Prion strains associated with iatrogenic CJD in French and UK human growth hormone recipients

Treatment with human pituitary-derived growth hormone (hGH) was responsible for a significant proportion of iatrogenic Creutzfeldt–Jakob disease (iCJD) cases. France and the UK experienced the largest case numbers of hGH-iCJD, with 122 and 81 cases respectively. Differences in the frequency of the three PRNP codon 129 polymorphisms (MM, MV and VV) and the estimated incubation periods associated with each of these genotypes in the French and the UK hGH-iCJD cohorts led to the suggestion that the prion strains responsible for these two hGH-iCJD cohorts were different. In this study, we characterized the prion strains responsible for hGH-iCJD cases originating from UK (n = 11) and France (n = 11) using human PrP expressing mouse models. The cases included PRNP MM, MV and VV genotypes from both countries. UK and French sporadic CJD (sCJD) cases were included as controls. The prion strains identified following inoculation with hGH-iCJD homogenates corresponded to the two most frequently observed sCJD prion strains (M1CJD and V2CJD). However, in clear contradiction to the initial hypothesis, the prion strains that were identified in the UK and the French hGH-iCJD cases were not radically different. In the vast majority of the cases originating from both countries, the V2CJD strain or a mixture of M1CJD + V2CJD strains were identified. These data strongly support the contention that the differences in the epidemiological and genetic profiles observed in the UK and France hGH-iCJD cohorts cannot be attributed only to the transmission of different prion strains. Supplementary Information The online version contains supplementary material available at 10.1186/s40478-021-01247-x.


