofluorescence (2, 10, 11), and transmission electron microscopy (1, 8). However, information on light microscopy of acrosome morphology (12) received little attention. With the strict (Tygerberg) criteria (13), the size of the acrosome plays an important role in defining the ideal normal spermatozoon. Although slightly abnormal, as well as slightly and moderately elongated, sperm may have normal acrosomes, these cells are recorded as abnormal, according to strict criteria. These cells, however, still may contribute to the total acrosin activity and thus playa role in the fertility pathway (14). Therefore, the purpose of this study was to investigate the relationship between acrosin activity and [1] a novel criterion, namely, acrosomal morphology, expressed as an acrosome index; [2] percentage morphologically normal spermatozoa evaluated according to strict criteria; and [3] IVF rates. MATERIALS AND METHODS Semen samples were obtained from 33 husbands of a group of un selected couples attending the Tygerberg Hospital IVF program. Samples were collected on the day of follicle aspiration or, if this was not possible, within a few days of oocyte retrieval. An aliquot of each semen sample was used for a standard semen analysis and acrosin determination. Evaluation of sperm morphology received special attention and was extended to include a novel characteristic viz acrosomal morphology, expressed as the acrosome index. In Vitro Fertilization Protocol Routine protocols for ovulation induction were used, with administration of clomiphene citrate and hmg, as well as standard laboratory procedures as described previously (15). All semen samples for IVF were prepared using standard wash and swim-up techniques. Sperm concentrations of recovered motile fractions were adjusted to 100,000 or 500,000 in cases of severe teratozoospermia (,,;4% normal forms). Follicle aspirates were placed into a Falcon 3003 petri dish (Falcon Plastics, Oxnard, CA) to disperse cumulus cells, thus allowing oocytes to become 638 Menkveld et a1. Acrosome morphology and IVF clearly visible. Maturity ofthe oocytes was evaluated using a dissecting microscope (20x eyepiece and 8x objective). All oocytes were graded according to maturity as metaphase II, metaphase I, or prophase I. Three selected (metaphase II) oocytes were transferred after follicle aspiration during the GIFT procedure. All excess oocytes from the GIFT procedure and all oocytes from IVF procedures were inseminated 4 to 6 hours after aspiration. Oocytes were evaluated for fertilization rate 18 hours after insemination. Only patients, both IVF and GIFT, in whom two or more ova were inseminated were included in this study. Embryo formation and development were assessed after 45 hours of incubation. All normally developed not transferred embryos were cryopreserved using standard procedures. Semen Analysis Standard semen analyses were performed on an aliquot of each semen sample as described previously (16). A small drop of the thoroughly mixed specimen was placed on a glass slide, covered with a coverslip, and left for a few minutes at room temperature to settle. Slides were examined under phase-contrast microscopy with the aid of 16x and 40x objectives. Percentage of motile spermatozoa and speed of forward progression (on a scale from 0 to 4) was assessed on 2: 10 randomly selected highpower fields. Depending on an estimation of the sperm concentration, a 1:10, 1:20, or 1:100 dilution was prepared with the use of a glass tuberculin syringe (instead of the white blood cell pipette). Sperms were counted with the aid of an improved double-ruled Neubauer hemocytometer. Morphological Evaluation of Spermatozoa For the evaluation of sperm morphology, thin smears were prepared, air dried, and stained with a modified Papanicolaou staining method (17). Sperm morphology was evaluated according to strict (Tygerberg) criteria (13). To be classified as normal, according to strict criteria, the sperm head must have a smooth oval configuration, measuring between 3 to 5 flm in length and 2 to 3 flm in width. The acrosome must be clearly visible and well defined, comprising approximately 40% to 70% of the sperm head. Borderline normal head forms or spermatozoa with nearly oval heads and no other abnormalities are classified as abnormal. Cytoplasmic droplets (remnants) less than half the size of the sperm head may be present. Evaluations were performed with bright field illumination and 100x oil immersion objectives with a total magnification of 1,250x. For each semen sample, at least 100 but preferably 200 spermatozoa were evaluated. Spermatozoa were Fertility and Sterility 1 j

