Possible inflammatory pain biomarkers in postamputation pain

Postamputation pain, which includes both phantom pain (pain referred to the missing limb) and stump pain (pain in the residual limb), affects up to 60–80% of all amputees [1] and is often associated with sensory abnormalities such as hypoaesthesia or hyperalgesia and allodynia. Postamputation pain are notoriously difficult to treat, and although a large variety of treatment options have been attempted, no effective pharmacological treatment has yet been found [2]. The pathogenesis of postamputation pain is multifactorial, with both peripheral and central mechanisms playing an important role. Through the last decades, studies in both animals and humans have suggested that the immune system may play an important role in different neuropathic pain conditions, including postamputation pain [3], [4], [5], [6], [7].

Sensory afferent axons in the skin are in close proximity to cells that are responsible for the innate immune response, including Langerhans cells, keratinocytes, and mast cells [4]. Activation of nociceptors in the afferent neurons in the skin releases neuropeptides such as substance P (SP) and calcitonin gene-related peptide (CGRP). It has been shown that CGRP is increased in the keratinocytes of patients with type 1 complex regional pain syndrome (CRPS) and in patients with post-herpetic neuralgia [8]. A study in rodents showed that activation of the mast cell receptor Mrgprb2/X2 by the neuropeptide SP leads to cytokine release and recruitment of immune cells, further facilitating the inflammatory cascade and sensitization of peripheral afferents [9]. Another important cell line in neuroinflammation and pain transduction is macrophages. A recent study showed that macrophages in the skin play a critical role in angiotensin II-induced pain hypersensitivity in mice, and that the macrophage marker IbA1 is increased in skin biopsies in humans with painful diabetic polyneuropathy [10].

The intraepidermal nerve fiber density (IENFD) is decreased in patients with peripheral neuropathy (e.g. caused by diabetes or chemotherapy) compared with healthy controls. However, no difference in IENFD has been found between painful and non-painful neuropathies [11], [12], and IENFD is therefore not believed to be an appropriate biomarker of pain. Thus, neuropathic pain in patients with neuropathy cannot be explained by a decrease in intraepidermal nerve fibers (IENFs). The density of IENF and inflammatory biomarkers of pain has also been investigated in human studies after peripheral nerve injury due to surgery. In a study by Martinez and colleagues, the authors investigated if changes in IENF, TNF-α and neurotrophic growth factor (NGF) in skin biopsies 3 months after iliac crest bone harvest were related to neuropathic pain. While the authors did not find any correlation between IENF, TNF-α and NGF and postsurgical pain, they found that IENFD correlated with burning pain and pain evoked by compression and brushing [13]. In another study, by Springer and colleagues, there were no relation between the density of IENF in skin biopsies and chronic pain after thoracotomy [14].

In order to improve the mechanism-based treatment for postamputation pain, a better understanding of the underlying pathophysiological mechanisms is needed, including insight into the role of inflammation. In this study, we hypothesized (1) that the expression of inflammatory markers such as macrophages and the neuropeptides SP and CGRP would be more prominent in skin biopsies from the amputated site compared with the control site in amputees with different types and pain intensities of postamputation pain, and (2) that spontaneous pain intensity would be correlated to the level of macrophages and SP/CGRP expression.

Data were obtained as a part of a larger observational study, investigating if ultrasound-verified neuromas are more prevalent in amputees with postamputation pain (clinicaltrials.gov Identifier: NCT03051113). The study was approved by the Regional Research Ethics Committee of Central Denmark (1-10-72-416-14) and the Danish Data Protection Agency (1-16-02-307-15). Written informed consent was obtained from all patients according to the Declaration of Helsinki. Amputees were recruited from Aarhus University Hospital, Denmark, and from the Facebook group “Active Amputees in Denmark” from August 2017 to May 2018. Inclusion criteria were age >18 years and amputation of an extremity or part of an extremity >6 months ago.

The study consisted of a single one-day clinical assessment. The following baseline information was collected: year, cause, and localization of the amputation; type of postamputation pain if any (no pain, stump pain, phantom pain, or both stump and phantom pain). Amputees were asked to recall the average intensity of spontaneous stump and phantom pain during the past 7 days on an 11-point numerical rating scale (NRS) where 0=no pain and 10=worst imaginable pain. Patients also completed the Neuropathic Pain Symptom Inventory (NPSI) that includes 12 questions about neuropathic pain characteristics [15].

Evoked responses to cold (20 °C), warm (40 °C) (Rolltemp; Somedic AB, Hörby, Sweden), brush (SENSElab Brush no. 5; Somedic AB, Hörby, Sweden), and pinprick stimuli with a monofilament (Touch Test™, 60 g, Stoelting Co., Wood Dale, IL, USA) were obtained, and patients were asked to rate if the sensation was decreased (hypoaesthesia), normal, or increased (hyperaesthesia). Intensities of dysaesthesia and pain were rated on two separate 0–10 scales (0=no dysaesthesia/no pain and 10=worst imaginable dysaesthesia/worst imaginable pain). After the assessment of evoked responses, skin biopsies were taken. Subcutaneous anaesthesia was achieved with 1% lidocaine before the skin biopsy was taken under sterile conditions. The same small amount (0.5 mL) of lidocaine was injected subcutaneously in all amputees and the biopsies were taken adjacent (around 3–5 mm) from the injection site, to ensure uniformity across amputation site and control site. Two biopsy samples were taken: one from the area of the stump with spontaneous stump pain or with evoked dysaesthesia/allodynia and another from the corresponding area on the contralateral extremity using a sterile disposable 3 mm punch (Miltex, York, PA, USA). The biopsies were fixed in Zamboni’s fixative for 16–24 h and cryoprotected overnight, followed by embedding in O.C.T. (Fisher Scientific UK Ltd, Loughborough, UK) before being stored at −80 °C until immunostaining.