Introduction
Prion diseases (otherwise known as transmissible spongiform encephalopathies) are a group of transmissible neurodegenerative disorders that occur naturally in man and a range of other mammalian species, including scrapie in sheep and goats, and chronic wasting disease in deer and elk. Prion diseases are characterized by the accumulation of a misfolded form of the normal cellular prion protein (PrP C ) in the central nervous system (CNS) [1]. This misfolded protein (commonly referred to as PrP Sc ) is considered to be the major, if not the only, component of the transmissible agents or prions that are responsible for these diseases [2]. Serial passage of sheep scrapie in mice by intracerebral inoculation identified the presence of distinct strains of the transmissible agent based on the individual biological properties in the mice that become stable on serial passage [3]. These properties include the disease incubation period and neuropathological phenotype, particularly the distribution and severity of spongiform change in the brain. In the apparent absence of any nucleic acid associated with the transmissible agent, these variations in strain properties have been accounted for by conformational variability in PrP Sc that is selfpropagating [2].
Human prion diseases occur in sporadic, genetic and acquired forms. The commonest of these is the sporadic form of Creutzfeldt-Jakob disease (sCJD), which has a worldwide distribution at a remarkably consistent incidence of 1-2 cases per million population per annum. The acquired forms of human prion disease include kuru, iatrogenic Creutzfeldt-Jakob disease (iCJD) and variant Creutzfeldt-Jakob disease (vCJD) [4].
The clinicopathological phenotype of human prion diseases is variable, particularly in sCJD. One major determinant of this variability is the naturally occurring polymorphism at codon 129 in the human prion protein gene (PRNP), which can encode either methionine (M) or valine (V), resulting in three possible genotypes: MM, MV and VV. The second major determinant is the isoform of the protease-resistant core of PrP Sc (referred to as PrP res ) which may be detected by Western blot analysis in brain tissue from affected patients. PrP res occurs as two major isoforms in sCJD cases that differ in the molecular weight of the unglycosylated fragments, designated Type 1 (21 kDa) and Type 2 (19 kDa). Distinct subtypes of sCJD have been recognized according to their clinical and neuropathological features, which largely correspond to the six possible codon 129 genotype/PrP res type combinations (MM1, MV1, VV1, MM2, MV2 and VV2) [5]. However, both type 1 and type 2 PrP res can be detected within either the same or different brain areas in up to 35% of sCJD patients [6,7]. It has been proposed that the different sCJD subtypes could result from the propagation of different strains of sCJD prions in patients. Experimental transmissions of sCJD brain homogenates to transgenic mice have demonstrated the existence of at least five strains of sCJD prions [8][9][10][11][12][13].
Although the precise relationship between prion strain diversity and the clinicopathological subtypes of sCJD remains imperfectly characterized, two strains named M1 CJD and V2 CJD seem to account for the most frequent forms of sCJD; M1 CJD is predominantly found in the brain of MM/MV1 sCJD patients, while the V2 CJD strain is generally found in VV/MV2 sCJD patients [8,9,11,14]. However, in up to 35% of sCJD patients the co-existence of both M1 CJD and V2 CJD strains was demonstrated by bioassay [11].
The first reported case of iatrogenic CJD (iCJD), occurring in a recipient of a corneal graft from a donor who had died from sCJD, was published in 1974 [15].
Since then, cases of iCJD have also been identified in small numbers of patients on whom neurosurgical instruments or intracerebral electrodes previously used on the brains of patients with CJD were subsequently used [16]. Larger numbers of iCJD cases (over 200 in each group) have occurred in patients inoculated with human pituitary-derived growth hormone (hGH), and in patients who received an implant of lyophilised human dura mater (hDM) during neurosurgery. Treatment of primary or secondary growth hormone deficiency in children with hGH was commenced initially in the 1950s on a small scale in a few countries. The subsequent clinical success of this treatment resulted its use on a much larger scale and in more countries. The first case of iCJD in a patient treated with hGH was reported in 1985, since when the use of hGH has ceased in many countries and replacement therapy with biosynthetic growth hormone was instigated [16].
The incidence of hGH-iCJD varies from country to country and appears to be related to the scale of the local hGH production process and the likelihood of prion contamination in the selection of autopsy cases for pituitary collection, e.g. the collection of pituitary glands from elderly patients. Since 1985, over 240 cases of iCJD in hGH recipients have been reported in several countries, with the largest numbers of cases occurring in France and the United Kingdom (UK). Four cases of iCJD in human pituitary gonadotrophin recipients have also occurred, all in Australia; one of these patients died in the UK [16].