3 classified into one of seven classes as described previously (13). Acrosome Evaluation (Acrosome Index) Mter the initial morphology evaluation, an additional 100 spermatozoa were evaluated on the same slide and at magnification of 1,250x. Acrosomal morphology was assessed with regard to size, form, and detailed characteristics of Papanicolaou staining quality. These recordings did not reflect the acrosomal reacted status ofthe spermatozoa. The acrosome index, therefore, included the following four classes: normal, too small, amorphous, and staining defects. Acrosmal Evaluation Criteria Normal Acrosomes. The same criteria for a normal sperm head, regarding the acrosome, namely clearly visible and well defined with a smooth oval configuration and comprising 40% to 70% of the sperm head was used to classify the acrosome as normal. Normal acrosomal staining had to be a homogeneous light blue. Staining Defects. Irregular staining, the presence of three or more vacuoles or large cysts was considered as a staining defect. Furthermore, the equatorial junction had to form a regular clear line across the sperm head. Too Small. In cases where the acrosome covered <40% of the anterior part of a normal head it was considered as too small. Acrosomes of spermatozoa classified as too small or from spermatozoa with severe elongation, usually associated with small acrosomes, also were considered as too small. Amorphous. All other abnormalities, including acrosomes that were too large, were classified under amorphous. An acrosomal index was calculated using the following formula: acrosomal index (%) spermatozoa with normal acrosomes X 100 total number of spermatozoa evaluated A population of spermatozoa with mostly morphologically normal acrosomes is illustrated in Figure 1. Reproducibility of Acrosomal Evaluations To evaluate the between-observer reproducibility of the acrosome index, 15 semen smears for morphology evaluations were evaluated blindly by two observers. To determine the within-observer reproducibility of the acrosome index, the same 15 smears were re-evaluated blindly by one observer (R.M.). Three smears were re-evaluated a total of 10 times Figure 1 A population of spermatozoa with mainly morphologically normal acrosomes. Note oval forms and distinct acrosomes with homogeneous staining. Abn, abnormal acrosomes. each to determine the correlation coefficient of the method. Acrosin Assay Acrosin activity was determined using the wellestablished method of Kennedy et al. (4). An aliquot of semen (::::;250 JLL) containing 2 to 10 X 10 6 spermatozoa was placed on top of 500 JLL of an 11% Ficoll solution in a 15-mL test tube. One control and three tests per patient were used. The assay tubes were centrifuged for 30 minutes at 1,000 X g. The supernatant was drawn off, leaving the pellet and ±0.1 ml solution. A volume of 100 JLL 500 mm benzamidine solution was added immediately to the controls. A volume of 1 ml 23 mm N-a-benzoyl-DL-argininep-nitroanilide (BAPNA; Sigma Chemical Co., St. Louis, MO) in 10% dimethylsulfoxide/0.1 % Triton X- 100 solution then was added to controls and tests. The tubes were left for 3 hours at room temperature with periodic vortexing to mix sperm and assay solution. Mter 3 hours, 100 JLL 500 mm benzamidine solution was added to all tests. The controls and tests tubes were centrifuged for 10 minutes at 1,000 X g. The supernatants were poured in 1 ml cuvettes and read at an optical density (OD) of 410 nm. The acrosin activity was calculated as follows: acrosin activity (JLIU/10 6 sperm) [(mean OD(test) - OD(control)] X ,851 X number of sperm cells on Ficoll in millions Reproducibility of Acrosin Assay To monitor the reproducibility ofthe assay, duplicate assays were performed on 11 semen samples, using two different semen volumes to overlay the Menkveld et al. Acrosome morphology and NF 639

