Immunohistochemical study of the cavernous nerves in the periprostatic region


Anthony J. Costello, Department of Urology, Royal Melbourne Hospital, Parkville, Melbourne, Victoria 3052, Australia.


Study Type – Therapy (case series)
Level of Evidence 4

What's known on the subject? and What does the study add?

The anatomy of the periprostatic tissues, in particular the fascial layers and neurovascular structures, remain a somewhat enigmatic and controversial area. Despite the considerable advances made in this area since the original dissections of Walsh and Donker almost thirty years ago, the precise location and function of these structures remains uncertain. The topic is of tremendous importance as we continue to refine surgical techniques to allow men the maximum opportunity to recover erectile function following radical prostatectomy. Yet many of these surgical refinements are not based on definitive knowledge of the relevant anatomy and clearly more knowledge is required.

For the first time, we have not just characterized the quantitative nature of the periprostatic nerves, but we also offer some insight into the likely functional nature of these nerves by selectively staining the sympathetic and parasympathetic nerves using immunohistochemical methods.


• To characterize the immunohistochemical nature of sympathetic and parasympathetic nerves surrounding the prostate.


• Using serial sectioning, four male cadavers were investigated using a combination of haematoxylin and eosin staining and immunohistochemistry. Both the sympathetic and parasympathetic contributions to the autonomic nervous system in the periprostatic region were assessed by staining analysis, the number of nerves fibres was quantified, their position relative to the prostate recorded and their function inferred.

• The fascial architecture of the neurovascular bundle (NVB) was also quantified.


• Approximately 27.8% of all nerve fibres identified were found on the anterior half of the prostate, above the 3 to 9 o’clock level. At the base, mid, and apex of prostate, parasympathetic fibres accounted for 4%, 5% and 6.8% of the nerves located on the anterolateral aspect of the prostate, respectively.

• Sympathetic nerves found above the 3 to 9 o’clock level represented ≈15% of the total number of nerves.

• When staining the periprostatic fascia, the classical NVB exhibited a distinct fascial architecture with three separate compartments.


• A tiny minority of nerves in the anterior periprostatic region are functionally significant parasympathetic nerves.

• There is little anatomical evidence to support higher incisions in the lateral prostatic fascia to spare cavernous nerve fibres, although such approaches may reduce the risk of traction injury on the more posterolaterally located NVB.

• The presence of distinct fascial compartments in the NVB is also confirmed.


haematoxylin and eosin


neuronal nitric oxide synthase


neurovascular bundle.


In recent years, our understanding of cavernous nerve anatomy and its implications for radical prostatectomy has undergone a considerable transformation. The original investigation by Walsh and Donker [1] into the causes of iatrogenic impotence localized the pro-erectile cavernous nerves posterolaterally to the prostate within a well defined neurovascular bundle (NVB). Using this knowledge, they adapted a radical prostatectomy technique to spare these structures with impressive outcomes [2]. Subsequently, however, both the exact positioning or the nerves and the concept of a well defined NVB has been challenged by the findings of several studies [3–6].

Some studies have located nerve fibres on the anterolateral aspect of the prostate, outside of the NVB [6], or have characterized a ‘spray-like’ distribution of autonomic nerves at the base and apex of the prostate [4]. It was hypothesized that a proportion of these anteriorly placed nerves are parasympathetic in nature and contribute functionality to the pro-erectile cavernous nerves. Although no evidence exists regarding the actual functionality of these nerves, some surgeons have developed nerve-sparing techniques aimed at preserving these structures [7,8]. This novel nerve-sparing approach, dubbed the ‘Veil of Aphrodite’ technique [9] or ‘curtain dissection’[7], releases the lateral prostatic fascia high on the aspect of the prostate and proceeds in an intrafascial plane [10]. Early reports from these centres have indicated improved postoperative potency using this technique [11,12]. However, there is little anatomical evidence to justify this approach, given that the higher placed nerves are most likely destined to innervate the prostatic stroma and not the cavernosal tissue of the penis [13].