Overall, 1849 patients from the UK and Ireland were treated with hGH produced in the UK from 1959 until 1985. Since 1985, 81 deaths from iCJD have occurred in this cohort. It has been established that one particular preparation of UK pituitary-derived hGH (the Hartree-modified Wilhelmi (HWP) preparation) had been administered to all hGH recipients who had developed iCJD, albeit from different batches, in varying quantities and over different time periods [17]. In France, 122 cases of iCJD in 1443 hGH recipients have occurred between 1991 and 2019. All were treated between December 1983 and July 1985, which has been identified as a high-risk period for hGH-iCJD in France [18,19].
Deaths from hGH-iCJD continue to occur in France and in the UK, with the most recent death occurring in 2019 and 2020 respectively, 35 years since hGH therapy ceased in these countries. A precise calculation of incubation periods in human pituitary hormone recipients is difficult, since the patients are often treated over a period of years; the time period from the mid-point of pituitary hormone treatment to the onset of clinical symptoms of iCJD is therefore used as an estimate for the incubation period.
A major factor influencing incubation periods in hGH-iCJD is the PRNP codon 129 polymorphism of the recipient. It has long been recognized that homozygosity at this locus may predispose to both iCJD and sporadic CJD (sCJD) [20]. In the healthy UK and French populations, the distribution of the individual genotypes at codon 129 of the PRNP gene is similar at around 40% MM, 10% VV and 50% MV [16,21].
Variations in the frequency of the three PRNP codon 129 polymorphisms (MM, MV and VV) and estimated incubation periods associated with each of these three genotypes have been reported in hGH-iCJD cohorts in different countries [16].
In France, the mean incubation period for hGH-iCJD has been estimated at 14.2 years. However, incubation periods differed according to the patient's age at start of treatment, with the shortest incubation periods observed in the group of patients aged between 10 and 13.7 years when treatment started. Incubation periods were also influenced by the patient's PRNP codon 129 genotype, with mean values of 12.6 years in MM and VV homozygotes and 17.6 years in MV heterozygotes [18].
The risk of developing iCJD in the UK hGH recipients was greatest in those patients who started treatment at ages 8-10 years [17]. Estimated incubation periods in the UK hGH-iCJD patients range from 7 to 40 years; these prolonged incubation periods are reminiscent of those occurring in kuru, where incubation periods of over 40 years have been reported [22]. The incubation periods in the UK hGH-iCJD patients are also influenced by the patient's PRNP codon 129 genotype, but in contrast to the French patients, UK VV homozygous patients had a mean incubation period of 14.3 years, while in MV heterozygous patients the mean incubation period was 23.4 years. The longest incubation periods in UK patients occurred in codon 129 MM homozygotes (mean value 30.8 years) [22].
These observations raised the hypothesis that different prion strains (most likely originating from undetected cases of sCJD in the donor population) contaminated the hGH preparations administered to recipients in the UK and in France, resulting in different incubation periods for each of the PRNP codon 129 subgroups between these countries [16,23]. This study aims to test this hypothesis by performing experimental transmissions of 22 hGH-iCJD brain homogenates from patients with all three codon 129 genotypes from France (11 cases) and the UK (11 cases) into a panel of well characterized human-PrP-expressing transgenic mice (tgHu), following an established methodology that has been used extensively to identify sCJD prion strains in a range of tissue samples from France, the UK and Spain [11,14]. Sporadic CJD cases originating from France and the UK were also used as controls in this study. This is the first study, to our knowledge, where the strain characteristics of prions in hGH-iCJD cases have been defined and compared to sCJD prion strains.
In clear contradiction to the initial hypothesis, the prion strains that were identified in the UK and the French hGH-iCJD cases were not radically different. In the vast majority of the cases originating from both countries, the V2 CJD strain or a mixture of M1 CJD + V2 CJD strains were identified by the strain typing in tgHu. These data strongly support the contention that differences observed in the epidemiological profiles between hGH-iCJD cases in the UK and France cannot be attributed solely to the transmission of different prion strains. Concerning the human CJD samples, in all cases informed consent for research was obtained and the material used had appropriate ethical approval for use in this project.