4 Table 1 Semen Parameters Values*, Acrosomal Index, and Acrosin Activity of the 33 patients Included in This Investigation Fertilization rate No. of patients Volume (ml) Motility (% motile) Forward progression (0 to 4) Sperm count (X10 6 /ml):j: Morphology (% normal):j: Acrosome index (% normal acrosomes) Acrosin activity Fertilization Ova fertilized Ova inseminated Fertilization rate (%) <50% ± ± ± ± 60.6 (3.16 to ) 2.3 ± 2.0 (1 to 7) 4.3 ± ± ± ± ± :50% Mean P valuet ± ± 1.4 NS 49.4 ± ± 13.9 NS 2.58 ± ± 0.38 NS ± ± 60.6 NS (26.88 to ) 11.8 ± ± 6.8 < (1 to 22) 25.9 ± ± 13.0 < ± ± 15.2 NS 5.4 ± ± 3.8 < ± ± 4.6 NS 82.9 ± ± 41.1 < * Values are means ± SD. t Between <50% and 2:50% fertilization groups as calculated by Mann-Whitney U-test. NS, not significant. :j: Values in parentheses are ranges. Ficoll solutions, resulting in two different amounts of recovered spermatozoa. A fresh donor semen sample was included in each assay bach to monitor the variability of the assay over time. Statistical Analysis Results were calculated and presented by descriptive statistics (means ± SD). Correlations were evaluated using linear regression analysis between acrosin activity, acrosome index, and percentage morphologically normal spermatozoa and IVF rates of metaphase II oocytes. To obtain a cutoff value for acrosin activity, normal morphology, and acrosome index a receiver operating characteristic (ROC) curve analysis (MedCalc; MedCalc Software, Mariakerke, Belgium) was performed with fertilization rate cutoff values of ;:::50% (normal) and <50% (low), based on previous results as published by Franken et al. (18). Moreover, to evaluate the contribution of the individual parameters to the explanation of the variation in the fertilization rates, a multiple linear regression analysis was performed. The parameters tested were male age, female age, cause of infertility, volume, motility, forward progression, total count, morphology, acrosome index, and acrosin activity. RESULTS Results of semen analyses and the other variables viz acrosin activity and acrosome index of the 33 patients, divided into two groups according to fertilization rates of <50% or ;:::50%, are represented in Table 1. The fertilization rate for the <50% group was 8.1% ± 13.9% (mean ± SD) and for the ;:::50% group was 82.9% ± 18.3% (P < ). Based on 640 Menkveld et al. Acrosome morphology and NF actual numbers, 15 of 119 ova (12.6%) were fertilized in the <50% fertilization group and 87 of 111 ova (78.4%) were fertilized in the ;:::50% group with an overall average fertilization rate of 44.3%, i.e., 102 of 230 ova. No statistically significant differences were found for sperm concentration, motility, and forward progression between the <50% and ;:::50% fertilization rate groups. The ranges for the sperm counts in the <50% and ;:::50% fertilization groups were 3.16 to X 10 6 /ml and to X 10 6 /ml, respectively. The morphology in the <50% and ;:::50% fertilization groups ranged between 1% to 7% and 1 % to 22% morphologically normal spermatozoa, respectively. Statistically significant differences were found between the two groups for sperm morphology with means of 2.3% ± 2.0% and 11.8% ± 6.7% normal, respectively (P < ), as well as for the acrosome index (percent normal acrosomes) with means of 4.3% ± 3.5% and 25.9% ± 9.6%, respectively (P < ). Regression analysis of the acrosome index scores for the 15 samples evaluated independently by two observers showed a correlation coefficient of r = (P < ). The duplicate evaluations of the same 15 smears by one observer resulted in mean acrosome index scores of 16.6% ± 10.0% and 16.1% ± 9.6%, respectively (P = ) and the regression analysis showed a correlation coefficient of r = (P < ). Coefficients of variation of the three smears, on which the acrosome index was evaluated 10 times each, were 5.1 %, 6.9%, and 10.9%, respectively. No statistically significant difference was found for acrosin activity between the <50% and ;:::50% Fertility and Sterility

5 Table 2 Correlation Coefficients Between Acrosome Index, Acrosin Activity, Sperm Morphology, and IVF Rates Morphology Acrosome index Fertilization rate * P < 0.000l. t P < Acrosin activity * * t Morphology * * Acrosome index * 50 xi 0 ~ 40 0 z'" -0 ~ ~ ~a; ~bo 0<:: 15 <i'!:. 10., ACROSIN ACTIVITY (1'Iu/10 sperms) fertilization groups with acrosin activities of ± 14.0 and ± 15.9 J.LlUlI0 6 sperm, respectively. Acrosin activities ranged between 3.0 and 85.4 j.liui 10 6 sperm with a mean of ± 20.7 j.liu/l0 6 sperm. The 11 duplicate acrosin assays resulted in mean acrosin activities of ± 13.2 and ± 13.0 j.liull0 6 sperm, respectively (P = ) and a correlation coefficient of r = (P < ). Coefficients of variation of 12.8% and 16.4% were obtained for acrosin activities of four and three fresh semen samples of two semen donors analyzed over a period of 6 and 12 weeks, respectively. Correlation coefficients between morphology, acrosome index, acrosin activity, and fertilization rates are represented in Table 2. Positive correlations were found between all four variables with the best and highly significant