The present study focussed on characterizing the position and nature of the autonomic nerves surrounding the prostate using specific immunohistochemical stains to clarify the likely functionality of the more anteriorly located nerves, therefore providing more evidence for the different nerve-sparing approaches.


Ethical approval for the present study was obtained from the Human Research and Ethics Committee of the Royal Melbourne Hospital and the University of Melbourne. Four blocks of pelvic tissue were serially histologically sectioned, from the prostatic base proximally to the apex distally. The blocks of tissue were from the pelves of two embalmed cadavers and two fresh cadavers, which had been hemisectioned. The hemisected pelvic blocks were then divided into 4-mm sections and embedded into paraffin for histological analysis (Fig. 1).

Figure 1.

Stylized representation of sectioning process. Lines indicate sites of sectioning (4 mm).

Each cadaver yielded ≈100 blocks of paraffin-embedded tissue: 50 from the right side and 50 from the left. Slides were cut at 5 µm and mounted for each block of tissue. All of these initial slides were stained with haematoxylin and eosin (H&E) and, based on a primary analysis, 50 representative blocks from only the fresh cadavers were selected. From these 50 blocks, a further two slides were cut at 3 µm and mounted from each block to undergo immunohistochemical staining using both tyrosine hydroxylase (sympathetics) and neuronal nitric oxide synthase (nNOS) (parasympathetics).

To localize the parasympathetics an antibody directed against nNOS was used. nNOS is a 150-kDa protein that is found in peripheral parasympathetic nerves and catalyses the formation of nitric oxide [14]. Nitric oxide is released by parasympathetics and is a potent vasodilator, implicated in the physiology of the male erection. The antibody itself was an undiluted mouse monoclonal antibody (Abcam, ab50267; Sapphire Bioscience Pty Ltd, Waterloo, NSW, Australia), which was used at a dilution of 1 : 500.

To localize the sympathetic nerves, we used a primary antibody directed against tyrosine hydroxylase. Tyrosine hydroxylase is the rate-limiting enzyme in the synthetic pathway of norepinephrine, which is a neurotransmitter found in peripheral sympathetic nerves and their associated ganglia. The antibody itself was an undiluted rabbit polyclonal antibody (Abcam, ab112; Sapphire Bioscience Pty Ltd), which was used at a dilution of 1 : 750.

Analysis of the stained slides and counting of the nerve fibres was conducted by an independent pathologist who was trained in urological pathology (J.P.). Each slide was analyzed according to nerve fibre number, type (somatic, parasympathetic or sympathetic) and position relative to the prostate. The analysis allocated nerves as they were observed to a particular sector as has been described previously [5,6] (Fig. 2).

Figure 2.

Method of sector analysis.


Immunohistochemical staining was successful using both tyrosine hydroxylase to identify sympathetic nerves and nNOS to identify parasympathetic nerves (Figs 3,4). Somatic nerves were identified on H&E stained sections.

Figure 3.

Immunohistochemical and haematoxylin and eosin (H&E) staining of sympathetic nerves. a, H&E staining of sympathetic ganglion lateral to the prostate. b, Tyrosine hydroxylase staining (brown) of same sympathetic ganglion.

Figure 4.

Immunohistochemical and haematoxylin and eosin (H&E) staining of parasympathetic nerves. a, H&E staining of parasympathetic nerves posterolateral to the prostate. b, neuronal nitric oxide synthase staining of same parasympathetic nerves posterolateral to the prostate.


The distribution and proportion of total nerve fibres is shown in Table 1. Parasympathetic nerves accounted for 43.3% of all nerve fibres present at the prostatic base, and 45.5% at the prostatic apex. Parasympathetic nerves found above the 3 to 9 o’clock level accounted for only 4%, 5% and 6.8% of the total number of nerves at the base, mid-prostate and apical regions, respectively.

Table 1.  Distribution of total nerve fibres expressed as number and percentage of total fibres at each level
  • *

    Nerve fibres in sectors 1, 2,7 and 8.