Ethical statement
France: human brain samples were obtained from the Brain Bank of CHU of Toulouse, Cardiobiotec (Centre de Resources Biologiques des Hospices Civils de Lyon) and former French national reference laboratory (CEA, Fontenay aux roses) under approval number AC62009-973/20-01-2010. Samples were pseudo-anonymized before dispatch.
UK: Human brain samples were obtained from the National CJD Research and Surveillance Unit Brain and Tissue Bank in Edinburgh, UK, which is part of the MRC Edinburgh Brain Bank. For the purposes of this study, samples were pseudo-anonymized using a Brain Bank reference number. All UK cases had informed consent for research and their supply and use in this study was covered by Ethics Approval (LREC 2000/4/157: National Creutzfeldt-Jakob disease tissue bank: acquisition and use of autopsy material for research on human transmissible spongiform encephalopathies, Professor James Ironside, amended date: 9th October 2007).

hGH-iCJD cases
22 cases of hGH-iCJD, each of which had frozen tissue (200 to 500 mg) available (cerebral cortex or cerebellar cortex), were included in this study. The patients originated from France and the UK. A diagnosis of iatrogenic CJD was made in each case according to established international criteria (ECDC 2021: https:// www. ecdc. europa. eu/ en/ infec tious-disea ses-public-health/ varia ntcreut zfeldt-jakob-disea se/ eu-case-defin ition). The UK cases are part of a larger series of UK hGH-iCJD cases that has been extensively characterised in terms of neuropathological features, PrP res profiles and prion protein amplification properties [24]. None of the patients had a familial history of prion disease and, in each case, the entire PRNP coding sequence was analysed [25,26]. For each patient the age at death and the estimated incubation period (based on individual hGH treatment histories) are indicated (Table 1).

Transgenic mouse lines
Tg340 and tg361 mouse lines that express human PrP methionine at codon 129 or valine at codon 129, respectively, in a prion protein gene knockout (PrP Ko ) background, were generated as previously described [9,27]. Both tg340 (tgMet) and tg361 (tgVal) are homozygous for human PRNP gene.

Mouse bioassays
Six-to ten-week-old female mice were anesthetized and inoculated with 2 mg of brain equivalent (20 µL of a 10% cerebral or cerebellar cortex homogenate) in the right parietal lobe using a 25-gauge disposable hypodermic needle.
Mice were observed daily and their neurological status was assessed weekly by qualified veterinarians. Three signs of neurological dysfunction (tremor, ataxia, difficulty righting from supine position, rigidity of tail, kyphosis, paralysis of the lower limbs or bradykinesia,) were necessary to score a mouse positive for prion disease [28].
When clinically progressive TSE disease was evident, the animals were euthanized and their brains harvested. Half of the brain from those animals that had displayed TSE clinical signs was fixed by immersion in 10% formol saline and the other half was frozen at − 20 °C. Tissues from animals found dead were frozen (no formalin fixation). Incubation period was expressed as the mean of the survival (time to death) days post inoculation (dpi) of all the symptomatic mice scored positive for PrP res , with its corresponding standard deviation.