correlation (r = ; P < ) between acrosome index and morphology as illustrated in Figure 2. A lower correlation (r = ; P < ) was found between acrosome index and fertilization rates. A positive correlation (r = ; P < ) also was found between acrosome index and acrosin activity as illustrated in Figure 3. As depicted in Figure 2, all patients with a fertilization rate of ~50% had an acrosome index of~15% > C) 9 10 o :J: ~ 5 o :::i: ACROSOME INDEX (% normal acrosomes) Figure 2 Correlation between acrosome index and sperm morphology. Correlation coefficient (r) = (P < ) and y = x (P < )..., fertilization rates <50%;, fertilization rates 2> 50%. Figure 3 Correlation between acrosome index and acrosin activity. Correlation coefficient (r) = (P < ) and y = x (P < )..., fertilization rates <50%;., fertilization rates 2> 50%. normal acrosomes. This is in agreement with the cutoff point of > 10.0% calculated by means of the ROC curve analysis (Fig. 4). With an acrosome index of> 10% normal acrosomes, the calculated specificity and sensitivity was 100% for the prediction offertilization rates of ~50%. Also depicted in Figure 2 are three patients with 5;4% morphologically normal spermatozoa, i.e., poor prognosis morphology group, but with an acrosome index of ~ 15% with fertilization rates of ~50% in all these cases. By ROC curve analysis (Fig. 4) the sperm morphology cutoff point for the prediction of a fertilization rate outcome of ~50% was at >4% morphological ~ > E 0) c:: CD ~ 60 z 0 i= u «a: 40 II.. w > i= en 20 0 a.. w ::> a: I- 0 IJ ~# 1(,10) # I., I ~ (.4) # # I... ; # I'" f / FALSE POSITIVE FRACTION (100-Specificity) Figure 4 Receiver operating characteristic (ROC) curves for acrosome index ( ), acrosin activity (- 0 -), and sperm morphology ( ) with respect to ferti~i~~tion outcom.e of 2>50%. With an acrosome index> 10.0%, sensitivity and specificity is 100%, respectively, with an acrosin activity of > 18.0!LIUI 10 6 sperm, sensitivity = 63% and specificity = 76%, and with >4% morphologically normal spermatozoa, the sensitivity = 87% and specificity = 88%. Menkveld et al. Acrosome morphology and NF 641

6 normal spermatozoa, with a sensitivity of 87% and a specificity ofb8%. A cutoff point for acrosin activity was indicated at > 18 /LIU/10 6 sperm, with a positive predictive rate of 62.5%. With an acrosin activity of 2=19.3 /LIU/l0 6 sperm, the fertilization rates were always 2=75%. However, lower acrosin activity did not exclude fertilization rates of 2=50%, as six men with acrosin activities between 6.4 and 10.5 /LIU/10 6 sperm fell in this group. Despite the low correlation (r = ; P < 0.005) between acrosin activity and fertilization rates, a multiple linear regression analysis showed that the acrosome index and acrosin activity significantly contributed to the explanation of variation of the fertilization rates with a coefficient of (P < 0.005) by the acrosome index and with a coefficient of (P < 0.03) by the log acrosin activity. The regression was significant with a multiple R value of 0.86 (P < ). Because the morphology, as evaluated according to the strict criteria, did not improve the explanation offertilization rate variation in the statistical model, it was omitted. The nonsignificant contribution of morphology to the multiple colinearity statistical model could be explained by the high correlation (r = ; P < ) between morphology and acrosome index. A multiple analysis without the acrosome index as one of the variables produced an univariate R for morphology of 0.80, without acrosin activity as a significant contributor to the statistical model. Other parameters, i.e., male age, female age, cause of infertility, volume, motility, total count, forward progression, and morphology did not contribute significantly to the explanation of variation in the fertilization rates. DISCUSSION Patients included in this study represent an unselected population, as will be encountered in any assisted reproduction center. Apparently, normal women as well as women with a wide spectrum of causes for their infertility including anovulation, endometriosis, and tubal factors were included in this study. The men included in the study had mostly normal sperm concentrations of 2= 10 X 10 6 /ml, although a few oligozoospermic men were present in the <50% fertilization rate group. However, no statistical significant differences were found between the <50% and 2=50% fertilization rate groups, for mean sperm concentrations as well as for motility and speed of forward progression. Thus, the main