FRAC assay identifies a subgroup of infertile males with reduced centriole quality

We selected 15 males with different forms of infertility who underwent infertility treatments with either IVF or ICSI. Six male patients had a normal semen analysis, and nine had male factor infertility (Online Resource 3). Specifically, for semen analysis, two patients had asthenozoospermia, three patients had asthenoteratozoospermia, one patient had teratozoospermia, one patient had oligoteratozoospermia, one patient had oligoasthenozoospermia, and one patient had oligoasthenoteratozoospermia. Semen samples were analyzed through Fluorescence-Based Ratiometric Analysis of Sperm Centrioles (FRAC) using the three previously described biomarkers: POC1B, acetylated tubulin, and tubulin [26]. In total, 1578 spermatozoa were quantified, ranging from 71 to 127 per patient (105 ± 15). The FRAC findings from the 15 infertile patients were compared to the standard obtained from fertile males collected by the RSRSR biorepository described in our previous study by Jaiswal and colleagues [26].

The comparison identified seven infertile males with no outlier values and, therefore, with a negative (i.e., normal) FRAC score. We found eight infertile males with at least one outlier each and, thus, with positive (i.e., abnormal) FRAC values. Of the three proteins analyzed, acetylated tubulin had 12 outliers, with six in the proximal centriole, one in the distal centriole, and five in the axoneme. Tubulin had only two outliers, one in the proximal centriole and one in the axoneme. POC1B, a centriole-specific protein, had no outliers (Fig. 2). Similar profiles of outliers, where acetylated tubulin is the most common, were identified in our previous studies in humans and bovine [25, 26].

Fig. 2

The studied population had seven couples with negative FRAC values and eight with positive FRAC values. Graph depicting the average FRAC ratio from each male (± 95% confidence intervals) for the nine parameters (for example, acetylated tubulin PC is a parameter). Each parameter shows male patients 1 to 15 from left to right. Black bold bars at the center of each group of confidence intervals represent the reference average plus/minus two standard deviations. Arrows point at 95% confidence intervals up to one standard deviation (blue arrows) or more than two standard deviations (orange arrows) outside the reference range

Nucleolus Precursor Body polarization scoring was determined based on five zones

During zygote progression towards cleavage, pronuclei NPB transition from evenly distributed to enrichment at the pronuclei edges facing each other [36, 41]. We quantified NPB polarization in videos generated in the clinical setting from ICSI/IVF embryos based and adapted on the approach described by Cavazza and colleagues [37], Scott [36], and Tesarik and Greco [41]. We excluded from this calculation any embryo where a conclusive determination of polarized NPBs was impossible or if the video started too late to capture the two pronuclei (2PN) stage.

To quantify Nucleolus Precursor Body polarization, we divided each pronucleus into two zones: the half towards the junction (J) and the half away from the junction (A) (Fig. 3). Although the ideal conditions for observing pronuclei were when they were side-by-side with no overlap, that was often not the case in clinical videos. The two pronuclei overlap in most cases, creating an Overlap (O) zone. These divisions made five scorable zones: A1, J1, O, J2, and A2. Zones J1 and J2 were the pronuclei halves where the two pronuclei meet at their junction, and NPBs in these zones counted as polarized. Zone O was the two pronuclei overlap location, with NPBs counting as polarized here as well. Zones A1 and A2 were the pronuclei halves away from the junction, and NPBs in these zones counted as nonpolarized. If the overlap was more than 50% for one or both pronuclei of an embryo, their NPBs were not counted, and that embryo was excluded from analysis. We considered the center of the NPB if it was at the border of A and J zones or the border of J and O zones. For each embryo, we assigned an NPB polarization score using the equation below:

