Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

10.1 Early-Stage Malignant Mesothelioma, Including the Concept of Mesothelioma In Situ and the Distinction from Reactive Mesothelial Hyperplasia

A subserosal multipotential fibroblastoid cell (SMFC) has been invoked in the past as the stem cell for mesothelial renewal following injury resulting in destruction of the surface mesothelium and as the progenitor cell for the development of malignant mesothelioma (MM) [16,17]. According to this theory, an origin of MM from such SMFCs could explain the observation that the time required for mesothelial regeneration remains constant, irrespective of the area of the injury, and also the biphasic differentiation characteristic of approximately 30% of MMs, within a range of about 25–35% [54,59]. (If this model is correct, it follows that MM is an invasive neoplasm ab initio, with no in situ phase of development.) Based, in part, on experimental models of mesothelial healing following injury without disruption of the submesothelial basal lamina [177,178,181], and on detection of early-stage MMs of epithelial type – where mesothelial atypia appeared to be predominantly in situ, in the absence of any radiological or gross anatomical evidence of pleural thickening or nodularity – Whitaker et al. [180] refocused upon the mesothelium itself as the reserve cell for “normal” mesothelial cell turnover and for healing, and as the progenitor cell for MM, advancing the concept of mesothelioma in situ (MMIS). (For a detailed discussion of mesothelial cell turnover and renewal, see Whitaker et al. [181]; the constancy of the time for mesothelial healing according to this model is largely explicable by detachment of mesothelial cells from viable mesothelium and their random reimplantation over the denuded area.) These authors [180,183] defined MMIS as the replacement of benign surface mesothelium by mesothelial cells with markers of malignancy – with the consequent problem of identifying an acceptable and consistently reproducible marker of neoplastic change. Whitaker et al. [180] described 22 cases of mesothelial proliferation that had presented in a “conventional” fashion, in the form of a pleural effusion with either no identifiable pleural tumor or only tiny nodules at thoracoscopy (Fig. 10.1). The diagnosis in a number of cases was established by existing cytologic criteria. Whitaker et al. [180] suggested that the markers for MMIS in pleural biopsies included the following [60,61] (Figs. 10.2–10.5):

Fig. 10.1
figure 1

Pleurectomy specimen from a patient who presented with a massive pleural effusion. No distinctive abnormality was seen at thoracoscopy but multiple random biopsies revealed an extensive atypical mesothelial proliferation, in situ in most areas of the biopsies, but with small foci of invasion. A pleurectomy was subsequently carried out, and in the surgical specimen, small foci of white invasive tumour were found, some of which extended into sub-pleural adipose tissue (From Battifora and McCaughey [11], Fig. 4-6. ; figure originally contributed by Dr. Douglas Henderson, Adelaide, Australia)

Full size image
Fig. 10.2
figure 2

Atypical mesothelial proliferation at the surface of a pleural biopsy, with the formation of at least two small papillary structures. Invasive mesothelioma was found in other areas of the same biopsy

Full size image
Fig. 10.3
figure 3

Atypical mesothelial proliferation in a pleural biopsy, with an exophytic papillary architecture at the surface. The lesion is entirely in situ in distribution in this field, but superficial invasion into the submesothelial fibrous tissue was found in other areas of this biopsy (Reproduced from Hammar et al. [54], p 643; Fig. 43.95A. ©Springer Science+Business Media 2008. With kind permission of Springer Science+Business Media)

Full size image
Fig. 10.4
figure 4

Exophytic in situ mesothelial atypia: superficial but undoubted invasion was found in other areas of the same biopsy. Positive labeling of the lesional cells for CK5/6 (Reproduced from Hammar et al. [54], p 620; Fig. 43.61. ©Springer Science+Business Media 2008. With kind permission of Springer Science+Business Media)

Full size image
Fig. 10.5
figure 5

Same pleural biopsy shown in Fig. 10.3, depicting the mesothelial atypia at higher magnification (Reproduced from Hammar et al. [54], p 643; Fig 43.95B. ©Springer Science+Business Media 2008. With kind permission of Springer Science+Business Media)

Full size image

  • Abnormal architecture of the mesothelium at the surface of the affected pleural tissue. Such architectural abnormalities included noninvasive, linear, papillary, and tubulopapillary patterns, sometimes with a complex exophytic architecture (Figs. 10.2–10.4).

  • Substantial cytological atypia (Fig. 10.5). However, these authors [183] also considered that other cases might occur where there is substantially less cytological atypia, so that such cases would be diagnosable (if at all) only by ancillary techniques: among those techniques they included strong linear labeling for epithelial membrane antigen (EMA) – see later discussion on Diagnostic Biomarkers.

  • Absence of background inflammation as an incitement for mesothelial hyperplasia.

The major problem in translating this concept into diagnosis in practice is that there is overlap in the degree of cytological atypia between benign reactive mesothelial proliferations (RMPs) versus mesothelioma [20,26,27,61]. In the absence of any consistently reliable immunohistochemical or molecular biomarker for discrimination between benign and malignant applicable to everyday diagnosis, Whitaker et al. [180] and Henderson et al. [60,61] emphasized that the only consistently reliable marker for mesothelioma as opposed to RMP is the presence of acceptable neoplastic invasion (Fig. 10.6) – as opposed to benign entrapment of mesothelium within pleural fibrous tissue as a consequence of inflammation – either in the same biopsy, a different biopsy taken at a different time, or at autopsy (see also [54]). Accordingly, Henderson et al. [60] commented in 1997:

Fig. 10.6
figure 6

Early-stage invasive mesothelioma of epithelial type, with infiltration into the sub-mesothelial fibrous tissue. This pattern is considered inconsistent with benign mesothelial entrapment as part of a fibro-inflammatory process, although there was no evidence of invasion into subpleural adipose tissue. There is only low-grade cytological atypia. This biopsy showed no evidence of exudative inflammation (Reproduced from Hammar et al. [54], p 645; Fig. 43.100. ©Springer Science+Business Media 2008. With kind permission of Springer Science+Business Media)

Full size image

We caution against rash or premature diagnosis of mesothelioma in situ from conventional light microscopy examination of biopsy tissue, taking into account the overlap in the cytologic abnormalities that occur in reactive mesothelioses versus mesothelioma. However, [findings suggestive of a component of MMIS] (especially in conjunction with effusion fluid cytology) may delineate “at risk” patients with “early” stage disease who require further investigation and follow-up. Because of the minimal and perhaps predominantly in situ tumor burden, the mesotheliomas may also be amenable to new modalities of therapy, and some of our “in situ” patients have had prolonged survivals.