male factor in both the <50% and 2=50% fertilization groups was isolated teratozoospermia «15% morphological normal spermatozoa). However, a statistical significant lower mean for the percentage of nor- 642 Menkveld et al. Acrosome morphology and NF mal sperm morphology was found in the < 50% fertilization group compared with the 2=50% fertilization group. This once again underlines the important and well documented (5, 15, 19) role played by sperm morphology in the prediction and outcome of IVF rates. The important role played by sperm morphology may be due to a functional relationship between sperm morphology and some of the five major events taking place in the fertilization pathway (14). These events are the acrosome reaction with eventual release of acrosin, sperm-zona pellucida binding, sperm-zona pellucida penetration, sperm-oolemma fusion, and sperm decondensation. It already has been documented that sperm morphology is related closely to sperm-zona pellucida binding (18,20,21). The results of this study further link these sperm functional events with sperm morphology in two ways. The first is in the relationship between sperm and acrosomal morphology, expressed as the acrosome index. This is illustrated by the strong positive correlation (r = ; P < ) found between these two variables. This could be expected as, with an increase in the percentage of morphologically normal spermatozoa, there also must be an increase in the number of spermatozoa with normal acrosomes. However, the higher correlation obtained between acrosome index and fertilization rates compared with morphology and fertilization rates may indicate that the presence of morphologically normal acrosomes are of more importance for normal spermzona pellucida interaction than normal sperm morphology alone. This may be due to the fact that, although mainly ideal morphological normal spermatozoa are found tightly bound to the zona pellucida (20, 21), slightly amorphous as well as slightly and moderately elongated spermatozoa, possessing morphologically normal acrosomes, also are found tightly bound to the zona pellucida, although to a lesser extent. Severely abnormal or elongated spermatozoa are not found bound to the zona pellucida. Furthermore, Liu and Baker (21) found that small oval formed and pyriform spermatozoa with acrosomal areas shaped similar to those of morphological normal spermatozoa could bind to the zona pellucida. Furthermore, Heywinkel et al. (22) found that acrosomes of elongated spermatozoa are capable of undergoing the acrosome reaction. This is of importance as more and stronger evidence that only spermatozoa with normal and intact acrosomes, i.e., unreacted, are capable to bind to the zona pellucida now is becoming available (21, 23). The establishment of the actual percentage of morphologically normal acrosomes available in a semen sample therefore will be a more accurate estimate of the Fertility and Sterility

7 number of spermatozoa potentially capable to bind to the zona pellucida compared with normal sperm morphology, and thus also superior in the prediction of the expected IVF rates. The acrosome index, therefore, does not reflect merely another aspect of a measured semen variable(s), in this case, sperm morphology. This differentiates the acrosome index from most other functional or diagnostic tests (24). An additional important fact is that, although the numbers of this study are small, no overlap of the acrosome index values occurred between the <50% and :?:50% fertilization rate groups. The repeatability of the acrosome index evaluation within and between observers was within acceptable limits and compared well with the good repeatability of morphology evaluation according to strict criteria as reported previously (13). The acrosome index therefore can be regarded as a valid additional parameter for predicting IVF rates. The greatest potential use of the acrosome index is in the group of men with severe teratozoospermia, i.e., :;;4% morphologically normal spermatozoa, as reported previously (17). Acrosomal evaluation in this group will enable the observer to identify teratozoospermic patients with sufficient numbers (> 10%) of spermatozoa with normal acrosomes who thus will have a good chance to achieve a fertilization success of :?:50% with IVF. Men with an acrosome index of :;; 10% may be considered for intracytoplasmic sperm injection. The second link between sperm functional events and sperm morphology is through the relationship between acrosomal morphology (and thus sperm