Fig. 3

Three scenarios of NPB polarization in the pronuclei of zygotes. a–c Represented model (top row) and annotated image (bottom row). Each pronucleus was divided into two halves, separated into either an “A” zone, the half that was furthest away from the junction, or a “J” zone, the half that was towards the junction; a a scenario of full NPB polarization in the two non-overlapping pronuclei. There were six NPBs (orange) in total, all in either J1 or J2 zones; b a scenario of full NPB polarization in the two overlapping pronuclei. There were nine NPBs in total, with five being in either J1 or J2 zones and four NPBs (green) being in the “O” zone, the zone created by the overlap of the two pronuclei; c a scenario showing partial NPB polarization in the two pronuclei with overlap. There were nine NPBs in total, with three NPBS being in the “J1” or “J2” zones, two NPBs in the “O” zone, and four NPBs (blue) being in the “A2” zone

$$\text=\frac1+\text+\text2}1+\text1+\text+\text2+\text2}*100$$

A polarization score at or above 91% was counted as polarized, and anything below 91% was classified as non-polarized. This 91% cutoff was obtained from pre-analysis, where over half of abnormal FRAC couples were above the cutoff (Online Resource 4). NPBs were most polarized anywhere from 18.5 to 25.6 h (21.9 ± 2.1 h) after fertilization and 1.8 to 8.9 h (4.5 ± 2.0 h) before pronuclear breakdown. Additionally, the average amount of NPBs in an embryo belonging to a normal FRAC couple was 9.1 ± 2.5, and to an abnormal FRAC couple was 11.4 ± 3.5 (p < 0.0001, T-test).

Three patterns of NPB polarization progression were observed in the zygote

The patterned NPB polarization approach calculated the NPB polarization rates from the first NPB appearance to their disappearance shortly before the first cell division in 10-min frame intervals. We visualized zygote centrosome function from as early as 15 h post-fertilization (18.1 ± 2.0) up to 40 h (25.6 ± 3.9) post-fertilization. This comprehensive approach allowed for a pattern of NPB polarization formation to be seen (Online Resource 5). Three different patterns were observed:

1.

Full polarization patterns with consistently high NPB polarization rates above or equal to 91% were observed in 44 embryos.

2.

Eventual polarization patterns in which NPB polarization rates rose from low (below 91%) to high (equal to or greater than 91%) were observed in 35 embryos.

3.

Reduced polarization patterns with consistently low NPB polarization rates (below 91%) were observed in 61 embryos.

An interesting pattern was observed in one embryo from couple 6, where it went from highly polarized (≥ 91%) to low-to-mid-polarized (< 91%). Additionally, in the reduced polarization patterns, 11 embryos had pronuclei go above the 91% polarization cutoff during polarization but still started and ended well below the cutoff. Eight of these embryos belonged to abnormal FRAC couples, and only three belonged to normal FRAC couples.

Some couples with abnormal FRAC have reduced NPB polarization

NPB polarization patterns were further examined for each couple to look for predictors of spermatozoa and zygotic defects (Fig. 4a). The couples with normal FRAC values (couples 1, 2, 4, 7, 8, 9, and 11) had similar NPB polarization pattern distributions, with the reduced NPB polarization pattern being at or below 44% (26 ± 17%).

Fig. 4

Half of the abnormal FRAC couples had high rates of reduced polarization pattern. a Individual breakdown for polarization patterns for each couple. Couples 1, 2, 4, 7, 8, 9, and 11 had normal FRAC scores. Most of their embryos are in either the full (green) or eventual (orange) polarization pattern. Couples 3, 5, 6, 10, 12, 13, 14, and 15 had abnormal FRAC values. Couples 3, 5, 6, and 10 had similar pattern distributions as those in the normal FRAC category. In contrast, couples 12, 13, 14, and 15 had a majority (couples 12 and 14) or all (couples 13 and 15) of their embryos classified as reduced (red) polarization. A dotted red box is placed around the four abnormal FRAC couples with the lowest patterned NPB polarization rates (or the highest rates of the reduced polarization pattern); b most outlier values are concentrated in the acetylated tubulin biomarker. PC, proximal centriole; DC, distal centriole; Ax, axoneme