Vacuolar lesion profiles
Hematoxylin-Eosin-stained paraffin-embedded brain tissue sections were used to establish standardized vacuolar lesion profiles in mice as previously described [3,31]. Each lesion profile was based on data obtained from 4 to 6 animals.

Reference M1 and V2 CJD strains
Successive 1/10 dilutions of 10% brain homogenate (frontal cortex) from a sCJD MM1 and a sCJD VV2) case were inoculated intra-cerebrally to tgMet (n = 6) and tgVal (n = 6). These data were already presented in a previous study [32] as cases 1 and 6, respectively. The brain of the last PrP Sc positive MM1 in tgMet and VV2 in tgVal mice (in highest dilution groups) were used to re-inoculate groups of tgMet (n = 12) and tgVal (n = 12). The brain of these tgMet and TgVal animals were pooled to constitute two stocks of reference material (named M1 CJD and V2 CJD ). The V2 CJD stock homogenate was end point titrated by bioassay in tgVal (inoculation of successive 1/10 dilutions of 10% homogenates-Additional file 1: Table S1).

Data availability
The authors confirm that the data supporting the findings of this study are available within the article and its supplementary material.

Results
22 hGH-iCJD cases were selected, which originated either from France (n = 11) or the UK (n = 11) and included all three possible PRNP genotypes at codon 129 (Homozygous Met 129 or Val 129 , Heterozygous Met 129 / Val 129 ) ( Table 1). In both the UK and France, the availability of hGH-iCJD cases from certain PRNP genotypes (such as MM in the UK or VV in France) with suitable biological (frozen) material was limited. Information on the selected 22 cases, including the estimated incubation period of the disease is included in Table 1.
From the frozen material available, tissue homogenates (10% weight volume) corresponding to either cerebral (frontal, parietal or temporal) cortex or cerebellar cortex from each hGH-iCJD case were transmitted (two iterative passages) to mice homozygous for either methionine (tgMet) or valine (tgVal) at codon 129 of human PrP (Table 1). These two tgHu PrP mouse lines have been shown to express approximately fourfold more PrP C in their brain than human brain tissue [9]. In parallel, French (n = 7) and UK (n = 4) cases of sCJD, classified as MM1, MV1, MV2 and VV2, were transmitted to tgMet and TgVal. (Table 2). Data from these sCJD transmissions have been previously reported in a recent study that aimed to type the prion strains in the brain of MM/MV1 and VV/MV2 sCJD affected patients [11]. Based on the transmission properties of incubation period, lesion profile and biological cloning of prion component present in different brain isolates, this study allowed the identification of two sCJD strains, named as M1 CJD and V2 CJD , that could be present as either pure components or as a mixture in the sCJD isolates ( Table 2).
In the French and UK hGH-iCJD isolates, three distinct transmission patterns were observed in mice inoculated with the panel of tissue homogenates.
This pattern was very similar to that observed in tgMet and tgVal inoculated with the V2 CJD strain (Table 3, Fig. 1) which was observed in the brain of some French and UK MV2 and VV2 sCJD-affected patients (cases 30-32) ( Table 2). The vacuolar lesion profile in the brain of the mice (after second passage) confirmed the propagation of the V2 CJD prion strain in both the tgMet and tgVal that had been inoculated with these 13 hGH-iCJD cases (Fig. 3).
The second transmission pattern observed was observed in one MM UK hGH-iCJD case (case 7). It was characterized by a short incubation period in tgMet (~ 200 dpi), and a longer one in tgVal (~ 300 dpi) ( Table 1, Fig. 1). Irrespective of the inoculated mouse line, a type 1 PrP res was observed in the brain of the inoculated animals (Table 1, Fig. 1). These features were similar to those associated with the transmission of the M1 CJD strain (Table 3) which could be observed in French and UK MM1 and in a part of the MV1 sCJD affected patients (cases 23-26) ( Table 2). The vacuolar lesion profile in the brain of the inoculated mice (after second passage) confirmed the propagation of M1 CJD prion strain in tgMet and tgVal inoculated with hGH-iCJD case 7.
In another case (case 22, French origin, VV genotype), available transmission results and lesion profiles in tgMet and tgVal, were also compatible with the presence of M1 CJD strain. However, incomplete second passage of this isolate in tgVal at the time of writing precluded drawing a definitive conclusion on the nature of the prion strain.
A third transmission profile was identified in French MM (n = 2, cases 4, 5) and MV (n = 1, case 12) cases as well as in UK MM (n = 1, case 6), MV (n = 1, case 13) and VV (n = 2, cases 20, 21) patients. This transmission pattern was characterized on second passage by a short incubation period in tgMet (~ 200-300 dpi), and a short incubation period in tgVal (about 160-200 dpi) (Table 1, Fig. 1). A type 1 PrP res accumulated in the brain of tgMet while type 2 PrP res was observed in tgVal mice (Fig. 2). In these isolates, the vacuolar lesion Transgenic mice that express the Met 129 (tgMet), Val 129 (tgVal) human PrP were inoculated intra-cerebrally (20µL per mouse) with human growth hormone (hGH) iatrogenic Creutzfeldt-Jakob brain tissue homogenates (frontal, temporal or parietal cortex: (cort) or cerebrellar cortex: (cereb)) from patients originating (orig.) from France (Fr), or the United Kingdom (UK). The hGH-iCJD patients displayed different PRNP genotypes at codon 129 (MM: homozygous Met 129 , VV: homozygous Val 129 , MV: heterozygous Met/Val 129 ). For each patient, the country of origin, the year of death and age in years at death are indicated. The estimated duration of the incubation period in years (based on the hGH treatment history) is also indicated. After the first and second passages, brain tissue from the clinically affected mice were pooled and used for a next passage in the same line. The PrP res WB isoforms (type 1 or type 2) identified in mouse brains are reported for each two passages. Survival times are shown as mean ± standard deviation (SD). 100% attack rate transmission were observed in all cases   heterozygous Met/Val 129 ) and PrP res Western blot isoforms (type 1 or type 2). After the first passage, brain tissue from clinically affected mice were pooled and used for a second passage in the same line. The PrP res WB isoforms (type 1 or type 2) identified in mouse brains are reported for each two passages. Survival times are shown as mean ± standard deviation (SD). 100% attack rate transmission were observed in all cases. ND: not done. At the exception of isolate 24, transmission data have already been use in a previous publication. The prion strain(s) identified in each isolate (strain typing based on survival time and vacuolar lesion profile in the brain) are indicated in the  1 Survival times of human PrP-expressing mice (tgHu) inoculated with growth hormone CJD cases originating from France and UK. Transgenic mice that express the Met 129 (tgMet), Val 129 (tgVal) human PrP were inoculated intra-cerebrally (6 mice, 20µL per mouse) with a 10% brain homogenate from hGH-iCJD cases of French and UK origin. Two iterative passages were performed in each mouse line (Table 1). For each passage, results are presented according to the country of origin (Fr and UK) and the PRNP codon 129 genotype of patients (homozygous Met129: MMheterozygous Met129/Val129: MV-homozygous Val129: VV). Survival time (mean ± SD in days post inoculation) in tg Met (○) and tg Val (∆). White/ black symbols correspond to a type 1/type 2 PrP res (as assessed by Western blot) in the brain of the mice respectively. Survival times in tgMet and TgVal associated with M1 CJD and V2 CJD cloned strains as well as an artificial mixture of V2 CJD + M1 CJD (10 -4 diluted) are included as reference (see Table 2 ) profiles in the TgMet matched with the profiles observed in mice inoculated with M1 CJD strain while in the tgVal mice it corresponded to the V2 CJD strain (Fig. 3). Strikingly, the transmission of artificial mixtures containing V2 CJD cloned strain and a low dilution (10 -3 to 10 -4 ) of the M1 CJD cloned strain in tgMet and TgVal mice resulted in similar transmission patterns (Table 3, Fig. 3).