Some authorities [27], including the International Mesothelioma Panel [20] consider that noninvasive atypical mesothelial proliferations should be designated simply as an atypical mesothelial proliferation (AMP). We would discourage use of the term atypical mesothelial hyperplasia, because by definition hyperplasia denotes a benign process and in effusion fluid cytology specimens invasion cannot be assessed and it often cannot be assessed in small or superficial biopsies. Even so, complex exophytic mesothelial proliferations (Figs. 10.3 and 10.4) do not usually occur as part of benign inflammation-induced mesothelial proliferations; such appearances (Figs. 10.2–10.4) raise a suspicion of MM where an invasive component (if present) has not been sampled by the biopsy. Hammar et al. [54] consider that such complex and exophytic AMPs should not be dismissed as benign; they require close clinical follow-up and/or further cytologic or biopsy investigation. That is, a noninvasive AMP in biopsy tissue or an effusion fluid cytology specimen does not by itself represent a treatable disorder – unless carefully correlated with the clinical and in particular the radiological findings or in exceptional circumstances where biopsy is contraindicated – instead, it is a finding that requires follow-up and/or further investigation.

Although it has been claimed that there is no direct proof that in situ mesothelial atypia together with areas of invasive MM represents a single neoplastic lesion [27], Simon et al. [151] reported a single case of “mesothelioma in situ” in association with focal early-stage invasive MM. They investigated the lesion by laser microdissection and comparative genomic hybridization and found similar chromosomal alterations in both the areas of in situ mesothelial atypia and in the foci of early invasive mesothelioma. Accordingly, in the areas of “mesothelioma in situ” they recorded losses at 3p, 5q, 6q, 8p, 9p, 15q, 22q, and Y, with a gain on 7q; in the area of early invasive mesothelioma there were losses at 3p, 5pq, 6q, 8p, 9p, 15q, and 22q with no gains; more advanced mesothelioma showed losses at 1p, 4pq, 6q, 9p, 13q, 14q, and 22q, with gains at 1q, 7pq, and 15q. In a study of 31 cases of MM for EMA, p53 and bcl-2 expression, Cury et al. [34] reported that the seven cases of MM with “… both in situ and invasive mesothelioma, the in situ elements showed similar staining patterns to the invasive epithelioid elements” (see following discussion).

Hammar et al. [54] continue to regard “mesothelioma in situ” as a useful concept for the development of MM. By refocussing attention on the mesothelium itself as the target for neoplastic transformation, this model foreshadows the potential for diagnosis of noninvasive mesotheliomas, with the hope of more effective therapy in the future. They [54] continue to believe that the expression “mesothelioma in situ” represents a valid retrospective diagnosis in cases where at least early-stage invasive MM has been demonstrated.

Hammar et al. [54] set forth the following guidelines and caveats as useful in the differential diagnosis of mesothelial lesions where the discrimination between MM and hyperplasia is problematic:

  • Correlation of the histologic appearances with the findings on pleural effusion fluid cytology and with any abnormalities revealed by imaging studies, such as chest radiographs or CT scans: in this context, the radiologic investigations in some cases can constitute a surrogate for the histological identification of invasion, in a patient with an AMP as shown by cytological examination of effusion fluid [22] (see later discussion).

  • Invasion of subpleural adipose tissue (or deeper chest wall structures) or invasion into peripheral lung parenchyma by either an epithelioid or sarcomatoid mesothelial proliferation is usually a decisive indicator of malignancy, for either epithelial or sarcomatoid MM respectively (provided that benign displacement by antecedent procedures such as thoracentesis or biopsy can be ruled out). Immunohistochemical staining for cytokeratins can often highlight genuine neoplastic invasion (especially for desmoplastic MMs [54,98], for assessment of invasion into subpleural adipose tissue).

  • Even in the absence of infiltration into subpleural tissues, MM is still diagnosable from superficial invasion within the pleural fibrous layer, provided that the pattern of infiltration is characteristic or diagnostic of neoplastic invasion, as opposed to a tangential plane of section through pleural tissue folded upon itself, artefact or benign entrapment of mesothelial cells as part of an organizing fibro-inflammatory process (please see below). In our experience this problem represents one of the frequent reasons for referral of biopsy tissue for further opinion: “it looks like it ought to be a mesothelioma, but I can’t find invasion into fat.”

  • Hammar et al. [ 54 ] emphasize the importance of correct orientation for pleural biopsy tissue as a prelude to histological sectioning, so that the tissue is embedded on edge with en profile sectioning (en face sections are frequently problematical as to what represents true invasion as contrasted to a tangential plane of section). Whenever sufficient pleural membrane is available (for example, pleurectomy/decortication specimens and some video-assisted thoracoscopy [VAT] biopsies) and especially when the tissue is received unfixed, it is useful to prepare a Swiss Roll from the biopsy, followed by fixation and then slicing of the Swiss Roll like a loaf of bread, so that the pleural membrane is sectioned en profile. This exercise has the added benefit that large areas of the pleura can be sampled, with a minimal number of tissue blocks. Whenever there is any doubt as to whether the histological appearances represent pseudo-invasion versus genuine neoplastic invasion, the appearances should be considered inconclusive [54].

  • Is it benign inflammation-induced entrapment of mesothelium or MM? Most inflammation-driven reactive mesothelial hyperplasias are noninvasive, but hyperplastic mesothelial cells can become entrapped within some organizing serosal inflammatory processes – an occurrence that requires distinction from genuine invasion. A florid fibrinous or neutrophilic inflammatory reaction is one marker for the likelihood of benign entrapment (but cases of proven invasive MM with prominent associated exudative inflammation are encountered occasionally). Hammar et al. [54] suggest that such benign entrapment results from burying of the plane where the surface mesothelium is normally located, by a layer of inflammatory exudate that extends over the surface of the membrane, with subsequent organization; in other words, it is the surface of the pleura that has moved inward, into the lumen of the serosal cavity – a process that they [54] sometimes liken to the shrinking of the Aral Sea (the Aral Sea Effect). For the distinction between entrapment and invasion, immunohistochemical staining for cytokeratins (or calretinin) is often of value, because it delineates a clear linear boundary between the entrapped mesothelial cells versus the deeper tissues [54].

