Dear Editor,

Bortezomib is a new chemotherapeutic agent approved for the treatment of multiple myeloma (MM) [1]. Length-dependent painful neuropathy, mainly related to the loss of c-fibers, but also to deficit in Aδ and Aβ fibers [2, 3], is a common and clinically significant drug-related adverse event and the first cause of dose reduction or treatment suspension [4]. Large and small fibers damage has been related to the duration of bortezomib treatment in rat models, where a short-term treatment induced, predominantly, Schwann cells and myelin damage [5], whereas a long-term exposure produced a prominent axonopathy of the unmyelinated fibers [6]. A small fiber neuropathy (SFN) was previously described in patients treated with bortezomib [4, 7], however, though autonomic symptoms like syncope, impotence, gastrointestinal and urinary disturbances have been reported [4, 8, 9], autonomic neuropathy has never been characterized.

We report three women affected by MM with IgG kappa paraproteins and treated with bortezomib (patients 1 and 2) or an association of bortezomib and thalidomide (patient 3) (Table 1). After 3–5 months of therapy, all patients developed neuropathic pain and discontinued bortezomib therapy (Table 1). Symptoms decreased slightly in patients 1 and 2 after 2 months from suspension, whereas patient 3 experienced a marked improvement. Patients underwent neurological examination and a study of motor (tibial, peroneal, median, ulnar nerves) and sensory (sural, median, ulnar nerves) nerve conductions to evaluate large myelinated nerve fibers. To evaluate somatic and autonomic peripheral small nerve fibers, a 3 mm punch biopsy was taken from the distal leg (10 cm above the lateral malleolus) and thigh (15 cm above the patella). Specimens were fixed, sectioned and incubated with primary antibodies, as previously described [10]. Primary antibodies included the pan-neuronal marker protein gene product 9.5 (PGP 9.5) and specific autonomic antibodies like vasoactive intestinal peptide (VIP) to mark cholinergic fibres innervating sweat glands, and dopamine beta hydroxylase (DBH) to visualize adrenergic fibers innervating blood vessels and arrector pilorum muscles. The ENFs was calculated per linear millimeter of epidermis. As previously described [10], autonomic innervation was graded from 0 (no innervation) to 4 (normal pattern, morphology and quantity of fibers) by a semi-quantitative scale showing a low interobserver and intraobserver variability [10]. Fifteen age-matched healthy subjects (62 ± 9 years) served as controls. Somatic and autonomic skin innervation was compared between patients and controls using the Z statistics. A Z score <−1.65 with a p < 0.05 was considered significant.

Table 1 Patients clinical features
Full size table

Nerve conduction studies demonstrated a sensory axonal neuropathy in patients 1 and 2, with absent-evoked sensory action potential in sural nerves, whereas they were normal in patient 3 (Table 1).

Skin biopsies showed a significant ENFs reduction in thigh (5 ± 2; n.v. 23.8 ± 4.5) and leg (2.5 ± 1.3; n.v. 15.9 ± 3.9) in all patients, predominantly at the distal leg site (Table 2).

Table 2 ENF density and semiquantitative score of cholinergic and adrenergic innervations
Full size table

Adrenergic (DBH-ir) fibers were heavily expressed in arrector pilorum muscles, along major axis of blood vessels and around arteriovenous anastomoses (AVAs) in normal controls, while patients showed sparse adrenergic fibers, significantly reduced compared with controls and predominantly in the leg, often displaying morphological abnormalities (i.e., swelling and fragmentation) and deranged pattern of innervation (Table 2). Cholinergic (VIP-ir) fibers were abundant around sweat glands, encircling sweat tubules in normal controls, whereas patients showed poor and deranged cholinergic innervation associated with morphological nerve fibers abnormalities, predominantly in the leg (Table 2).

We studied somatic and autonomic innervation in three patients with relapsed MM treated with bortezomib or bortezomib + thalidomide. As already reported [7], we found a peripheral neuropathy involving large and/or small fibers predominantly in the leg, suggesting a length-dependent neuropathic damage [4, 7]. We also disclosed a severe reduction of autonomic fibres, both adrenergic and cholinergic, innervating epidermal annexes. Skin innervation abnormalities were likely induced by bortezomib therapy, because our patients developed neuropathic pain and autonomic symptoms after 3–5 months from the start of treatment and all patients reported an improvement of those symptoms after bortezomib withdrawal. However, we are not able to exclude that MM was a direct cause of peripheral neuropathy, since we have not done a skin biopsy before the start of bortezomib treatment. Involvement of the autonomic system by bortezomib-induced neuropathy has already been described [4, 8, 9], however, this is the first study reporting an objective analysis of peripheral autonomic fibres in patients treated with bortezomib. We found a reduced autonomic innervation score in both symptomatic and asymptomatic patients (number 2 of Table 1) suggesting that skin autonomic fiber loss may not directly express a specific clinical symptom and may precede clinical autonomic failure. Additional studies including a larger sample of patients, particularly with untreated MM, are needed to confirm the main results of our study.

In conclusion, our immunofluorescence study shows that bortezomib can cause a neuropathy involving somatic as well as autonomic small fibres with a length-dependent pattern.