morphology) and acrosin activity, as a positive correlation (r = ; P < ) was found between these two variables. This relationship could be expected as several acrosomal enzymes, including proacrosin, which, when released from the acrosome during the acrosome reaction, is transformed to the active form acrosin, originate from the acrosome. The reduction, absence, or inhibition of acrosin activity is related to decreased sperm-zona binding and penetration of the zona pellucida (23, 25). Although a low correlation (r = ; P < 0.05) between acrosin activity and IVF rates was obtained in this study it, is important to remember that the fertilization process is a multifactorial event influenced by the results of histopathological parameters such as the acrosome index and functional tests such as the acrosin assay. To give a more dynamic view ofthe fertilizing capacity of a semen sample, it seems therefore more appropriately described by multidimensional statistics rather than by univariate statistics. The results of the multiple linear regression analysis showed that the acrosome index and the acrosin activity significantly contributes to the ex- planation of variation of the fertilization rates despite the low correlation obtained with univariate statistics. A cutoff point of> 18.0 /-tiu/10 B sperm with regard to an expected fertilization rate of :?:50% for acrosin activity was found with ROC curve analysis. A cutoff value for normal acrosin activity was set by Kennedy et al. (4) at :?:25 /LIU/10 B sperm. They furthermore proposed three diagnostic groups for the prediction of expected IVF rates viz :?:25 pju/10 B sperm = normal, 14 to 24 /LIU/10 B sperm = intermediate (grey), and /L13 /LIU/10 B sperm = low. Such prognostic cutoff points were not clear in our study, as for six men with an acrosin activity between 6.4 and 10.5 /LIU/ lob sperm the fertilization rate was :?:50%. With an acrosin activity of :?:19.3 /LIU/10 B sperm, the fertilization rate was :?:75%, except for four men in whom the fertilization rate was :;;40% only. It was interesting to note that three of these four cases presented with large (:?:5 /Lm) headed, but oval shaped, spermatozoa and, consequently, large acrosomes. All four cases fell in the severe teratozoospermia or poor prognosis (:;;4% morphologically normal spermatozoa) group. These four cases, especially the case with an acrosin activity of 53.3 /LIU/10 B sperm (case 4) are also responsible for the lack of a statistically significant difference for acrosin activity between the <50% and :?:50% fertilization rates groups and for the low correlation between acrosin activity and fertilization rates found in our study. This study also illustrated that it is possible to evaluate acrosomal morphology of spermatozoa using bright field microscopy. Originally, most of the studies on acrosome status have been performed by transmission electron microscopy (1, 8) or fluorescent staining techniques (2, 10, 11) to determine the presence of acrosin or the intactness and size of the acrosomes. Light (12) or phase-contrast microscopy (9, 10) was regarded to be insufficient to study acrosomal morphology in detail. However, with good optics at 1,000x or preferably 1,250x magnification and with good staining techniques, e.g., Papanicolaou (17), valuable information on acrosome morphology can be obtained. This also has been demonstrated indirectly by Jeulin et al. (12) and previously was acknowledged by another researcher (1). In summary, in the whole array of sperm functions, normal sperm morphology, the acrosome, and acrosin activity thus plays an vital role. The evaluation of acrosome morphology (i.e., acrosome index) therefore may become a very important guideline in the diagnosis of male fertility potential. However, when comparing the role of acrosome index, sperm morphology, and acrosin activity, especially on ground of the ROC curve analysis results (Fig. 4), it must be concluded that acrosin activity is not as Menkveld et al. Acrosome morphology and NF 643