  • There is a consensus that neoplastic invasion remains the mainstay for diagnosis of early-stage MMs of epithelial type [20,26,27,54,61,180] (Fig. 10.6): whenever there is any doubt as to whether genuine invasion is present or not, Hammar et al. [54] assign a less-than-definite confidence index for a diagnosis of MM (for example, “possible,” “probable,” or “highly probable,” depending on the degree of doubt) – on the principle that if the lesion is MM “it will declare itself as such soon enough, whereas, inappropriate overdiagnosis of mesothelioma can lead to erroneous cytotoxic chemotherapy or even radical surgery, together with the anguish that a diagnosis of mesothelioma usually entails” (primum non nocere).

  • Even when invasion cannot be found in a biopsy sample, there are several findings in combination that are suspicious of MM – requiring clinical follow-up or further investigation – although each is nondiagnostic by itself [54]. Such findings include:

    • The extent of the mesothelial proliferation

    • A complex exophytic or papillary architecture at the surface of the pleura, in the absence of exudative inflammation

    • Prominent cytological atypia

    • Focal necrosis within sheets of proliferative mesothelial cells in the pleura

    • Prominent intracytoplasmic vacuoles devoid of mucin-like content

    • Strong thick linear labeling for EMA with antibodies based on the E29 clone (see later discussion)

A consensus document from the International Mesothelioma Interest Group (IMIG) states that a diagnosis of MM “… has to be made with certainty …”, so “that a cytologic suspicion of MM is followed by tissue confirmation that must be supported by both clinical and radiological data” [69]. However, the 2007 statement on MM from the British Thoracic Society (BTS) [22] takes a less restrictive approach to diagnosis: “If the clinical, radiological, and cytological results … support a diagnosis of mesothelioma, then this can be accepted.… A biopsy is required if the diagnosis is not clear after the pleural tap and a CT scan.”

10.2 Biomarkers for Early-Stage Epithelioid Malignant Mesothelioma Versus Reactive Mesothelial Hyperplasia

As indicated in the preceding discussion, there is a consensus at present that neoplastic invasion represents the only consistently reliable marker for the discrimination between benign versus malignant mesothelial proliferations. Nonetheless, the potential of several biomarkers has been investigated, for the diagnosis of epithelioid MM as opposed to RMPs, for example, in effusion fluid cytology preparations – with mixed results.

10.2.1 Epithelial Membrane Antigen (EMA)

In their paper emphasizing the concept of mesothelioma in situ, Whitaker et al. [180] observed thick linear labeling of the mesothelial cells for EMA in 17 of 22 such cases (see also Wolanski et al. [183] and Segal et al. [144]); in contrast, proven benign reactive mesothelial proliferations usually showed no significant labeling or only patchy weak labeling [60]. These findings seem to be applicable only to EMA antibodies based upon the E29 clone. In this context, Saad et al. [134] studied EMA expression in 20 cases of reactive mesothelial proliferation (RMP) and 20 cases of MM, using antibodies based on the Mc5 and E29 clones: for the Mc5 clone, 14/20 cases of MM (70%) and 12/20 cases of RMP (60%) showed positive staining. However, for the E29 clone, the corresponding results were 15/20 for MM (75%) and 0/20 for RMP. Saad et al. [134] concluded that EMA antibodies based on the E29 clone are a reliable discriminator between RMP and MM, and Simon et al. [151] commented along similar lines.

Cury et al. [34] investigated EMA, p53 and bcl-2 expression among 31 cases of MM (plus four biopsies initially reported as suspicious, from patients who later developed overt MM) and 20 cases of RMP, as well as 14 cases of benign pleural fibrosis (BPF). Thirty-four out of 35 cases of MM showed diffuse linear staining for EMA (97%). Of the 20 cases of RMP, 5 (25%) showed “focal weak staining” for EMA, and 6/14 cases of BPF also stained for EMA (43%). They [34] concluded that “… strong diffuse linear staining for EMA is a good marker of malignancy when differentiating epithelioid malignant mesothelioma and mesothelioma in situ from reactive mesothelial hyperplasia, although weak focal staining may occur in reactive conditions.”

Attanoos et al. [6] investigated 60 cases of pleural MM and 40 cases of RMP for desmin, EMA), p53, bcl-2, P-glycoprotein and platelet-derived growth factor receptor (PDGF-R) β-chain: 48/60 MMs were positive for EMA (80%) in comparison to 8/40 RMPs (20%); 6/60 MMs (10%) showed expression of desmin, versus 34/40 RMPs (85%). These authors [6] concluded: “Desmin and EMA appear to be the most useful markers in distinguishing benign from malignant mesothelial proliferations. Desmin appears to be preferentially expressed in reactive mesothelium and EMA appears to be preferentially expressed in neoplastic mesothelium.”

In summarizing the usefulness of EMA immunostaining for the distinction between MM and RMP, the following points and caveats seem to be worth emphasis: [144]:

  • Diffuse strong thick linear staining of single cells and cell groups for EMA is a useful pointer on a probability basis for mesothelial neoplasia – especially in effusion cytology – but it is not decisively diagnostic in isolation. In some studies [164,179,180,183], about 75–90% of MMs or more showed this pattern of EMA labeling [69], whereas labeling in RMPs is usually undetectable or weak [40,90,99,150,151,164,174,179,182].

  • EMA staining can be used as a cytology screening test for patients with a pleural effusion and a past history of asbestos exposure or for effusions that appear to contain “reactive” mesothelial cells [144] – that is, as an indicator for further investigation and follow-up of the patient.

  • Negative EMA staining does not exclude a diagnosis of MM, and in biopsy tissue undetectable EMA expression is not uncommon in the deep zones of invasive MMs [54].