In contrast, couples with abnormal FRAC values had two distinct populations observed. Couples 3, 5, 6, and 10 had a distribution of embryos similar to couples with normal FRAC values, having a reduced NPB polarization embryo pattern at or below 50% (37 ± 13%). Couples 12, 13, 14, and 15 had a distribution of embryos that heavily favored the reduced NPB polarization pattern, with couples 12 and 14 at 71% and 88%, respectively, and couples 13 and 15 at 100% of their embryos ascribed to this pattern. This finding suggests that couples 12, 13, 14, and 15 likely had a spermatozoa centriolar defect that manifested as a zygotic phenotype. Indeed, a T-test comparison of the patterned NPB polarization approach of couples 3, 5, 6, and 10 to that of couples 12, 13, 14, and 15 found significant differences in the three patterns and their distribution (Full polarization pattern, p = 0.011; eventual polarization pattern, p = 0.005; reduced polarization pattern, p = 0.001). Importantly, there was not a statistically significant difference between the ages of the two sub-populations in abnormal FRAC couples (couples 3, 5, 6, and 10; 31.5 ± 1.6 years male age and 30.6 ± 3.4 years female age; couples 12, 13, 14, and 15; 33.0 ± 3.0 years male age and 32.2 ± 3.8 years female age; p = 0.408, male years, and p = 0.555, female years).

No significant differences in the individual NPB polarization time courses were observed between couples 3, 5, 6, and 10 and couples 12, 13, 14, and 15 (Online Resource 6).

We also analyzed the FRAC data to distinguish any biomarker differences between the couples with abnormal FRAC and the subgroups with abnormal or normal NPB polarization (Fig. 4b). Although the majority of outlier values were observed in acetylated tubulin in both subgroups, no biomarker differences that could further separate the two abnormal FRAC subgroups were observed.

Fifty-six percent of abnormal FRAC embryos were classified as having reduced polarization pattern

When examined at a couple level, testing for statistical correlation, patterned NPB polarization was found to be significantly different between couples with normal versus abnormal FRAC values (p < 0.0001, chi-square test of independence) (Fig. 5a). A higher rate of reduced polarization patterns in abnormal FRAC couples primarily drives the difference.

Fig. 5

Normal FRAC embryos had fewer reduced polarization pattern rates when compared to abnormal FRAC embryos. a Patterned NPB polarization of couples with normal FRAC and abnormal FRAC. Each column represents an average ± 1 SD. Each dot represents a couple (seven per column in normal FRAC and eight per column in abnormal FRAC). Each couple is color-coded to match those in Fig. 2. A dotted red box is placed around the four abnormal FRAC couples with the lowest patterned polarization rate (or the highest rates of the reduced polarization pattern). b Patterned NPB polarization of embryos belonging to normal FRAC couples (n = 62 embryos) and abnormal FRAC couples (n = 78 embryos). Each column represents the overall percentage observed in our study. A chi-square test of independence was used for embryo population analysis. FP, full polarization; EP, eventual polarization; RP, reduced polarization

The normal FRAC couples had 62 embryos, and abnormal FRAC couples had 78 embryos, for a total of 140 embryos. When examined at an embryo level, patterned NPB polarization was found to be significantly different between the embryos of normal FRAC couples and abnormal FRAC couples (p = 0.0024, chi-square test of independence) (Fig. 5b). Analysis of the results found the largest difference in the reduced polarization (RP) pattern between normal (27%, n = 17 embryos) and abnormal (56%, n = 44 embryos) FRAC couples. Smaller differences were found in the full polarization (FP) and eventual polarization (EP) patterns between normal (FP: 39%, n = 24; EP: 34%, n = 21) and abnormal (FP: 26%, n = 20; EP: 18%, n = 14) FRAC couples.

A receiver operating characteristic (ROC) curve test was calculated on the 15 couples for the full, eventual, and reduced polarization scores. The area under the curve (AUC) value for full polarization was 0.786, eventual polarization was 0.732, and reduced polarization was 0.839. An AUC value of 0.7 to 0.8 is moderate, 0.8 to 0.9 is excellent, and > 0.9 is outstanding [43]. Thus, our study found a moderate to excellent correlation.