Representativeness of the panel of hGH-iCJD cases
hGH-iCJD cases are considered to be the consequence of childhood treatment with human growth hormone, prepared using pituitary glands collected from donors that were either affected with or at a late stage of incubation of sCJD. France (n = 122) and the UK (n = 81) account for the largest numbers of hGH-iCJD patients identified worldwide. The epidemiological profile of the hGH-iCJD cases in these countries differs in the distribution of PRNP genotypes at codon 129 and in the mean incubation period in each genotype group. These differences between the UK and France have been proposed to have resulted from contamination with different prion strains [16,23,24].
Bioassays in reporter animal models, followed by the phenotyping of the propagated prions (vacuolar lesion profile in the brain and incubation period) remains the gold standard approach for the characterization of prion strains. Over the last decade, transgenic mice that express different human PrP variants at codon 129 (tgHu) have been shown to be valuable models in which to discriminate between different human prion agents and have provided valuable insights in the diversity of the prion strains responsible for sCJD [8][9][10][11][12][13]. These mouse models have allowed the identification of five different prion strains associated with sCJD among which two, named M1 CJD and V2 CJD , are thought responsible for the most frequent clinico-pathological forms of sCJD (observed in MM/MV1 and VV/MV2 patients, respectively) [8][9][10][11]. These same animal models also demonstrated that in around 30% of the MV1, Fig. 2 PrP res western blot profiles in the brains of human PrP-expressing mice (TgHu) inoculated with human growth hormone iatrogenic CJD (hGH-iCJD) cases. Transgenic mice that express the Met 129 (tgMet) or Val 129 (tgVal) human PrP were inoculated intra-cerebrally (6 mice, 20µL per mouse) with a 10% brain homogenate from hGH-iCJD cases of French and UK origin. Two iterative passages were performed in each line (Table 1). After each passage and in each mouse line the isoform (type 1/type 2) of the PrP res was determined in mice brain by SDS-PAGE and WB with the anti-PrP monoclonal antibody Sha31 (epitope YEDRYYRE). A PrP res type 1 isoform (MM1 sCJD isolate) and type 2 isoform (VV2 sCJD isolate) were included as controls on each gel. The WB results obtain in each line are reported in Table 1  Table 3 Bioassay transmission in tg Hu mice of artificial V2 CJD /M1 CJD strains mixture M1 CJD and V2 CJD strains were obtained by the endpoint titration of a MM1 and VV2 sCJD isolate in Met 129 (tgMet) or Val 129 (tgVal) human PrP-expressing mice respectively (see Additional file 1: Table S1). Brains from tgMet (inoculated with M1 CJD strain) and tgVal (inoculated with V2 CJD strain) were used to produce stock solutions (10% tissue homogenates). 1/10 dilution series (in phosphate buffer saline) of each stock solution were prepared. M1 CJD /V2 CJD strain mixtures were obtained by mixing equal volume of each component at the chosen dilutions. Samples were then transmitted (two iterative passages) to tgMet and tgVal (intra-cerebral route, 20µL per mouse). Brains from first passage positive mice (PrP res presence in the brain) were pooled and used for the second passage. The PrP res WB isoforms (type 1 or type 2) identified in mouse brains are reported.
Survival times (time to death in days) are shown as mean ± standard deviation (SD). n/n0: number of diseased/number of inoculated mice. *PrP res WB profile of the mice in the group were not homogenous and the individual incubation periods of each animal are presented. ND: not done. These data were already used in a previously published study [9] Artificial strain mixture composition MV2 and VV2 sCJD patients both M1 CJD and V2 CJD prions were present as a mixture [11]. Transmission of the 11 French and 11 UK hGH-iCJD cases in mice expressing valine 129 (tgVal) and methionine 129 (tgMet) human PrP variants in this study, represents an unprecedented effort to document the nature of the prion strains responsible for this iatrogenic form of the disease. Our study was primarily designed to test the hypothesis that different prion agents were responsible for the hGH-iCJD outbreaks in France and in the UK. While the design of the experiments was fit for this purpose, the materials (cost and duration of bioassay, availability of material from patients) and ethical considerations limited the number of hGH-iCJD cases that could undergo strain typing by bioassay.

Prion strains in the UK and French hGH-iCJD cases
The prion strains that were identified in brain isolates prepared from French and UK hGH-iCJD cases following bioassay in tgHu mice models correspond to prion strains that were previously identified in sCJD patients in France and in the UK using the same mouse models [11].
In contradiction to the initial hypothesis, the prion strains that were identified in the UK and the French GHcases were not radically different. In the vast majority of the UK (10 out 11) and French (10 out 11) cases, the V2 CJD strain or a mixture of M1 CJD + V2 CJD strains were identified by the strain typing bioassays.
A pure V2 CJD strain was observed in all three PRNP genotypes in French hGH-iCJD patients and in two PRNP genotypes (MV and VV) in the UK patients. In both countries, M1 CJD + V2 CJD strain mixtures were identified in the three different PRNP genotypes. A pure M1 CJD strain was identified in a single UK patient (case 7, MM genotype) and suspected in a French patient (case 22, VV genotype -definitive transmission results not yet available).
These data strongly support the contention that the difference of epidemiological profile observed between the UK and the French hGH-iCJD outbreaks cannot be attributed to the transmission of different prion strains.