 Total, n (%)  111 (43.3) 99 (38.7)46 (18)256 (100)
 Total above 3 to 9 o’clock, n (%)* 10 (4) 39 (15.2)21 (8.2) 70 (27.4)
 Total, n (%)144 (44.7)125 (38.8)53 (16.5)322 (100)
 Total above 3 to 9 o’clock, n (%)* 18 (5) 45 (14)24 (7.5) 87 (26.5)
 Total, n (%)100 (45.5) 86 (39.1)34 (15.5)220 (100)
 Total above 3-9 o’clock, n (%)* 15 (6.8) 34 (15.5)16 (7.3) 65 (29.6)

The proportion of nerves characterized as sympathetic remained relatively constant from base to apex: 38.7–39.1%. Sympathetic nerves found above the 3 to 9 o’clock level represented ≈15% of the total number of nerves at any given prostatic level.

Somatic nerves corresponded to 18%, 16.5% and 15.5% of the total nerve fibres at the base, mid-prostate and apex, respectively. Somatic nerves found above 3 to 9 o’clock level represented 7.5% of the total nerves at each level of the prostate.

Some 25% (27.4% base, 26.5% mid-prostate and 29.6% apical) of all the nerve fibres identified were found on the anterior half of the prostate, above the 3 to 9 o’clock level.


At the prostatic base, parasympathetic nerves accounted for only 14.3% of the nerves located on the anterior aspect of the prostate. However, these nerves correspond to only 9% (Table 2) of the total number of parasympathetic nerves found at the base. A significant proportion of the parasympathetic nerves at the base were localized to the region of the NVB (sectors 4 and 5). At this level, 69.4% of the parasympathetic nerves were found within the anatomically defined NVB.

Table 2.  Distribution of parasympathetic nerve fibres (fresh cadavers) as a percentage of total parasympathetic fibres at that level
PositionIn NVB*Above 3 to 9 o’clock
  • *

    Within sectors 4 and 5.

  • †Within sectors 1, 2, 7 and 8. NVB, neurovascular bundle.

Prostate base69.4 9
Mid-prostate65.3 11.1
Prostate apex5715

Similarly, at the mid-prostate, 65.3% of the parasympathetic nerves were found inside the NVB. Only 11.1% of parasympathetics at this level were found above the 3 to 9 o’clock junction, which, as a proportion of the total nerves at this level, represents 18.8% of those nerves found on the anterior aspects of the prostate.

At the apex, 57% of all parasympathetic fibres were found within the region of the NVB, with 15% found above the 3 to 9 o’clock level. Importantly, the proportion of parasympathetic nerves found within the NVB dropped slightly from the base to the apex (see Table 3).

Table 3.  Functional nerve distribution in anterolateral sectors as a percentage of fibres in sectors 1, 2, 7 and 8
Prostate base14.355.730
Prostate apex23.152.318.6


The fascial architecture of the posterolaterally located NVB was analyzed in embalmed cadavers. In total, 32 representative tissue blocks were used. The findings are summarized in Table 4.

Table 4.  Fascial architecture of neurovascular bundle
ArchitectureNumber of slides
No distinct fascial compartments 3
Two fascial compartments 8
Three fascial compartments18
More than three fascial compartments 3

In 18 of the 32 slides examined, the NVB exhibited a distinct fascial architecture with three separate compartments. Figure 5 illustrates the compartmental architecture found within the NVB of a slide. Three compartments containing nerves and blood vessels can be clearly seen. This fascial tissue is composed of fused levator ani, Denonvilliers’ and prostatic fascia.

Figure 5.

Compartmental architecture of the neurovascular bundle (mid-prostate). Left: haematoxylin and eosin slide showing compartmental neurovascular bundle architecture. Right (overlay): prostate (green), fascial bands (blue), nerves (yellow), para-rectal tissue (grey), levator ani musculature (spotted pink).