  • Lymphoplasmacytic cells often show positive EMA staining, so that it is imperative to show that the cell proliferation is mesothelial in character [144].

10.2.2 GLUT-1

GLUT-1 is one of a family of 14 glucose transmembrane transporters that facilitate the entry of glucose into cells [77]. Although immunohistochemically undetectable in normal epithelial tissues and benign tumors, GLUT-1 is expressed in a variety of malignant neoplasms. In one study on pleural effusion fluids [2], GLUT-1 was expressed in 28/39 of cases of malignant effusion (72%) – 100% from the ovary, 91% from the lung, 67% from the gastrointestinal tract, and 12% from the breast – but none (0/25) of the benign effusions expressed GLUT-1.

Kato et al. [77] studied GLUT-1 expression in 48 cases of MM, 40 RMPs, and 58 cases of carcinoma of lung. GLUT-1 expression as demonstrated by linear membrane-related staining was observed in all 48 epithelioid, biphasic, and sarcomatoid MMs, whereas GLUT-1 was undetectable in all 40 RMPs: in the 11 biphasic MMs, staining for GLUT-1 was found in the epithelioid areas in 10 (91%) and in the sarcomatoid areas in 7 (64%). GLUT-1 staining was also found in 56/58 carcinomas of lung (96.5%). The authors [77] concluded that GLUT-1 is a sensitive and specific discriminator between MM and RMP, but it cannot distinguish MM from lung carcinomas. Husain et al. [69] also refer to the abstract for a study carried out by Acurio et al. [1], which revealed negative reactions in all 40 benign mesothelial tissues (20 normal and 20 RMPs); of the 45 MMs, 9 were negative (20%), 34 showed weak positivity (53%) and 12 were strongly positive (27%). Husain et al. [69] concluded that GLUT-1 staining when positive is a helpful marker for MM in comparison to RMP, but it is unhelpful when negative.

Shen et al. [147] compared EMA with GLUT-1 (both monoclonal and polyclonal antibodies) and the X-linked inhibitor of apoptosis protein (XIAP) in 35 MMs and 38 cases of “benign effusion” and they concluded that EMA “… is a better marker than XIAP or GLUT-1 for the diagnosis of MM.”

10.2.3 Bcl-2

Bcl-2 is a proto-oncogene that inhibits apoptosis and thereby promotes survival of individual cells. Detectable overexpression [42] and direct mutations of bcl-2 in MM are rare [110] (unlike many other tumors, including follicular lymphoma and even lung carcinoma [12,44,111], where overexpression is common and may be predictive of a poor prognosis). Segers et al. [145] investigated bcl-2 expression in 62 cases of MM and 44 cases of non-neoplastic mesothelium: cytoplasmic staining was found in 5 MMs (8%) and the benign cases were “… not immunoreactive.” All 15 pleural MMs and 15 RMPs studied by Attanoos et al. [6] were negative for bcl-2, and these authors concluded that bcl-2 is of “… no use in distinguishing reactive from neoplastic mesothelium, although more formal evaluation of these markers is required.”

10.2.4 p53

The tumor suppressor gene p53 is an inducer of cell cycle arrest and is maintained at low levels in normal unstressed cells, whereas “stress” can induce increased levels of p53 and result in cell cycle arrest and apoptosis. P53 is rarely detectable in normal cells (related to its short half-life) but increased expression of p53 is common in malignant tumors, related to mutations that render p53 nonfunctional and resistant to degradation, as opposed to an increase in functional p53. In MM, such mutations of p53 are rare [121], but the p53 pathway is affected by numerous mutations.

The presence of p53 has been reported in between 25% and 97% of MMs, whereas p53 was found in between 0% and 82% of reactive mesothelial lesions examined [6,23,39,71,83,101,102,109,128,143]. For example, Cury et al. [34] found positive nuclear staining for p53 in 30/31 cases of MM (97%), with greater frequency of positivity in epithelioid than in sarcomatoid tissue, and “occasional nuclear positivity” was found in 13/20 RMPs (65%). Therefore, this antibody does not appear to be useful for the distinction of benign from malignant mesothelial lesions. A relationship between p53 expression and prognosis has not been identified.

10.2.5 X-Linked Inhibitor of Apoptosis Proteins (XIAP)

Wu et al. [184] reported that labeling for XIAP (a member of a family of inhibitors of apoptosis proteins: IAPs) also shows promise in distinguishing benign from reactive pleural effusions. In a study of 116 samples of cell block material from 82 pleural effusions, 22 ascites, 11 pelvic/peritoneal washes, and 1 pericardial effusion, these authors [184] found positive particulate cytoplasmic staining for XIAP in 4/5 MMs, as well as variable positivity in 33–100% of carcinomas according to the site of origin – for example, in all 13 ovarian carcinomas and 9/11 carcinomas of lung (82%) – but all 4 colonic carcinomas were negative and the 35 benign effusions were “virtually XIAP-negative except for two cases (6%).” In a further study on XIAP, Wu et al. [186] found that all nine samples of normal mesothelium were negative, and only one of 13 RMPs showed weak positivity in less than 10% of cells; of 31 MMs, 25 (81%) displayed XIAP positivity. Wu et al. [186] concluded that strong staining for XIAP allowed a distinction between MM and RMPs, especially for small samples and problematical cases.

Lyons-Boudreaux et al. [96] investigated XIAP (and other markers that included calretinin, D2-40, WT1 and MOC31) in five MMs, 48 adenocarcinomas, and 19 benign effusions and found that most MMs stained for XIAP (80%) as well as some adenocarcinomas (51%) and rare benign effusions (11%). They [96] concluded that XIAP is not a sensitive marker for malignancy and has limited value in cytology.

As indicated above, Shen et al. [147] found EMA to be a better marker than XIAP for MM versus RMP.

Based on studies of mesothelial cell lines, XIAP has been mooted (along with IAP-1 and IAP-2, and p21/WAF1, p27/KIP1 and survivin) as a potential target for treatment of MM using the proteasome inhibitor bortezomib alone or in combination with standard chemotherapy [48].