The biopsies were stained using the following antibodies on free floating 50 μm thick sections: PGP 9.5 (1:2,000; Zytomed, Berlin, Germany), SP (1:1,000; Immunostar, Hudson, WI, USA), CGRP (1:500; Immunostar, Hudson, WI, USA) and IbA1 (1:500; Wako Chemicals, Richmond, VA, USA). The PGP 9.5 and IbA1 stainings are described in details elsewhere [10], [16]. The SP and CGRP stainings were performed following the IbA1 staining protocol.

The IENFD was estimated following the standard counting protocol [17] and the IbA infiltration as described in Shepherd et al. [10]. Briefly, PGP 9.5+ nerve fibers crossing the epidermal-dermal junction were quantified in at least three 50 μm thick sections and the number divided with the section’s length. IbA1 infiltration was quantified using ImageJ (NIH and LOCI, WI), where a threshold RBG intensity was set prior to the analysis, and the area of the region of interest that exhibited fluorescence above the threshold was reported as a percentage value [10]. Approximately 0.5 mm² of the dermis was measured. The dermal length of SP- and CGRP-positive fibers was obtained using the 3D stereological sampling technique global spatial sampling (GSS), as described in details elsewhere [18]. Briefly, GSS removes the known estimation bias when counting tubular objects (e.g. nerve fibers) that do not have an isotropic orientation, resulting in an uneven chance of being counted. This is done by generating virtual isotropic planes using the NewCAST software (Visiopharm, Hoersholm, Denmark), thus ensuring that all fibers have the same chance of being counted in an unbiased manner [16]. Additionally, the total density of peptidergic fibers can be obtained by adding CGRP+ fibers with SP+ fibers.

Eleven amputees were included in the study. Subject characteristics are summarized in Table 1.

Figure 1 shows skin biopsy stainings of CGRP (A1-2), SP (B1-2), and IbA1 (C1-2). No significant differences were found in SP, GCRP, total peptidergic fiber counts or IbA1 between the amputated site and the control site (Table 2). Furthermore, no correlation was found between the worst spontaneous pain intensity the last 7 days (NRS, 0–10) and IbA1 or peptidergic fiber count (Fig. 2). Type of spontaneous pain (stump or phantom) and evoked responses including hypoaesthesia, hyperaesthesia, and dysaesthesia/allodynia to brush, cold, warm, or pinprick were not associated to IbA1, SP, and CGRP. As expected, no differences in IENFDs were found between the amputated site and the control site.

Fig. 1:

Immunofluorescence images showing examples of increased immunostainings in three patients at the amputation site (upper panel) compared with the control site (lower panel). A1+A2: Increased CGRP+ fibers from the amputated site (A1) compared with the control site (A2) in the same patient. B1+B2: Increased SP+ nerve fibers from the amputated site (B1) compared with the control site (B2) in the same patient. C1+C2: Increased IbA infiltration from the amputated site (C1) compared with the control site (C2) in the same patient. Scale bar: 100 μm (except for C1 and C2, where the scale bar represents 200 μm).

Fig. 2:

(A) Correlation analysis (Spearman correlation) between worst pain intensity the last 7 days (stump or phantom pain) and total peptidergic fibers (mm⁻²) (p=0.5). (B) Correlation analysis (Spearman correlation) between worst pain intensity the last 7 days (stump or phantom pain) and skin IbA1 immuno-staining density (% of ROI) (p=0.71).

Overall, we did not identify either IbA1, CGRP, or SP as possible markers of postamputation pain in skin biopsies of amputees. The strengths of our study include a robust study design and a cohort of amputees that were carefully examined for both spontaneous and evoked pain. However, we acknowledge that our study has some limitations. First, and most importantly, our study only included a limited number of amputees who were hetereogenous with respect to type, site, and cause of amputation. Second, we only examined three inflammatory markers, and it cannot be excluded that other neuropeptides and cytokines are involved. Third, we did not examine if there was any systemic activation of the immune system as we only investigated inflammatory markers in the skin. In fact, a proinflammatory systemic profile has been reported in patients with neuropathic pain including amputees with postamputation pain, with elevated levels of tumor necrosis factor (TNF) and certain cytokines and interleukins [7], [19], [20].

Even though this study did not find any correlation between postamputation pain and macrophages and the neuropeptides SP and CGRP, we find that the method of this study is a novel and promising way of investigating and understanding changes in the skin in amputees with postamputation pain. In future studies, staining of cytokines and other inflammatory mediators in skin biopsies in a larger and maybe more homogeneous cohort of amputees could provide new insight into the mechanisms of postamputation pain.