8 sensitive as normal sperm morphology, which again is not as sensitive as the acrosome index in predicting the outcome of expected IVF rates. This is in agreement with the conclusion of Liu and Baker (5). Our results suggest that the acrosome index is a simple way of gaining more information on the potential functional ability of spermatozoa and therefore for the prognosis of a specific male's fertility potential. The greatest potential use of the acrosome index as a guideline for the fertility prognosis is in the group of men with severe teratozoospermia, i.e., :54% morphologically normal spermatozoa. However, further investigations between samples with an acrosome index of < 15% and :2: 15% may be warranted. Acknowledgments. We thank Kobie van der Merwe, M.Med.; Mr. Frik Stander, Ms. Amanda de Villiers, Ms. Karin Smith, and Evelyn Erasmus, M.Sc., of the Reproductive Biology Unit, Tygerberg Hospital, for skillful technical assistance; and Erik Schoute, M.Sc., IVF Center and Department of Epidemiology and Biostatistics, Free University of Amsterdam, for advice on statistical analysis. REFERENCES 1. Schill W-B. Some disturbances ofacrosomal development and function in human spermatozoa. Hum Reprod 1991;6: Senn A, Germond M, De Grandi P. Immunofluorescence study of actin, acrosin, dynein, tubulin and hyalurondase and their impact on in-vitro fertilization. Hum Reprod 1992; 7: Kruger TF, Haque D, Acosta AA, Pleban P, Swanson RJ, Simmons KF, et al. Correlation between sperm morphology, acrosin, and fertilization in an IVF program. Arch Androl 1988;20: Kennedy WP, Kaminski JM, van der Ven HH, Jeyendran RS, Reid DS, Blackwell J, et al. A simple clinical assay to evaluate the acrosin activity of human spermatozoa. J Androl 1989; 10: Liu DY, Baker HWG. Relationships between human sperm acrosin, acrosomes, morphology and fertilization in vitro. Hum Reprod 1990;5: Tummon IS, Yuzpe AA, Daniel SAJ, Deutsch A. Total acrosin activity correlates with fertility potential after fertilization in vitro. Fertil Steril 1991;56: De Jonge CJ, Tarchala SM, Rawlins RG, Binor Z, Radwanska E. Acrosin activity in human spermatozoa in relation to semen quality and in-vitro fertilization. Hum Reprod 1993; 8: Reichart M, Lederman H, Har-Even D, Kedem P, Bartoov B. Human sperm acrosin activity with relation to semen parameters and acrosomal ultrastructure. Andrologia 1993;25: Talbot P, Chacon R. A new procedure for scoring acrosome reactions of human sperm. Gamete Res 1980;3: Liu DY, Baker HWG. The proportion of human sperm with poor morphology but normal intact acrosomes detected with pisum sativum agglutinin correlates with fertilization in vitro. Fertil Steril 1988;50: Albert M, Gallo JM, Escalier D, Parseghian N, Jouannet P, Schrevel J, et al. Unexplained in-vitro fertilization failure: implication of acrosomes with small reacting region, as revealed by a monoclonal antibody. Hum Reprod 1992;7: Jeulin C, Feneux D, Serres C, Jouannet P, Guillet-Rosso F, Belaisch-Allart J, et al. Sperm factors related to failure of human in-vitro fertilization. J Reprod Fertil 1986; 76: Menkveld R, Stander FSH, Kotze TJvW, Kruger TF, van Zyl JA. The evaluation of morphological characteristics of human spermatozoa according to stricter criteria. Hum Reprod 1990;5: Wassarman PM. Fertilization in mammals. Sci Am 1988; 259: Kruger TF, Menkveld R, Stander FSH, Lombard CJ, Van der Merwe JP, van Zyl JA, et al. Sperm morphological features as a prognostic factor in in vitro fertilization. Fertil Steril 1986;46: Menkveld R, Kruger TF. Semen analysis-the Tygerberg experience. In: Acosta AA, Swanson RJ, Ackerman JB, Kruger TF, van Zyl JA, Menkveld R, editors. Human spermatozoa in assisted reproduction. Baltimore: Williams & Wilkins, 1990: Menkveld R, Oettle EE, Kruger TF, Swanson RJ, Acosta AA, Oehninger SC. Atlas of human sperm morphology. Baltimore: Williams & Wilkins, Franken DR, Acosta AA, Kruger TF, Lombard CJ, Oehninger S, Hodgen GD. The hemizona assay: its role in identifying male factor infertility in assisted reproduction. Fertil Steril 1993;59: 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: Menkveld R, Franken DR, Kruger TF, Oehninger S, Hodgen GD. Sperm selection capacity of the human zona pellucida. Mol Reprod Dev 1991;30: Liu DY, Baker HWG. Morphology of spermatozoa bound to the zona pellucida of human oocytes that failed to fertilize in vitro. J Reprod Fertil1992;94: Heywinkel E, Freundl G, Hofmann N. Acrosome reaction of spermatozoa with different morphology. Andrologia 1993; 25: Takano H, Yanagimachi R, Urch UA. Evidence that acrosin activity is important for the development of fusibility ofmammalian spermatozoa with oolemma: inhibitor studies using the golden hamster. Zygote 1993; 1: Jeyendran RS, Zaneveld LJD. Controversies in the development and validation of new sperm assays. Fertil Steril 1993; 59: Liu DY, Baker HWG. Inhibition of acrosin activity with a trypsin inhibitor blocks human sperm penetration of the zona pellucida. BioI Reprod 1993;48: Menkveid et al. Acrosome morphology and NF Fertility and Sterility