10.2.6 P-Glycoprotein (P-170)

P-glycoprotein plays a role in cell membrane transport, and its expression has been associated with resistance to chemotherapy [146]. Expression of P-170 glycoprotein has not been identified in normal mesothelium, but it has been found in a high proportion of MMs [146], albeit with no apparent effect on patient survival [152]. Ramael et al. [129] detected P-170 in most cases of MM studied, whereas it was not found in normal mesothelium, and Segers et al. [146] found that 54/57 mesothelioma cases showed immunoreactivity for P-170. In a study of 36 cases of MM in comparison to normal mesothelium, Soini et al. [152] detected P-170 in 61% of the MMs but not in normal mesothelial cells. However, in a later study of 15 MMs and 15 RMPs, Attanoos et al. [6] reported that P-glycoprotein was expressed in only 2/15 of the MMs (13%) and none of the RMPs: they [6] concluded that P-glycoprotein (as well as bcl-2 and PDGF-R β-chain) appeared to be of no value for the distinction of MM from RMP, although further studies were required.

10.2.7 Neural Cell Adhesion Molecules (NCAMs): CD56

Neural cell adhesion molecules (NCAMs) corresponding to CD56 antigen are a family of closely related cell surface glycoproteins, thought to play a role in the development of neural cells and the interactions between them. Lantuéjoul et al. [89] studied 26 cases of epithelial, biphasic, and sarcomatoid MM for NCAM reactivity using the 123C3 antibody, in comparison to normal mesothelium and 50 non-small cell lung carcinomas divided evenly between adenocarcinomas and squamous cell carcinomas. Although normal mesothelium was “negative,” staining for NCAM was recorded in 19 of the 26 MMs of all histological subtypes (73%). Although this finding raises the possibility that CD56 may be useful for discrimination between RMPs versus MM, there appears to be too little data on NCAM/CD56 expression in MM and mesothelial hyperplasia to justify inclusion of NCAM/CD56 antibodies in everyday diagnostic practice, until further and more extensive studies become available.

10.3 Screening for Malignant Mesothelioma and Prognostic Biomarkers: Serum Levels of Soluble Mesothelin-Related Proteins (SMRPs), Osteopontin (OPN), Megakaryocyte Potentiating Factor (MKPF) and CA125

10.3.1 Introductory Remarks on Screening for Malignant Mesothelioma

As a matter principle and logic, screening for any disease such as cancer is justifiable only when a certain set of circumstances prevail, apart from any considerations of cost [142]:

  • The disease occurs with reasonable frequency in the population for which screening is proposed (i.e., it must not be one of great rarity). Because MM is rare in the general population – with an annual incidence of about one case or less per million of the population without identifiable asbestos exposure [22] – screening would be justifiable only for high-risk populations such as middle-aged to older men with substantial (usually occupational) exposure to asbestos.

  • The disease in question must result in substantial morbidity or mortality (clearly the case for MM).

  • The screening procedure(s) must have reasonable specificity and sensitivity for the detection of the cancer at early presymptomatic stage; in other words, the procedure should have a reasonable positive predictive value for the detection of the cancer in question.

  • Ideally, the screening procedure(s) should be noninvasive or only minimally invasive: as a follow-on to this principle, the morbidity and even mortality from the screening test(s) – and any subsequent test(s) necessary to establish a definitive diagnosis for those who test positively for the initial screening – must be taken into consideration and balanced against the potential benefits of therapy for any early-stage disease so detected.

  • One or more effective therapeutic interventions exist for the early-stage cancer, with substantially improved outcomes in comparison to the prognosis for those whose cancer is diagnosed at a later and symptomatic stage. (Apart from radical pleuropneumonectomy – applicable for only a minority of MM patients, even when the disease is detected at an early stage – this is not the case for MM and present-day chemotherapy results in only a slight improvement in medial/mean survival times [142], but this situation may change).

Therefore, screening specifically for MM, even in high-risk groups, seems unjustifiable at present [21,51,52,123,142,165] – although groups with past occupational asbestos exposure may be under intermittent clinical and radiological surveillance (or screening) for other asbestos-related disorders such as asbestosis and lung cancer [35,37,43,62,63,85,86,127,153,162,192], (even so, the value of screening programs for lung cancer among former asbestos workers remains debatable [100]). The anguish that can result from a false-positive screening result for MM and the consequent requirement for further investigative procedures also needs to be taken into account [13].

10.3.2 Radiological Screening for MM

The radiographic appearances of pleural MM can vary from essentially normal with early-stage disease, to complete opacification of the affected hemithorax, with confluent nodular pleural thickening sometimes accompanied by extension along interlobar fissures and encasement of the lung, often with contraction of the hemithorax; depending on the size of the MM and its associated effusion, the mediastinum may be displaced to one side or the other [22,92,104]. A pleural effusion of variable volume without pleural thickening is often the only detectable radiological abnormality in cases of symptomatic early-stage MM [180], and as such the finding of an effusion by itself lacks specificity.

Conventional chest x-rays (CXRs) and computerized tomography (CT) have not been shown to be effective screening procedures for early-stage MM [142]. For example, Fasola et al. [43] studied 1,045 asbestos-exposed workers aged 40–75 years (median 58 years), using CXRs and low-dose CT (LDCT) scans. Pleural abnormalities were identified in 70% by LDCT (44% by CXR); ten non-small cell lung carcinomas and one thymic carcinoid tumor were found (1%) but no case of pleural MM was diagnosed. There were “11 false-positive results.”

10.3.3 Soluble Mesothelin-Related Proteins (SMRPs)

A significant recent development for the investigation of MM has been the demonstration of elevated serum SMRP levels in MM patients [29,31,131,132], and a commercially marketed test for SMRP is now available in the form of a two-step immunoenzymatic assay in an ELISA format (mesomark™) [ 14 ].