The quest for improved outcomes subsequent to nerve-sparing radical prostatectomy continues. Younger, better-informed patients and competing surgical and other technologies have ensured that the ‘trifecta’ outcome remains of paramount importance. Although our understanding of the complex neuroanatomy and fascial investments of the prostate has improved considerably [15], there remains some uncertainty about the precise autonomic nerve supply in the context of nerve-sparing radical prostatectomy. Much of the complexity relevant to the urological surgeon centres on the course of the pro-erectile parasympathetic fibres, also known as the cavernous nerves. Injury to these structures is invariably linked to erectile dysfunction post-radical prostatectomy.

Walsh and Donker’s original dissections in fetal cadavers showed a NVB located posterolaterally [1] and led to the ‘standard’ technique of nerve-sparing, which releases the NVB from the posterolateral groove between the prostate and the rectum [2]. Although these anatomical findings have been confirmed and refined [13,16], there remains speculation that nerves with pro-erectile functional relevance exist more anterolaterally above the 3 and 9 o’clock positions [3,6,17]. In their studies of resected prostates, Eichelberg et al.[6], Kiyoshima et al.[3] and Tewari et al.[4] have identified nerves in a higher, more anterolateral position than previously identified. It was speculated that a proportion of these nerves were parasympathetic in nature and, as a result, contributed fibres to the pro-erectile fibres of the cavernous nerves. Similarly, in a study of non-nerve sparing radical prostatectomies, Ganzer et al.[5] used novel planimetry software to characterize the topographical anatomy of periprostatic and capsular nerves. The percentage of total nerve surface area was highest dorsolaterally (84.1%, 75.1% and 74.5% at the base, middle and apex, respectively) but variable. In addition, there was a significant decrease in total nerve surface area from the base over the middle towards the apex of prostate as the periprostatic nerves penetrated the capsule to directly innervate the prostate. Evidence of the functionality of these anterolaterally placed nerves is limited [18] and, in the absence of entire anatomical blocks of tissue to determine their course, speculation about the value of preserving more anteriorly located nerves remained unsubstantiated.

Nevertheless, there have been modifications to the standard nerve-sparing approach to radical prostatectomy in an attemot to improve erectile function outcomes based on the prospect of more functionally relevant pro-erectile nerves in these areas. Although a high incision of the lateral prostatic fascia (‘curtain dissection’) was first proposed by Lunacek et al.[7] based on anatomical studies, it was the description of the ‘Veil of Aphrodite’ technique by Menon et al.[8–10] that provoked the recent interest in this modification. The Veil technique describes the higher release of prostatic fascia on the anterolateral aspect of the prostate with preservation of the prostatic fascia (i.e. intrafascial dissection) [9]. Their outstanding report of 97% potency in men undergoing this technique attracted much attention and controversy [11] and has not been reproduced elsewhere. Chabert et al.[19] subsequently adopted this technique for nerve-sparing laparoscopic radical prostatectomy and failed to show any value in the ‘curtain dissection’ technique, raising the question of whether there is sufficient anatomical evidence to justify this approach.

The present study has characterized the periprostatic nerves, including the cavernous nerves, by immunohistochemical analysis of sympathetic and parasympathetic nerves. We confirmed that autonomic nerve fibres were present on the anterolateral aspects of the prostate between the prostate and lateral prostatic fascia; however, only a small proportion of these were parasympathetic nerves likely to be of functional relevance. At the mid-prostate level, only 18.8% of the nerves found on the anterior aspect of the prostate were parasympathetic in nature, with the most parasympathetic nerves (68%) being found in the previously defined NVB [1] posterolateral to the prostate. Most of the nerve fibres found on the anterolateral regions examined were sympathetic in nature. The sympathetic nerves contribute significant innervation to the prostatic stroma, are responsible for innervation to the vascular structures in the region, and, as such, may extend outside the typical NVB, which is predominantly parasympathetic by immunohistochemical characterization. The external urethral sphincter located immediately distal to prostatic apex, receives input from the autonomic sympathetic nerves [16,20]. The sympathetics may course over the anterior aspects of the prostate to provide innervation to the anterior external urethral sphincter.