V2 dominant presence is concordant with previous investigations in the UK
In the UK, several studies have addressed the question surrounding the prion strains responsible for hGH-iCJD, by the characterisation of the neuropathology features and/or the PrP res biochemical properties in the brain of hGH-iCJD affected individuals, including those cases investigated by bioassay in this study [22,24,33]. So far, no comprehensive description of the neuropathological and biochemical PrP res properties in the French hGH-iCJD cohort is available in the literature.
In a large number of VV and MV UK hGH-iCJD patients, neuropathological features that are considered characteristic of VV2 sCJD cases (including plaque-like deposits) were identified. Along with the biochemical profile of the UK VV and MV hGH-iCJD cases, this leads to the proposal that pituitary glands collected from VV2 sCJD case(s) might have played a preponderant role in the UK hGH-iCJD outbreak [22,24].
The results of this study supports this hypothesis, with the unambiguous identification of the V2 CJD strain that was obtained by bioassay of brain homogenate from 6 out of the 9 VV and MV UK hGH-iCJD cases.
Similarly, the identification of MM1-like sCJD neuropathological, biochemical and protein misfolding cyclical amplification (PMCA) features reported in case hGH-iCJD 21 in the earlier study by Ritchie et al. concurs with the pure M1 CJD strain phenotype we observed in tgHu mice inoculated with cerebral cortex homogenate from the same patient (case 7 in this study). The other UK PRNP codon 129 MM hGH-iCJD case reported by  had atypical neuropathological and biochemical features along with PMCA features that suggested the influence of a V2 strain. This case was also included in the present study as case 6, which had transmission characteristics different from case 7, suggesting a mixed M1 CJD /V2 CJD strain effect. The transmission findings in cases 6 and 7 therefore provide a direct in vivo correlation that confirms the in vitro findings of Ritchie et al. [24].

Coexistence of several strains in French and UK hGH-iCJD cohorts
The identification of mixed M1 CJD and V2 CJD strains in both the UK and the French hGH-iCJD cases raises the question of the origin of this strain diversity.
The pooling of pituitary glands during hGH production from several individuals affected with different strains of sCJD may be one explanation for this phenomenon. However, since the co-existence of M1 CJD and V2 CJD strains have been reported in up to 35% of sCJD cases, the use of pituitary gland from a single donor may also explain the presence of the two strains [7,11,34].
In France, 12 batches of extracted hGH, all produced between January 1982 and December 1985, were identified as the potential source of infection. According to available data, a batch was in average produced from 1000 to 1500 pituitary glands, meaning that tissue from up to 18 000 individuals may have contributed to the atrisk batches [18].
In the UK it is estimated that around 200,000 pituitary glands were used to produce the hGH extracts that were used to treat patients that later developed hGH-iCJD [17].
The sCJD prevalence in industrialized countries is estimated to be approximatively 1.5 cases per million people and per year [35]. In people aged over 50, sCJD frequency significantly increases to reach 5.5 to 7.5 case per million people per year in individuals aged between 65 and 79 [36,37]. However, in France and in the UK (each with a similar population size), between 600,000 and 700,000 deaths are recorded each year, with 80-130 sCJD cases diagnosed annually. Based on these figures, the possibility exists that more than one infectious pituitary gland could have entered in the production of the at-risk hGH batches in both France and the UK.