Mesothelin is a cell-surface glycoprotein on normal mesothelial cells and can be found in several cancers [105,114,188], including mesotheliomas with an epithelioid component [87,105,114,188], ovarian adenocarcinomas [32,133,188,189], squamous and large cell carcinomas and adenocarcinomas of lung [64,87,105], pancreatic adenocarcinomas [9,55], and some gastrointestinal cancers [133]. The protein product of the mesothelin gene appears to be a 69–71 kDa polypeptide anchored to the cell membrane by a glycosyl-phosphatidyl-inositol (GPI) linkage [97,133,139]; this anchored protein can be cleaved by a protease to yield a 31 kDa soluble protein called megakaryocyte potentiating factor (MKPF) secreted into the blood [97,133,139], and a 40 kDa protein named mesothelin, attached to the cell membrane [139]. The normal biological function of mesothelin is unclear and mice with a knock-out of the mesothelin gene(s) show no obvious phenotypic abnormality [189]. Although attached to the cell membrane, mesothelin can be shed like other cell membrane proteins and Robinson et al. [29,31,131,132] have described a 42–44 kDa soluble mesothelin/MKPF-related protein (SMRP) in sera from patients with pleural MM and also ovarian carcinoma. The process underlying the release of SMRP from cell membranes may be related to an abnormal splicing event that leads to synthesis of a secreted protein (release) or to enzymatic cleavage of membrane-bound mesothelin (ectodomain shedding), and Sapede et al. [139] found evidence that both mechanisms are implicated.

Robinson et al. [31,131] detected SMRP using the OV569 monoclonal antibody – which is used together with another monoclonal antibody, 4H3, for the commercially marketed mesomark™ test [14]. However, others [56,148,149] appear to have used different antibodies to mesothelin, making it difficult to compare their results with those for other studies where the mesomark™ test [14] was used, for example, Scherpereel et al. [141] and Park et al. [122]. Robinson et al. [131] found elevated blood SMRP levels in 37/44 patients previously diagnosed with MM (sensitivity = 84%) as opposed to one of 22 lung cancers (histologic types not specified) and seven out of 40 asbestos-exposed control patients (three of these subjects developed MM 15–19 months after the SMRP sample had been taken). In a more recent (2006) publication from the same laboratory, Creaney et al. [31] reported the results as nanoMoles (nM), with a mean value of about 15.3 ± 20.5 nM in the mesothelioma group, in comparison to a level of approximately 0.9 ± 0.8 nM for healthy controls.

Beyer et al. [14] investigated serum SMRP levels in 409 apparently healthy individuals, 177 patients with nonmalignant disorders and 500 cancer patients (88 of whom had pleural MM). The 99th percentile level for the reference group was 1.5 nM/L, in comparison to a mean level of 7.5 nM/L (95% CI = 2.8–12.1) for the 88 mesothelioma patients. The SMRP levels were increased in 52% of the MM patients and 5% of asbestos-exposed individuals.

In another series, Scherpereel et al. [141] reported blood SMRP levels in 74 mesothelioma patients, 35 patients with secondary carcinomas in the pleura and 28 cases of benign pleural abnormalities associated with asbestos exposure (BPA). They [141] found that serum SMRP levels were significantly higher for epithelioid MMs than for biphasic or sarcomatoid MMs. They [141] also found that the median value for patients with pleural MM was 2.05 ± 2.5 nM/L, in comparison to a level of about 1.0 ± 1.8 nM/L for the metastatic carcinoma group, and in the BPA cases the level was approximately 0.55 ± 0.6 nM/L. Scherpereel et al. [141] commented that serum SMRP levels had a poor capacity for discrimination between pleural MM and secondary carcinoma, related to high SMRP levels in some of the carcinoma patients. They [141] commented further that pleural biopsy tissue remained the “gold standard” for the diagnosis of pleural MM, and in 2007 Scherpereel and Lee [140] added the comment that the “… proposed markers [SMRP, osteopontin and megakaryocyte potentiating factor] have insufficient accuracy to replace cytohistology as the gold standard for diagnosis for mesothelioma.”

In a large-scale prospective study of serum SMRP concentrations among 538 asbestos-exposed subjects attending the Dust Diseases Board in Sydney, Australia, Park et al. [122] found a mean SMRP levels of 0.8 ± 0.45 nM in 223 healthy asbestos-exposed individuals; 15 had elevated SMRP levels (2.8%); [30] one subject had lung cancer, but none was diagnosed with MM (individuals with SMRP levels ≥2.5 nM were investigated further by CT scanning and positron-emission tomography). Subjects with pleural plaques had a slightly higher mean concentration of SMRP than those without – a finding thought to be explicable by low-grade pleural inflammation related to the plaques [122]. Park et al. [122] concluded that a high false-positive was observed for SMRP levels and that it seems “… unlikely to prove useful for screening for MM.”

In 2009, Creaney et al. [30] reviewed the usefulness of blood SMRP levels for detection of MM, in comparison to osteopontin and megakaryocyte potentiating factor, and they concluded that at present soluble mesothelin remains the best biomarker for MM, but is beset with “… a lack of sensitivity for early-stage disease and for all malignant mesothelioma histologies ….”

In 2007, Creaney et al. [32] had also reported mesothelin levels in effusion fluids from 52 patients with pleural MM, as opposed to 56 patients with cancers other than mesothelioma and 84 with benign pleural effusions. Significantly greater pleural fluid concentrations of mesothelin were found in the MM patients than in either of the other two groups, with a specificity of 98% and a sensitivity of 67% for the MM group in comparison to those with non-neoplastic effusions. In seven of ten cases, mesothelin levels were elevated before the diagnosis of MM was made (by 0.75–10 months); four out of eight such cases had elevated mesothelin concentrations in the effusion fluid but not in the serum. The highest mesothelin levels were found in peritoneal fluid in patients with ovarian carcinoma. Significant differences in the mean mesothelin values in pleural effusion fluid were found for epithelial (47 ± 1.0 nM), biphasic (30 ± 0.8), and sarcomatoid (4.5 ± 1.4) MMs; for pleural sarcomatoid MMs the mesothelin concentrations were not significantly different from those in patients with nonmalignant effusions. MM patients with high concentrations of mesothelin in effusion fluid had a median survival of 14 months, as opposed to 8 months for those with low mesothelin levels – probably reflecting MMs with an epithelial component as opposed to sarcomatoid mesotheliomas.