At the apex, the absolute number of the parasympathetic fibres above the 3 to 9 o’clock junction increased slightly. This is consistent with studies that show the cavernous nerves ascending to assume a higher position distal to the apex [17,21]. Takenaka et al.[17] claimed the cavernous nerves assume a higher 2 to 10 o’ clock positioning at the apex. The present study did not confirm these findings; however, careful apical dissection and ligation of the dorsal venous complex is of critical importance for preserving the neural anatomy.

The anteriorly placed parasympathetic nerve fibres are likely to be destined for innervation of the prostatic stroma and not the corpora cavernosa of the penis. This observation is supported by two pieces of evidence. First, the total number of visible nerve fibres was smaller at the prostatic apex as compared to the base. This decrease in the number of fibres has been reported previously [3,5,6,22] and may be related to a significant proportion of nerves penetrating the prostate and other structures, including seminal vesicles, levator ani and rectum. In the fresh cadavers included in the present study, 134 nerve fibres were located at the base compared to 115 at the apex. Of these, the absolute number of parasympathetic nerves found on the anterior half of the prostate at the apex was eight, further supporting this hypothesis.

The second piece of evidence supporting the prostatic innervation hypothesis of these nerve fibres is the fascial architecture of the NVB itself. In 12 microdissected cadavers Costello et al.[3] reported distinct compartments within the NVB, each supplying separate structures. A similar compartmental structure of the NVB was reported in the present study. In 18 of 32 of representative tissue block sections, the predominant fascial architectural pattern within each NVB observed involved three distinct compartments, the most medial of which conveyed a significant proportion of the parasympathetic nerve fibres. This indicates that the cavernous nerves probably remain within a distinct structural compartment and the higher parasympathetic nerves do not contribute to the pro-erectile innervation of the penis. This provides the putative basis for an incremental approach to nerve-sparing radical prostatectomy, whereby a small amount of extracapsular tissue may be resected in the posterolateral region to ensure adequate cancer control in patients with a suspicion of high-volume or extracapsular disease in this area. Within the compartmentalized fascial architecture of the NVB, the main parasympathetic supply to the cavernosal nerves is located slightly lateral to the autonomic nerves supplying the prostate. An incremental approach would sacrifice these nerves and preserve the main cavernosal supply. Figure 6 illustrates this anatomy, showing the neural elements in the other compartments innervating levator ani, the rectum and the prostatic stroma.

Figure 6.

Illustrative representation of periprostatic autonomic innervations. Denonviller’s Fascia (DF), Levator ani (LA), lateral prostatic fascia (LPF), pararectal fat (PF), rectum and prostate (P), rectum (Rec).

Although the present studies have not confirmed the presence of significant numbers of parasympathetic nerves in the anterolateral periprostatic fascia, there may be other explanations for the improved potency outcomes when a high incision of this fascia is performed during nerve-sparing radical prostatectomy. We agree with Nielsen et al.[23] that the high release of the fascia in this apical region produces less traction on the posterolateral NVB and less thermal damage in this region. Thus, the neural elements below are protected from a neuropraxia-type injury.

The limitations of the present study include: (i) the relatively small number of cadaveric specimens; (ii) the lack of physiological correlation between immunohistochemically confirmed parasympathetic fibres and erectile function; and (iii) the lack of detail about the course of the autonomic nerves beyond the apex of the prostate.

In conclusion, the present studies confirm that autonomic nerve fibres do exist on the anterior aspects of the prostate, above the 3 to 9 o’clock level. However, only a few of these fibres correspond to the pro-erectile parasympathetic cavernous nerves, with most parasympathetic nerves existing in the posterolateral region of the prostate. A compartmentalized fascial architecture of the NVB is confirmed and supports the notion of incremental nerve-sparing. Further studies are required to correlate these immunohistochemical findings with functional outcomes.


The authors acknowledge Dr Levent Efe for his assistance with the graphical illustrations.


None declared.