Epidemiological profiles in both countries
Assuming that the same prion strains are responsible for the UK and French hGH-iCJD cases, the question remains of how the epidemiological differences occurring between the two hGH-iCJD case cohorts may be explained.
The frequency of Met/Met 129 , Met/Val 129 and Val/ Val 129 individuals in the general population and in their distribution in sCJD cases are similar in France and in the UK [16]. The hGH therapeutic scheme in both countries, including the route of administration (in general intramuscular) and the average age of administration were also similar, making these parameters unlikely explanations for the observed epidemiological differences [17][18][19]23].
In experimental TSE models, longer and more widespread incubation periods are classically observed when animals are exposed to decreasing doses of infectivity [38]. The average incubation duration for hGH-iCJD patients from each PRNP codon 129 genotype were significantly longer in the UK patients (VV: 14.3 years-MV: 23.4 years-MM: 30.8 years) when compared with French patients (VV: 9 years-MV: 17.6 years-MM: 12 years) [16,22]. These elements support the hypothesis that French hGH-iCJD affected patients could have been exposed to higher infectious doses than the UK patients.
Although the relevance of findings in experimental TSE models to the physiopathology of CJD in humans is uncertain, they could provide some valuable insights on the relative abilities of sCJD strains to propagate in hosts expressing different human PrP variants.
The end point titrations of V2 CJD and M1 CJD in tgMet and tgVal indicated that the M1 CJD strain displays a 1000-fold lower capacity to propagate in Val 129 than in Met 129 PrP-expressing host (Additional file 1: Table S1). Conversely, one infectious dose (ID 50 ) of V2 CJD strain (as measured by the intracerebral route in tgVal) transmits with shorter incubation period in tgVal than 10 7.2 ID 50 M1 CJD strain (as measured by the intracerebral route in tgMet) [11,32].
When tgHu mice are co-infected with M1 CJD and V2 CJD prions, the nature of the strain(s) that propagate in the brain of the recipients depends on the PRNP genotype at codon 129 and to the specific amount of M1 CJD and V2 CJD infectivity in the inoculum, resulting in the presence of either M1 CJD + V2 CJD strain mixture or of a pure M1 CJD or V2 CJD strain in the infected host brain (Table 3) [11]. Based on these elements, a scenario where M1 CJD and V2 CJD would display different abilities to propagate in an individual according to his genotype at codon 129 and the infectious titre of each strain seems plausible. At low infectious doses, as seems possible in the UK hGH situation, the V2 CJD strain would dominate, mostly in VV individuals.
The isolation of a sub-dominant M1 CJD strain in homozygous (cases 4 and 5) and heterozygous (case 12) Met 129 patients indicates that French hGH-iCJD patients were exposed to this prion strain. The transmission of the M1 CJD and V2 CJD strain in tgMet mice support the view that the M1 CJD strain, even if present at very low level in the extractive hGH, should propagate as the dominant strain in the homozygous Met 129 patients. The identification of the V2 CJD strain as dominant prion strain in the brain of the homozygous Met 129 French hGH-iCJD patients is therefore surprising.
These observations could be explained by the effects of a peripheral (non-CNS) route of transmission that would differentially impact on the relative efficiency of M1 CJD and V2 CJD transmission by the intramuscular route of exposure in hGH recipients, resulting in a slow or inefficient transmission of the M1 CJD strain.
In a given host, the inoculation by a peripheral route of two prion strains can result in radically different transmission profile, depending on their capacity to replicate early in the lymphoid tissues before neuroinvasion. The inoculation of the HY prion strain in hamsters by the intracerebral, the intraperitoneal or the oral route resulted in all three instances in a highly efficient transmission (100% attack rate) of the disease. The inoculation of the DY prion strain in hamsters by the intra cerebral route also resulted in a prion disease. In sharp contrast, the inoculation of a high infectious dose of DY prions in hamsters by the intraperitoneal or the oral route failed to transmit the disease [39,40]. HY but not DY are able to replicate at detectable levels in the lymphoid tissue. Interestingly, V2 CJD strains appeared to replicate at higher levels than M1 CJD strain in the lymphoid tissue of tgMet mice [10].
Further investigations will be necessary to confirm the influence of the intramuscular route of infection in modulating M1 CJD and V2 CJD strain transmission efficiencies. Beyond this, the characterization of the PrP res and measurement of the infectivity levels in the pituitary glands of sCJD patients and the impact of hGH extraction processes on M1 CJD and V2 CJD strains from the pituitary glands (removal of infectivity) would be very helpful for appreciating the potential exposure conditions to prions in hGH-treated patients. Ultimately, the direct characterisation of prion infectivity levels and strains in the French and UK hGH-batches that appear to have resulted in the transmission of hGH-iCJD would be very informative.
Although numerous elements are still lacking to fully understand the difference existing between the French and UK hGH-iCJD case cohorts, the results that we report in this study indicate that the nature of the prion strains involved are not fundamental to the observed differences in epidemiological profiles.
Finally, the number of hGH-iCJD cases that we investigated (n = 22) should be considered as relatively limited when compared to the total number of hGH-iCJD cases identified in France (n = 122) or the UK (n = 81). Therefore, we cannot be certain that the panel of hGH-iCJD cases that we investigated necessarily provides a full picture of the prion strain diversity in the French and the UK hGH-iCJD cohorts.