Therefore, the following conclusions can be drawn:

  • Blood SMRP levels are elevated in most cases of epithelioid MMs, but other cancers can also be associated with elevated serum SMRP concentrations, including lung and, in particular, ovarian cancers, as well as apparently benign disorders.

  • The SMRP levels appear to be greatest for advanced-stage epithelioid MMs, with suboptimal sensitivity for the detection of early-stage MM.

  • Epithelioid MMs are associated with higher SMRP levels in serum and effusion fluid than biphasic or sarcomatoid MMs; for sarcomatoid MMs, the mean effusion fluid SMRP levels appear to be no greater than for benign effusions.

  • For patients with proven MM, low concentrations of SMRP in blood or effusion fluid appear to represent a marker for a poor prognosis, presumably correlating with the histological subtype and corresponding to predominantly sarcomatoid MMs.

  • As indicated in the 2007 BTS statement on MM [22], its diagnosis remains an essentially clinicopathological exercise.

  • Serum SMRP levels cannot replace cytologic or biopsy diagnosis of MM, except in unusual circumstances (e.g., a frail elderly patient whose physical condition contraindicates biopsy, or for whom past biopsies have been nondiagnostic, but who has high serum SMRP levels, such as levels >15 nM/L).

  • Serial assays of serum SMRP levels may find a role as an indicator of prognosis for MM and as a means to assess its progress or response to treatment.

10.3.4 Serum Osteopontin (OPN) Levels

The significance of serum osteopontin (OPN) levels as a marker for MM is more problematic and doubtful than testing for serum SMRP concentrations [140], with a reported sensitivity of about 47% for the detection of MM [33]. An acidic glycoprotein normally synthesized by osteoblasts – like angiopoietin-1 (ANG-1) also produced by osteoblasts – OPN (SPP1) [72] is said to be a “constraining factor” [57] on hemopoietic stem cell proliferation in the bone marrow. Elevated blood OPN levels have been recorded in patients with MM [124], but elevated levels have also been recorded in a variety of other disorders that include carcinomas of the head and neck region [41,173] and cervix [173], as well as lung [45], ovarian [7], gastric [185], and hepatocellular carcinomas [79]. Elevated OPN levels have also been found in patients with inflammatory bowel disease [106].

Therefore, it appears that serum OPN levels have poor sensitivity and specificity for the detection of MM [49,51,52], but serial serum OPN assays may find a role in assessment of the progress of MM and its response to treatment [24,50,130,140].

10.3.5 Megakaryocyte Potentiating Factor (MKPF)

As discussed in the preceding section on SMRP, MKPF appears to be closely related to SMRP [97,133,139]. It lacks specificity for the detection of MM [140], with poor sensitivity for the detection of non-epithelioid MMs [49], and Creaney et al. [33] found that it had a sensitivity of only 34%. Iwahori et al. [73] found that MKPF was of greater diagnostic value for MM than SMRP and that these two markers had about equal specificity. Even so, assays of serum MKPF appear to have no advantage over SMRP for the detection of MM; like SMRP and OPN, serial measurements of MKPF may be of value in assessment of the progress of MM and its response to treatment [113,130,140], perhaps in conjunction with those other markers and CA125.

10.3.6 CA125

Immunohistochemical investigation of tissue sections for CA125 has no value in the discrimination between MM and adenocarcinomas developing at various anatomic sites, such as those arising in the ovary, lung, and breast [5,10,87,195]. For example, Bateman et al. [10] found that 15/17 cases of MM labeled for CA125 (88%) in comparison to 7/14 cases of adenocarcinomas metastatic to lung and pleura (50%). Attanoos et al. [5] recorded positive immunostaining for CA125 in 19/20 ovarian papillary serous adenocarcinomas (95%) and 2/3 primary peritoneal serous adenocarcinomas, in comparison to 8/32 peritoneal MMs (all in females). In a Japanese study on 90 epithelioid MMs and 51 adenocarcinomas of lung, Kushitani et al. [87] found that 85% of the MMs and 80% of the adenocarcinomas were positive for CA125. In a further study on effusion fluids, Zhu and Michael [195] found positive staining of all 20 metastatic ovarian carcinomas for CA125, in comparison to 8/13 adenocarcinomas of lung (62%) and 6/13 cases of metastatic breast carcinoma (46%).

However, there is evidence that assays of serum CA125 levels are useful and sensitive for the assessment of the progression of MM and, therefore, its prognosis or for its response to treatment. Hedman et al. [58] found that serum CA125 concentrations increased as the disease progressed, whereas stable disease was accompanied by a decrease in CA125 levels. In a Turkish study on 11 peritoneal MMs, Kebapci et al. [78] found that the mean serum CA125 level was 230 U/mL, within a range of 19–1,000 U/mL (this study gave a normal reference range of 1.2–32 U/mL). In a later study from Italy on 60 cases of peritoneal MM, Baratti et al. [8] recorded a baseline sensitivity of 53% for serum CA125 in the MM patients: in patients who underwent debulking surgery the serum CA125 concentration fell in 21/22 patients who had elevated baseline levels, but it stayed high in all 9 patients with grossly persistent MM, and elevated CA125 levels developed in all 12 patients who developed progressive disease after the surgery and other treatment.

Therefore, there is reasonable evidence that serum CA125 levels represent a sensitive but nonspecific marker for MM, and that serial measurements of the serum levels are a useful means for monitoring the progression of MM or its response to therapeutic measures, especially when the results are correlated with other serum markers as discussed above.

10.3.7 Summary

The serum biomarkers discussed above have the advantage that they represent even less invasive studies than thoracentesis or serosal-surface biopsies, but they are beset with problems of specificity for MM and insensitivity for early-stage disease and non-epithelioid subtypes of MM. At present they cannot replace conventional cytological and biopsy diagnosis of MM, except as probability markers in unusual circumstances, for example, when biopsy is contraindicated. However, either individually or in combination, assays for these proteins may be useful for the monitoring of diagnosed MMs and for the assessment of responsiveness (or lack of it) to treatment strategies. As newer treatments are introduced for MM they may assume increasing importance to assess the effectiveness of such treatment, especially in clinical trials.

10.4 Aquaporins and Malignant Mesothelioma

Over recent years, it has been shown that the transport of water across cells is not explicable by simple diffusion driven by osmotic gradients, but instead is regulated and facilitated by a superfamily of membrane-related proteins known as the aquaporins (AQPs) [80,84]. The AQPs appear to represent an ancient group of proteins that developed at an early stage of evolution and they have been found not only in mammals, but also in amphibia, insects, plants, and microorganisms [18,82]. At least 13 AQPs have been identified (AQP0 to AQP12) [70], which show differential expression in various mammalian tissues [80,91,166,169,172]. As examples, AQP1 is expressed in the endothelial cells that line small blood vessels and it mediates proximal tubule fluid reabsorption in the kidney, the secretion of aqueous humor in the eye, and also cerebrospinal fluid, and lung water homeostasis [80]. AQP2 mediates vasopressin-dependent renal collecting duct water permeability [18,80] and AQP4 is abundant in brain [80], whereas AQP5 influences fluid secretion in salivary and lacrimal glands and is abundant in alveolar epithelium of the lung [80,166]. The importance of AQPs is demonstrated by the fact that water permeability driven by osmosis between the gas-exchange membranes of the lung is reduced by a factor of 10 if AQP1 or AQP5 are deleted, and it is reduced even more when AQP1 and AQP4 or AQP1 and AQP5 are deleted together [171]. In this context, the function of AQPs has been investigated using AQP-knockout mice, and Verkman et al. [166,167,171,172] have developed and studied transgenic mice that lack AQPs 1, 3, 4, and 5. Various phenotype abnormalities were found in the null mice: in the kidney, deletion of AQP1 or AQP3 resulted in polyuria, but AQP4 deletion resulted in a mild concentrating defect only. Deletion of AQP5 caused defective saliva production. In the brain, deletion of AQP4 conferred protection from brain swelling induced by acute water intoxication.

The lung expresses several AQPs: [19,171] AQP1 is found in vascular endothelium, whereas AQP3 appears to be localized to the epithelium lining large air passages and AQP4 in large and small airway lining cells. AQP5 has been found in alveolar epithelium. AQP1 has also been demonstrated in the mesothelium of the pleura and peritoneum in both experimental models [75,76,81,95,112,193], and for humans [36,47,88,93,155]. Song et al. [154] found that achievement of osmotic equilibrium for pleural fluid took place rapidly in wild-type mice (50% equilibration in <2 min) but was slowed in AQP1 null mice (to less than 25%).

More recently, the study of AQPs has moved from the realm of normal physiology to that of pathology [3,82,91], although the study of AQPs in various disease processes is still in its infancy. AQP2 is the vasopressin-regulated water channel implicated in some hereditary and acquired renal diseases affecting urine-concentrating ability [167]: AQP11-null mice die from uremia as a result of polycystic kidneys [70], whereas AQP2-null humans suffer from hereditary non-X-linked nephrogenic diabetes insipidus [170]. AQP4 appears to play an important role in cerebral edema [4,15,46,117,118,120,135,136,158], and antibodies to AQP4 are implicated in the pathogenesis of neuromyelitis optica [74,107,108,125,126,138,156,157,159,160,163,175,176].

In addition, there is evidence that AQP expression can influence the pathogenesis, growth, and metastatic potential of tumor cells that express AQP water channels (in both stromal vascular endothelium and/or the neoplastic cells themselves, in a variety of tumors) [25,38,65,103,115,116,137] and AQP1 appears to be related to angiogenesis in tumors [28]. For example, Hoque et al. [65] found that AQP1 as assessed by immunohistochemical staining – in several types of primary lung tumors that included 16 squamous cell carcinomas, 21 adenocarcinomas, and 7 so-called bronchioloalveolar carcinomas (BACs) – was overexpressed in 62% (13/21) and in 75% (6/8) of cases of adenocarcinoma and BAC, respectively, whereas all cases of squamous cell carcinoma and normal lung tissue were negative. The authors [65] concluded that: “Forced expression of full-length AQP1 cDNA in NIH-3T3 cells induced many phenotypic changes characteristic of transformation … although further details on the molecular function of AQP1 related to tumorigenesis remain to be elucidated, our results suggest a potential role of AQP1 as a novel therapeutic target for the management of lung cancer.”

Others have also suggested that AQPs may represent a target for treatment by AQP inhibitors/blockaders that have been identified [53,66–68,94,119,161,168,187,190,191,194], by way of target inhibition of AQPs themselves or growth factors such as vascular endothelial growth factor (VEGF) that appear to be closely associated with mesothelial-related growth.

We have recently carried out preliminary investigation of AQPs in pleural MM, based upon two observations: (1) as indicated above, AQP1 is expressed in the pleura and peritoneum, not only in the endothelium lining submesothelial blood vessels but also in the mesothelium itself and (2) even early-stage MMs, with apparently minimal tumor bulk, usually present with a pleural effusion that may be massive. Therefore, we postulated that pleural MMs may be accompanied by overexpression of AQP1 or the acquisition of other AQPs. Our preliminary immunohistochemical studies have identified consistent strong membranous expression of AQP1 with apical prominence by the tumor cells (labeling in stromal blood vessels is also seen) in epithelioid MMs (Fig. 10.7), with weaker and inconsistent expression of AQP9. Interestingly, so far we have found little or no labeling in sarcomatoid mesotheliomas or the sarcomatoid component of biphasic tumors. At present it is unclear whether this reflects approximately “normal” (or even subnormal) AQP1 expression per unit cell, within an expanded cell population, or whether it represents overexpression by individual tumor cells. However, in pleural effusion fluids, it appears that in malignant mesothelial cells this pattern is also seen (Fig. 10.8). Further work will be required to investigate the potential uses of this new marker.

Fig. 10.7
figure 7

AQP1 expression in an invasive MM of epithelioid type, predominantly membrane-related

Full size image
Fig. 10.8
figure 8

AQP1 expression in a pleural effusion fluid of MM of epithelioid type. Similar distribution in labelling as in the histological section (Fig. 10.7) is seen. Preliminary studies suggest that labelling may be more prominent in malignant lesions versus RMPs, but further studies are needed to confirm that impression

Full size image